AI Trajectory


Trajectory: (Expand/Collapse All)
Seq Timestamp Type Name Duration
0/0 2026/09/25 02:07 flow patch-triage
2m Args:
null

Results:
{
  "EnableConfigs": [
    "VXLAN",
    "INET",
    "IPV6"
  ],
  "FocusSymbols": [
    "vxlan_changelink",
    "vxlan_config_apply",
    "vxlan_dev_configure",
    "vxlan_xmit",
    "vxlan_rcv",
    "vxlan_vnilist_update_group",
    "vxlan_update_default_fdb_entry",
    "vxlan_vnifilter_dump",
    "vxlan_multicast_join",
    "vxlan_multicast_leave",
    "vxlan_free_dev"
  ],
  "KMSANReasoning": "The patch refactors vxlan device configuration (`vxlan-\u003ecfg`) from an embedded struct to a dynamically-allocated, RCU-protected pointer (`struct vxlan_config __rcu *cfg`) to allow atomic configuration updates under RCU. It also replaces `vxlan-\u003edefault_dst` with fields directly in `cfg` and adds RCU dereference helpers across vxlan core, mlx5, and mlxsw drivers.\n\nEvaluation of uninitialized memory risks:\n1. All dynamically allocated `struct vxlan_config` instances are created with `kmemdup` from a `struct vxlan_config conf` that is explicitly zero-initialized via `memset(conf, 0, sizeof(*conf))` in `vxlan_nl2conf()`.\n2. In `vxlan_fill_info()`, `struct ifla_vxlan_port_range ports` contains only two `__be16` fields (`low` and `high`) without padding, both of which are fully assigned before `nla_put()`.\n3. In `mlxsw_sp_bridge_vlan_aware_vxlan_join()`, `struct mlxsw_sp_nve_params params` has all its active fields explicitly assigned, and is solely consumed internally within the mlxsw driver without exposure to userspace.\n4. No uninitialized memory is read, copied to userspace, or branched upon.\n\nThe primary bug risks associated with this refactor involve RCU pointer lifetimes, use-after-free of old configuration objects during grace periods, and NULL pointer dereferences. These classes of bugs are thoroughly covered by KASAN, PROVE_RCU, and standard kernel debugging facilities, not KMSAN. Therefore, a dedicated KMSAN fuzzing session is not warranted.",
  "NeedsKMSAN": false,
  "Reasoning": "The patch refactors the VXLAN driver by converting the embedded vxlan_config in struct vxlan_dev to a dynamically allocated RCU-protected pointer (vxlan-\u003ecfg). It updates device configuration, changelink, transmission (vxlan_xmit), reception (vxlan_rcv), multicast, and VNI-filter operations to access configuration under RCU, and adds memory management (kmemdup, kfree_rcu, and priv_destructor vxlan_free_dev) and rollback handling. These reachable networking changes introduce subtle lifetime and concurrency considerations that warrant fuzzing via netlink and packet I/O.",
  "WorthFuzzing": true
}

1/1 2026/09/25 02:07 action read-patch-diff
0m Args:
null

Results:
{
  "PatchDiff": "commit c6e7846ceee7f17a881c3bd9be86fc5863645aba\nAuthor: syz-cluster \u003ctriage@syzkaller.com\u003e\nDate:   Fri Sep 25 02:07:19 2026 +0000\n\n    syz-cluster: applied patch under review\n\ndiff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c\nindex 7a18a469961db..467fbe43b89e9 100644\n--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c\n+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c\n@@ -241,9 +241,16 @@ static bool mlx5e_tc_tun_encap_info_equal_vxlan(struct mlx5e_encap_key *a,\n static int mlx5e_tc_tun_get_remote_ifindex(struct net_device *mirred_dev)\n {\n \tconst struct vxlan_dev *vxlan = netdev_priv(mirred_dev);\n-\tconst struct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n+\tconst struct vxlan_config *cfg;\n+\tint ifindex = 0;\n \n-\treturn dst-\u003eremote_ifindex;\n+\trcu_read_lock();\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\tif (cfg)\n+\t\tifindex = cfg-\u003eremote_ifindex;\n+\trcu_read_unlock();\n+\n+\treturn ifindex;\n }\n \n struct mlx5e_tc_tunnel vxlan_tunnel = {\ndiff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c\nindex 52c2fe3644d4b..21e7b499f3879 100644\n--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c\n+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c\n@@ -59,8 +59,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,\n \t\t\t\t\t   const struct mlxsw_sp_nve_params *params,\n \t\t\t\t\t   struct netlink_ext_ack *extack)\n {\n-\tstruct vxlan_dev *vxlan = netdev_priv(params-\u003edev);\n-\tstruct vxlan_config *cfg = \u0026vxlan-\u003ecfg;\n+\tconst struct vxlan_config *cfg;\n+\tstruct vxlan_dev *vxlan;\n+\n+\tvxlan = netdev_priv(params-\u003edev);\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \tif (vxlan_addr_multicast(\u0026cfg-\u003eremote_ip)) {\n \t\tNL_SET_ERR_MSG_MOD(extack, \"VxLAN: Multicast destination IP is not supported\");\n@@ -92,6 +95,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,\n \t\treturn false;\n \t}\n \n+\tif (test_bit(VXLAN_DEV_F_MDB, \u0026vxlan-\u003eflags)) {\n+\t\tNL_SET_ERR_MSG_MOD(extack, \"VxLAN: MDB entries are not supported\");\n+\t\treturn false;\n+\t}\n+\n \tswitch (cfg-\u003esaddr.sa.sa_family) {\n \tcase AF_INET:\n \t\tif (!mlxsw_sp_nve_vxlan_ipv4_flags_check(cfg, extack))\n@@ -148,8 +156,11 @@ static void mlxsw_sp_nve_vxlan_config(const struct mlxsw_sp_nve *nve,\n \t\t\t\t      const struct mlxsw_sp_nve_params *params,\n \t\t\t\t      struct mlxsw_sp_nve_config *config)\n {\n-\tstruct vxlan_dev *vxlan = netdev_priv(params-\u003edev);\n-\tstruct vxlan_config *cfg = \u0026vxlan-\u003ecfg;\n+\tconst struct vxlan_config *cfg;\n+\tstruct vxlan_dev *vxlan;\n+\n+\tvxlan = netdev_priv(params-\u003edev);\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \tconfig-\u003etype = MLXSW_SP_NVE_TYPE_VXLAN;\n \tconfig-\u003ettl = cfg-\u003ettl;\ndiff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c\nindex fe45e533a4b2e..de60b698bea98 100644\n--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c\n+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c\n@@ -2513,15 +2513,17 @@ mlxsw_sp_bridge_vlan_aware_vxlan_join(struct mlxsw_sp_bridge_device *bridge_devi\n {\n \tstruct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device-\u003edev);\n \tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n-\tstruct mlxsw_sp_nve_params params = {\n-\t\t.type = MLXSW_SP_NVE_TYPE_VXLAN,\n-\t\t.vni = vxlan-\u003ecfg.vni,\n-\t\t.dev = vxlan_dev,\n-\t\t.ethertype = ethertype,\n-\t};\n+\tstruct mlxsw_sp_nve_params params;\n+\tconst struct vxlan_config *cfg;\n \tstruct mlxsw_sp_fid *fid;\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tparams.type = MLXSW_SP_NVE_TYPE_VXLAN;\n+\tparams.vni = cfg-\u003evni;\n+\tparams.dev = vxlan_dev;\n+\tparams.ethertype = ethertype;\n+\n \t/* If the VLAN is 0, we need to find the VLAN that is configured as\n \t * PVID and egress untagged on the bridge port of the VxLAN device.\n \t * It is possible no such VLAN exists\n@@ -2704,15 +2706,17 @@ mlxsw_sp_bridge_8021d_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,\n {\n \tstruct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device-\u003edev);\n \tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n-\tstruct mlxsw_sp_nve_params params = {\n-\t\t.type = MLXSW_SP_NVE_TYPE_VXLAN,\n-\t\t.vni = vxlan-\u003ecfg.vni,\n-\t\t.dev = vxlan_dev,\n-\t\t.ethertype = ETH_P_8021Q,\n-\t};\n+\tstruct mlxsw_sp_nve_params params;\n+\tconst struct vxlan_config *cfg;\n \tstruct mlxsw_sp_fid *fid;\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tparams.type = MLXSW_SP_NVE_TYPE_VXLAN;\n+\tparams.vni = cfg-\u003evni;\n+\tparams.dev = vxlan_dev;\n+\tparams.ethertype = ETH_P_8021Q;\n+\n \tfid = mlxsw_sp_fid_8021d_get(mlxsw_sp, bridge_device-\u003edev-\u003eifindex);\n \tif (IS_ERR(fid)) {\n \t\tNL_SET_ERR_MSG_MOD(extack, \"Failed to create 802.1D FID\");\n@@ -2933,10 +2937,13 @@ static void __mlxsw_sp_bridge_vxlan_leave(struct mlxsw_sp *mlxsw_sp,\n \t\t\t\t\t  const struct net_device *vxlan_dev)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n+\tconst struct vxlan_config *cfg;\n \tstruct mlxsw_sp_fid *fid;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n \t/* If the VxLAN device is down, then the FID does not have a VNI */\n-\tfid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vxlan-\u003ecfg.vni);\n+\tfid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, cfg-\u003evni);\n \tif (!fid)\n \t\treturn;\n \n@@ -3029,11 +3036,13 @@ static void mlxsw_sp_fdb_vxlan_call_notifiers(struct net_device *dev,\n \tstruct switchdev_notifier_vxlan_fdb_info info;\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tenum switchdev_notifier_type type;\n+\tconst struct vxlan_config *cfg;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \ttype = adding ? SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE :\n \t\t\tSWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE;\n \tmlxsw_sp_switchdev_addr_vxlan_convert(proto, addr, \u0026info.remote_ip);\n-\tinfo.remote_port = vxlan-\u003ecfg.dst_port;\n+\tinfo.remote_port = cfg-\u003edst_port;\n \tinfo.remote_vni = vni;\n \tinfo.remote_ifindex = 0;\n \tether_addr_copy(info.eth_addr, mac);\n@@ -3236,8 +3245,10 @@ __mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp *mlxsw_sp,\n \n \tif (adding \u0026\u0026 netif_is_vxlan(dev)) {\n \t\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\t\tconst struct vxlan_config *cfg;\n \n-\t\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_LEARN))\n+\t\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\t\tif (!(cfg-\u003eflags \u0026 VXLAN_F_LEARN))\n \t\t\treturn -EINVAL;\n \t}\n \n@@ -3722,9 +3733,11 @@ mlxsw_sp_switchdev_vxlan_work_prepare(struct mlxsw_sp_switchdev_event_work *\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(switchdev_work-\u003edev);\n \tstruct switchdev_notifier_vxlan_fdb_info *vxlan_fdb_info;\n-\tstruct vxlan_config *cfg = \u0026vxlan-\u003ecfg;\n+\tconst struct vxlan_config *cfg;\n \tstruct netlink_ext_ack *extack;\n \n+\tcfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n+\n \textack = switchdev_notifier_info_to_extack(info);\n \tvxlan_fdb_info = container_of(info,\n \t\t\t\t      struct switchdev_notifier_vxlan_fdb_info,\n@@ -3851,11 +3864,15 @@ mlxsw_sp_switchdev_vxlan_vlan_add(struct mlxsw_sp *mlxsw_sp,\n \t\t\t\t  struct netlink_ext_ack *extack)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n-\t__be32 vni = vxlan-\u003ecfg.vni;\n+\tconst struct vxlan_config *cfg;\n \tstruct mlxsw_sp_fid *fid;\n \tu16 old_vid;\n+\t__be32 vni;\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tvni = cfg-\u003evni;\n+\n \t/* We cannot have the same VLAN as PVID and egress untagged on multiple\n \t * VxLAN devices. Note that we get this notification before the VLAN is\n \t * actually added to the bridge's database, so it is not possible for\n@@ -3935,12 +3952,16 @@ mlxsw_sp_switchdev_vxlan_vlan_del(struct mlxsw_sp *mlxsw_sp,\n \t\t\t\t  const struct net_device *vxlan_dev, u16 vid)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n-\t__be32 vni = vxlan-\u003ecfg.vni;\n+\tconst struct vxlan_config *cfg;\n \tstruct mlxsw_sp_fid *fid;\n+\t__be32 vni;\n \n \tif (!netif_running(vxlan_dev))\n \t\treturn;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tvni = cfg-\u003evni;\n+\n \tfid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vni);\n \tif (!fid)\n \t\treturn;\ndiff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c\nindex 347245cc1de4e..0e5ab0055ff8c 100644\n--- a/drivers/net/vxlan/vxlan_core.c\n+++ b/drivers/net/vxlan/vxlan_core.c\n@@ -110,20 +110,23 @@ static struct vxlan_dev *vxlan_vs_find_vni(struct vxlan_sock *vs,\n \t\tvni = 0;\n \n \thlist_for_each_entry_rcu(node, vni_head(vs, vni), hlist) {\n+\t\tconst struct vxlan_config *cfg;\n+\n \t\tif (!node-\u003evxlan)\n \t\t\tcontinue;\n+\n+\t\tcfg = rcu_dereference(node-\u003evxlan-\u003ecfg);\n+\n \t\tvnode = NULL;\n-\t\tif (node-\u003evxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER) {\n+\t\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER) {\n \t\t\tvnode = vxlan_vnifilter_lookup(node-\u003evxlan, vni);\n \t\t\tif (!vnode)\n \t\t\t\tcontinue;\n-\t\t} else if (node-\u003evxlan-\u003edefault_dst.remote_vni != vni) {\n+\t\t} else if (cfg-\u003evni != vni) {\n \t\t\tcontinue;\n \t\t}\n \n \t\tif (IS_ENABLED(CONFIG_IPV6)) {\n-\t\t\tconst struct vxlan_config *cfg = \u0026node-\u003evxlan-\u003ecfg;\n-\n \t\t\tif ((cfg-\u003eflags \u0026 VXLAN_F_IPV6_LINKLOCAL) \u0026\u0026\n \t\t\t    cfg-\u003eremote_ifindex != ifindex)\n \t\t\t\tcontinue;\n@@ -157,6 +160,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,\n \t\t\t  u32 portid, u32 seq, int type, unsigned int flags,\n \t\t\t  const struct vxlan_rdst *rdst)\n {\n+\tconst struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n \tunsigned long now = jiffies;\n \tstruct nda_cacheinfo ci;\n \tbool send_ip, send_eth;\n@@ -216,10 +220,10 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,\n \t\t\tgoto nla_put_failure;\n \n \t\tif (rdst-\u003eremote_port \u0026\u0026\n-\t\t    rdst-\u003eremote_port != vxlan-\u003ecfg.dst_port \u0026\u0026\n+\t\t    rdst-\u003eremote_port != cfg-\u003edst_port \u0026\u0026\n \t\t    nla_put_be16(skb, NDA_PORT, rdst-\u003eremote_port))\n \t\t\tgoto nla_put_failure;\n-\t\tif (rdst-\u003eremote_vni != vxlan-\u003edefault_dst.remote_vni \u0026\u0026\n+\t\tif (rdst-\u003eremote_vni != cfg-\u003evni \u0026\u0026\n \t\t    nla_put_u32(skb, NDA_VNI, be32_to_cpu(rdst-\u003eremote_vni)))\n \t\t\tgoto nla_put_failure;\n \t\tif (rdst-\u003eremote_ifindex \u0026\u0026\n@@ -227,7 +231,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,\n \t\t\tgoto nla_put_failure;\n \t}\n \n-\tif ((vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026 fdb-\u003ekey.vni \u0026\u0026\n+\tif ((cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026 fdb-\u003ekey.vni \u0026\u0026\n \t    nla_put_u32(skb, NDA_SRC_VNI,\n \t\t\tbe32_to_cpu(fdb-\u003ekey.vni)))\n \t\tgoto nla_put_failure;\n@@ -377,14 +381,15 @@ static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])\n \n /* Look up Ethernet address in forwarding table */\n static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,\n+\t\t\t\t\t    const struct vxlan_config *cfg,\n \t\t\t\t\t    const u8 *mac, __be32 vni)\n {\n \tstruct vxlan_fdb_key key;\n \n \tmemset(\u0026key, 0, sizeof(key));\n \tmemcpy(key.eth_addr, mac, sizeof(key.eth_addr));\n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA))\n-\t\tkey.vni = vxlan-\u003edefault_dst.remote_vni;\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA))\n+\t\tkey.vni = cfg-\u003evni;\n \telse\n \t\tkey.vni = vni;\n \n@@ -393,11 +398,12 @@ static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,\n }\n \n static struct vxlan_fdb *vxlan_find_mac_tx(struct vxlan_dev *vxlan,\n+\t\t\t\t\t   const struct vxlan_config *cfg,\n \t\t\t\t\t   const u8 *mac, __be32 vni)\n {\n \tstruct vxlan_fdb *f;\n \n-\tf = vxlan_find_mac_rcu(vxlan, mac, vni);\n+\tf = vxlan_find_mac_rcu(vxlan, cfg, mac, vni);\n \tif (f) {\n \t\tunsigned long now = jiffies;\n \n@@ -416,7 +422,7 @@ static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan,\n \tlockdep_assert_held_once(\u0026vxlan-\u003ehash_lock);\n \n \trcu_read_lock();\n-\tf = vxlan_find_mac_rcu(vxlan, mac, vni);\n+\tf = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan-\u003ecfg), mac, vni);\n \trcu_read_unlock();\n \n \treturn f;\n@@ -457,7 +463,7 @@ int vxlan_fdb_find_uc(struct net_device *dev, const u8 *mac, __be32 vni,\n \n \trcu_read_lock();\n \n-\tf = vxlan_find_mac_rcu(vxlan, eth_addr, vni);\n+\tf = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan-\u003ecfg), eth_addr, vni);\n \tif (f)\n \t\trdst = first_remote_rcu(f);\n \tif (!rdst) {\n@@ -798,6 +804,7 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,\n \t\t\t       u32 nhid, struct netlink_ext_ack *extack)\n {\n \tstruct nexthop *old_nh = rtnl_dereference(fdb-\u003enh);\n+\tconst struct vxlan_config *cfg;\n \tstruct nexthop *nh;\n \tint err = -EINVAL;\n \n@@ -826,7 +833,8 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,\n \t}\n \n \t/* check nexthop group family */\n-\tswitch (vxlan-\u003edefault_dst.remote_ip.sa.sa_family) {\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tswitch (cfg-\u003eremote_ip.sa.sa_family) {\n \tcase AF_INET:\n \t\tif (!nexthop_has_v4(nh)) {\n \t\t\terr = -EAFNOSUPPORT;\n@@ -863,12 +871,13 @@ int vxlan_fdb_create(struct vxlan_dev *vxlan,\n \t\t     u32 nhid, struct vxlan_fdb **fdb,\n \t\t     struct netlink_ext_ack *extack)\n {\n+\tconst struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n \tstruct vxlan_rdst *rd = NULL;\n \tstruct vxlan_fdb *f;\n \tint rc;\n \n-\tif (vxlan-\u003ecfg.addrmax \u0026\u0026\n-\t    vxlan-\u003eaddrcnt \u003e= vxlan-\u003ecfg.addrmax)\n+\tif (cfg-\u003eaddrmax \u0026\u0026\n+\t    vxlan-\u003eaddrcnt \u003e= cfg-\u003eaddrmax)\n \t\treturn -ENOSPC;\n \n \tnetdev_dbg(vxlan-\u003edev, \"add %pM -\u003e %pIS\\n\", mac, ip);\n@@ -1154,6 +1163,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,\n \t\t\t   __be32 *vni, u32 *ifindex, u32 *nhid,\n \t\t\t   struct netlink_ext_ack *extack)\n {\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tstruct net *net = dev_net(vxlan-\u003edev);\n \tint err;\n \n@@ -1170,7 +1180,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,\n \t\t\treturn err;\n \t\t}\n \t} else {\n-\t\tunion vxlan_addr *remote = \u0026vxlan-\u003edefault_dst.remote_ip;\n+\t\tconst union vxlan_addr *remote = \u0026cfg-\u003eremote_ip;\n \n \t\tif (remote-\u003esa.sa_family == AF_INET) {\n \t\t\tip-\u003esin.sin_addr.s_addr = htonl(INADDR_ANY);\n@@ -1190,7 +1200,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,\n \t\t}\n \t\t*port = nla_get_be16(tb[NDA_PORT]);\n \t} else {\n-\t\t*port = vxlan-\u003ecfg.dst_port;\n+\t\t*port = cfg-\u003edst_port;\n \t}\n \n \tif (tb[NDA_VNI]) {\n@@ -1200,7 +1210,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,\n \t\t}\n \t\t*vni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));\n \t} else {\n-\t\t*vni = vxlan-\u003edefault_dst.remote_vni;\n+\t\t*vni = cfg-\u003evni;\n \t}\n \n \tif (tb[NDA_SRC_VNI]) {\n@@ -1210,7 +1220,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,\n \t\t}\n \t\t*src_vni = cpu_to_be32(nla_get_u32(tb[NDA_SRC_VNI]));\n \t} else {\n-\t\t*src_vni = vxlan-\u003edefault_dst.remote_vni;\n+\t\t*src_vni = cfg-\u003evni;\n \t}\n \n \tif (tb[NDA_IFINDEX]) {\n@@ -1241,6 +1251,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],\n \t\t\t const unsigned char *addr, u16 vid, u16 flags,\n \t\t\t bool *notified, struct netlink_ext_ack *extack)\n {\n+\tconst struct vxlan_config *cfg;\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \t/* struct net *net = dev_net(vxlan-\u003edev); */\n \tunion vxlan_addr ip;\n@@ -1268,7 +1279,8 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],\n \t\treturn -EINVAL;\n \t}\n \n-\tif (vxlan-\u003edefault_dst.remote_ip.sa.sa_family != ip.sa.sa_family)\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tif (cfg-\u003eremote_ip.sa.sa_family != ip.sa.sa_family)\n \t\treturn -EAFNOSUPPORT;\n \n \tspin_lock_bh(\u0026vxlan-\u003ehash_lock);\n@@ -1405,18 +1417,21 @@ static int vxlan_fdb_get(struct sk_buff *skb,\n \t\t\t struct netlink_ext_ack *extack)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tconst struct vxlan_config *cfg;\n \tstruct vxlan_fdb *f;\n \t__be32 vni;\n \tint err;\n \n+\tcfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n+\n \tif (tb[NDA_VNI])\n \t\tvni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));\n \telse\n-\t\tvni = vxlan-\u003edefault_dst.remote_vni;\n+\t\tvni = cfg-\u003evni;\n \n \trcu_read_lock();\n \n-\tf = vxlan_find_mac_rcu(vxlan, addr, vni);\n+\tf = vxlan_find_mac_rcu(vxlan, cfg, addr, vni);\n \tif (!f) {\n \t\tNL_SET_ERR_MSG(extack, \"Fdb entry not found\");\n \t\terr = -ENOENT;\n@@ -1434,6 +1449,7 @@ static int vxlan_fdb_get(struct sk_buff *skb,\n  * and Tunnel endpoint.\n  */\n static enum skb_drop_reason vxlan_snoop(struct net_device *dev,\n+\t\t\t\t\tconst struct vxlan_config *cfg,\n \t\t\t\t\tunion vxlan_addr *src_ip,\n \t\t\t\t\tconst u8 *src_mac, u32 src_ifindex,\n \t\t\t\t\t__be32 vni)\n@@ -1452,7 +1468,7 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,\n \t\tifindex = src_ifindex;\n #endif\n \n-\tf = vxlan_find_mac_rcu(vxlan, src_mac, vni);\n+\tf = vxlan_find_mac_rcu(vxlan, cfg, src_mac, vni);\n \tif (likely(f)) {\n \t\tstruct vxlan_rdst *rdst = first_remote_rcu(f);\n \t\tunsigned long now = jiffies;\n@@ -1488,9 +1504,9 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,\n \t\t\tvxlan_fdb_update(vxlan, src_mac, src_ip,\n \t\t\t\t\t NUD_REACHABLE,\n \t\t\t\t\t NLM_F_EXCL|NLM_F_CREATE,\n-\t\t\t\t\t vxlan-\u003ecfg.dst_port,\n+\t\t\t\t\t cfg-\u003edst_port,\n \t\t\t\t\t vni,\n-\t\t\t\t\t vxlan-\u003edefault_dst.remote_vni,\n+\t\t\t\t\t cfg-\u003evni,\n \t\t\t\t\t ifindex, NTF_SELF, 0, true, NULL);\n \t\tspin_unlock(\u0026vxlan-\u003ehash_lock);\n \t}\n@@ -1518,6 +1534,7 @@ static bool __vxlan_sock_release_prep(struct vxlan_sock *vs)\n \n static void vxlan_sock_release(struct vxlan_dev *vxlan)\n {\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tstruct vxlan_sock *sock4 = rtnl_dereference(vxlan-\u003evn4_sock);\n #if IS_ENABLED(CONFIG_IPV6)\n \tstruct vxlan_sock *sock6 = rtnl_dereference(vxlan-\u003evn6_sock);\n@@ -1527,7 +1544,7 @@ static void vxlan_sock_release(struct vxlan_dev *vxlan)\n \n \tRCU_INIT_POINTER(vxlan-\u003evn4_sock, NULL);\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)\n \t\tvxlan_vs_del_vnigrp(vxlan);\n \telse\n \t\tvxlan_vs_del_dev(vxlan);\n@@ -1598,6 +1615,7 @@ static void vxlan_parse_gbp_hdr(struct sk_buff *skb, u32 vxflags,\n }\n \n static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,\n+\t\t\t\t\t  const struct vxlan_config *cfg,\n \t\t\t\t\t  struct vxlan_sock *vs,\n \t\t\t\t\t  struct sk_buff *skb, __be32 vni)\n {\n@@ -1623,10 +1641,10 @@ static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,\n #endif\n \t}\n \n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_LEARN))\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_LEARN))\n \t\treturn SKB_NOT_DROPPED_YET;\n \n-\treturn vxlan_snoop(skb-\u003edev, \u0026saddr, eth_hdr(skb)-\u003eh_source,\n+\treturn vxlan_snoop(skb-\u003edev, cfg, \u0026saddr, eth_hdr(skb)-\u003eh_source,\n \t\t\t   ifindex, vni);\n }\n \n@@ -1657,18 +1675,21 @@ static bool vxlan_ecn_decapsulate(struct vxlan_sock *vs, void *oiph,\n static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n {\n \tstruct vxlan_vni_node *vninode = NULL;\n-\tconst struct vxlanhdr *vh;\n-\tstruct vxlan_dev *vxlan;\n-\tstruct vxlan_sock *vs;\n-\tstruct vxlan_metadata _md;\n-\tstruct vxlan_metadata *md = \u0026_md;\n \t__be16 protocol = htons(ETH_P_TEB);\n+\tconst struct vxlan_config *cfg;\n \tenum skb_drop_reason reason;\n+\tconst struct vxlanhdr *vh;\n+\tstruct vxlan_metadata *md;\n+\tstruct vxlan_metadata _md;\n+\tstruct vxlan_dev *vxlan;\n \tbool raw_proto = false;\n-\tvoid *oiph;\n+\tstruct vxlan_sock *vs;\n \t__be32 vni = 0;\n+\tvoid *oiph;\n \tint nh;\n \n+\tmd = \u0026_md;\n+\n \t/* Need UDP and VXLAN header to be present */\n \treason = pskb_may_pull_reason(skb, VXLAN_HLEN);\n \tif (reason)\n@@ -1696,8 +1717,10 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \t\tgoto drop;\n \t}\n \n-\tif (vh-\u003evx_flags \u0026 vxlan-\u003ecfg.reserved_bits.vx_flags ||\n-\t    vh-\u003evx_vni \u0026 vxlan-\u003ecfg.reserved_bits.vx_vni) {\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\n+\tif (vh-\u003evx_flags \u0026 cfg-\u003ereserved_bits.vx_flags ||\n+\t    vh-\u003evx_vni \u0026 cfg-\u003ereserved_bits.vx_vni) {\n \t\t/* If the header uses bits besides those enabled by the\n \t\t * netdevice configuration, treat this as a malformed packet.\n \t\t * This behavior diverges from VXLAN RFC (RFC7348) which\n@@ -1709,12 +1732,12 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \t\treason = SKB_DROP_REASON_VXLAN_INVALID_HDR;\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_frame_errors);\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_errors);\n-\t\tvxlan_vnifilter_count(vxlan, vni, vninode,\n+\t\tvxlan_vnifilter_count(vxlan, cfg, vni, vninode,\n \t\t\t\t      VXLAN_VNI_STATS_RX_ERRORS, 0);\n \t\tgoto drop;\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_GPE) {\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_GPE) {\n \t\tif (!vxlan_parse_gpe_proto(vh, \u0026protocol))\n \t\t\tgoto drop;\n \t\traw_proto = true;\n@@ -1726,8 +1749,8 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \t\tgoto drop;\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_REMCSUM_RX) {\n-\t\treason = vxlan_remcsum(skb, vxlan-\u003ecfg.flags);\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_REMCSUM_RX) {\n+\t\treason = vxlan_remcsum(skb, cfg-\u003eflags);\n \t\tif (unlikely(reason))\n \t\t\tgoto drop;\n \t}\n@@ -1752,14 +1775,14 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \t\tmemset(md, 0, sizeof(*md));\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_GBP)\n-\t\tvxlan_parse_gbp_hdr(skb, vxlan-\u003ecfg.flags, md);\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_GBP)\n+\t\tvxlan_parse_gbp_hdr(skb, cfg-\u003eflags, md);\n \t/* Note that GBP and GPE can never be active together. This is\n \t * ensured in vxlan_dev_configure.\n \t */\n \n \tif (!raw_proto) {\n-\t\treason = vxlan_set_mac(vxlan, vs, skb, vni);\n+\t\treason = vxlan_set_mac(vxlan, cfg, vs, skb, vni);\n \t\tif (reason)\n \t\t\tgoto drop;\n \t} else {\n@@ -1780,7 +1803,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \tif (reason) {\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_length_errors);\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_errors);\n-\t\tvxlan_vnifilter_count(vxlan, vni, vninode,\n+\t\tvxlan_vnifilter_count(vxlan, cfg, vni, vninode,\n \t\t\t\t      VXLAN_VNI_STATS_RX_ERRORS, 0);\n \t\tgoto drop;\n \t}\n@@ -1792,7 +1815,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \t\treason = SKB_DROP_REASON_IP_TUNNEL_ECN;\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_frame_errors);\n \t\tDEV_STATS_INC(vxlan-\u003edev, rx_errors);\n-\t\tvxlan_vnifilter_count(vxlan, vni, vninode,\n+\t\tvxlan_vnifilter_count(vxlan, cfg, vni, vninode,\n \t\t\t\t      VXLAN_VNI_STATS_RX_ERRORS, 0);\n \t\tgoto drop;\n \t}\n@@ -1802,14 +1825,15 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n \tif (unlikely(!(vxlan-\u003edev-\u003eflags \u0026 IFF_UP))) {\n \t\trcu_read_unlock();\n \t\tdev_dstats_rx_dropped(vxlan-\u003edev);\n-\t\tvxlan_vnifilter_count(vxlan, vni, vninode,\n+\t\tvxlan_vnifilter_count(vxlan, cfg, vni, vninode,\n \t\t\t\t      VXLAN_VNI_STATS_RX_DROPS, 0);\n \t\treason = SKB_DROP_REASON_DEV_READY;\n \t\tgoto drop;\n \t}\n \n \tdev_dstats_rx_add(vxlan-\u003edev, skb-\u003elen);\n-\tvxlan_vnifilter_count(vxlan, vni, vninode, VXLAN_VNI_STATS_RX, skb-\u003elen);\n+\tvxlan_vnifilter_count(vxlan, cfg, vni, vninode, VXLAN_VNI_STATS_RX,\n+\t\t\t      skb-\u003elen);\n \tgro_cells_receive(\u0026vxlan-\u003egro_cells, skb);\n \n \trcu_read_unlock();\n@@ -1850,7 +1874,8 @@ static int vxlan_err_lookup(struct sock *sk, struct sk_buff *skb)\n \treturn 0;\n }\n \n-static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n+static int arp_reduce(struct net_device *dev, struct sk_buff *skb,\n+\t\t      const struct vxlan_config *cfg, __be32 vni)\n {\n \tstruct neigh_table *tbl = arp_table(dev_net(dev));\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n@@ -1864,7 +1889,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n \n \tif (!pskb_network_may_pull(skb, arp_hdr_len(dev))) {\n \t\tdev_dstats_tx_dropped(dev);\n-\t\tvxlan_vnifilter_count(vxlan, vni, NULL,\n+\t\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL,\n \t\t\t\t      VXLAN_VNI_STATS_TX_DROPS, 0);\n \t\tgoto out;\n \t}\n@@ -1905,7 +1930,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n \t\tneigh_ha_snapshot(ha, n, n-\u003edev);\n \n \t\trcu_read_lock();\n-\t\tf = vxlan_find_mac_tx(vxlan, ha, vni);\n+\t\tf = vxlan_find_mac_tx(vxlan, cfg, ha, vni);\n \t\tif (f)\n \t\t\trdst = first_remote_rcu(f);\n \t\tif (rdst \u0026\u0026 vxlan_addr_any(\u0026rdst-\u003eremote_ip)) {\n@@ -1931,11 +1956,11 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n \n \t\tif (netif_rx(reply) == NET_RX_DROP) {\n \t\t\tdev_dstats_rx_dropped(dev);\n-\t\t\tvxlan_vnifilter_count(vxlan, vni, NULL,\n+\t\t\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL,\n \t\t\t\t\t      VXLAN_VNI_STATS_RX_DROPS, 0);\n \t\t}\n \n-\t} else if (vxlan-\u003ecfg.flags \u0026 VXLAN_F_L3MISS) {\n+\t} else if (cfg-\u003eflags \u0026 VXLAN_F_L3MISS) {\n \t\tunion vxlan_addr ipa = {\n \t\t\t.sin.sin_addr.s_addr = tip,\n \t\t\t.sin.sin_family = AF_INET,\n@@ -2043,7 +2068,8 @@ static struct sk_buff *vxlan_na_create(struct sk_buff *request,\n \treturn reply;\n }\n \n-static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n+static int neigh_reduce(struct net_device *dev, struct sk_buff *skb,\n+\t\t\tconst struct vxlan_config *cfg, __be32 vni)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tconst struct in6_addr *daddr;\n@@ -2077,7 +2103,7 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n \t\t}\n \n \t\tneigh_ha_snapshot(ha, n, n-\u003edev);\n-\t\tf = vxlan_find_mac_tx(vxlan, ha, vni);\n+\t\tf = vxlan_find_mac_tx(vxlan, cfg, ha, vni);\n \t\tif (f)\n \t\t\trdst = first_remote_rcu(f);\n \t\tif (rdst \u0026\u0026 vxlan_addr_any(\u0026rdst-\u003eremote_ip)) {\n@@ -2096,10 +2122,10 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n \n \t\tif (netif_rx(reply) == NET_RX_DROP) {\n \t\t\tdev_dstats_rx_dropped(dev);\n-\t\t\tvxlan_vnifilter_count(vxlan, vni, NULL,\n+\t\t\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL,\n \t\t\t\t\t      VXLAN_VNI_STATS_RX_DROPS, 0);\n \t\t}\n-\t} else if (vxlan-\u003ecfg.flags \u0026 VXLAN_F_L3MISS) {\n+\t} else if (cfg-\u003eflags \u0026 VXLAN_F_L3MISS) {\n \t\tunion vxlan_addr ipa = {\n \t\t\t.sin6.sin6_addr = msg-\u003etarget,\n \t\t\t.sin6.sin6_family = AF_INET6,\n@@ -2115,9 +2141,9 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)\n }\n #endif\n \n-static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)\n+static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb,\n+\t\t\t       const struct vxlan_config *cfg)\n {\n-\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct neigh_table *tbl;\n \tstruct neighbour *n;\n \n@@ -2136,7 +2162,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)\n \t\ttbl = arp_table(dev_net(dev));\n \t\tpip = ip_hdr(skb);\n \t\tn = neigh_lookup(tbl, \u0026pip-\u003edaddr, dev);\n-\t\tif (!n \u0026\u0026 (vxlan-\u003ecfg.flags \u0026 VXLAN_F_L3MISS)) {\n+\t\tif (!n \u0026\u0026 (cfg-\u003eflags \u0026 VXLAN_F_L3MISS)) {\n \t\t\tunion vxlan_addr ipa = {\n \t\t\t\t.sin.sin_addr.s_addr = pip-\u003edaddr,\n \t\t\t\t.sin.sin_family = AF_INET,\n@@ -2164,7 +2190,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)\n \t\ttbl = nd_table(dev_net(dev));\n \t\tpip6 = ipv6_hdr(skb);\n \t\tn = neigh_lookup(tbl, \u0026pip6-\u003edaddr, dev);\n-\t\tif (!n \u0026\u0026 (vxlan-\u003ecfg.flags \u0026 VXLAN_F_L3MISS)) {\n+\t\tif (!n \u0026\u0026 (cfg-\u003eflags \u0026 VXLAN_F_L3MISS)) {\n \t\t\tunion vxlan_addr ipa = {\n \t\t\t\t.sin6.sin6_addr = pip6-\u003edaddr,\n \t\t\t\t.sin6.sin6_family = AF_INET6,\n@@ -2282,20 +2308,28 @@ static int vxlan_build_skb(struct sk_buff *skb, struct dst_entry *dst,\n \n /* Bypass encapsulation if the destination is local */\n static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,\n-\t\t\t       struct vxlan_dev *dst_vxlan, __be32 vni,\n-\t\t\t       bool snoop)\n+\t\t\t       struct vxlan_dev *dst_vxlan,\n+\t\t\t       const struct vxlan_config *src_cfg,\n+\t\t\t       __be32 vni, bool snoop)\n {\n+\tconst struct vxlan_config *dst_cfg;\n \tunion vxlan_addr loopback;\n-\tunion vxlan_addr *remote_ip = \u0026dst_vxlan-\u003edefault_dst.remote_ip;\n \tunsigned int len = skb-\u003elen;\n-\tstruct net_device *dev;\n+\tstruct net_device *dev = dst_vxlan-\u003edev;\n \n \tskb-\u003epkt_type = PACKET_HOST;\n \tskb-\u003eencapsulation = 0;\n-\tskb-\u003edev = dst_vxlan-\u003edev;\n+\tskb-\u003edev = dev;\n \t__skb_pull(skb, skb_network_offset(skb));\n \n-\tif (remote_ip-\u003esa.sa_family == AF_INET) {\n+\trcu_read_lock();\n+\tdst_cfg = rcu_dereference(dst_vxlan-\u003ecfg);\n+\tif (unlikely(!(dev-\u003eflags \u0026 IFF_UP))) {\n+\t\tkfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);\n+\t\tgoto drop;\n+\t}\n+\n+\tif (dst_cfg-\u003eremote_ip.sa.sa_family == AF_INET) {\n \t\tloopback.sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);\n \t\tloopback.sa.sa_family =  AF_INET;\n #if IS_ENABLED(CONFIG_IPV6)\n@@ -2305,27 +2339,20 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,\n #endif\n \t}\n \n-\trcu_read_lock();\n-\tdev = skb-\u003edev;\n-\tif (unlikely(!(dev-\u003eflags \u0026 IFF_UP))) {\n-\t\tkfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);\n-\t\tgoto drop;\n-\t}\n-\n-\tif ((dst_vxlan-\u003ecfg.flags \u0026 VXLAN_F_LEARN) \u0026\u0026 snoop)\n-\t\tvxlan_snoop(dev, \u0026loopback, eth_hdr(skb)-\u003eh_source, 0, vni);\n+\tif ((dst_cfg-\u003eflags \u0026 VXLAN_F_LEARN) \u0026\u0026 snoop)\n+\t\tvxlan_snoop(dev, dst_cfg, \u0026loopback, eth_hdr(skb)-\u003eh_source, 0, vni);\n \n \tdev_dstats_tx_add(src_vxlan-\u003edev, len);\n-\tvxlan_vnifilter_count(src_vxlan, vni, NULL, VXLAN_VNI_STATS_TX, len);\n+\tvxlan_vnifilter_count(src_vxlan, src_cfg, vni, NULL, VXLAN_VNI_STATS_TX, len);\n \n \tif (__netif_rx(skb) == NET_RX_SUCCESS) {\n \t\tdev_dstats_rx_add(dst_vxlan-\u003edev, len);\n-\t\tvxlan_vnifilter_count(dst_vxlan, vni, NULL, VXLAN_VNI_STATS_RX,\n+\t\tvxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL, VXLAN_VNI_STATS_RX,\n \t\t\t\t      len);\n \t} else {\n drop:\n \t\tdev_dstats_rx_dropped(dev);\n-\t\tvxlan_vnifilter_count(dst_vxlan, vni, NULL,\n+\t\tvxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL,\n \t\t\t\t      VXLAN_VNI_STATS_RX_DROPS, 0);\n \t}\n \trcu_read_unlock();\n@@ -2333,6 +2360,7 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,\n \n static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,\n \t\t\t\t struct vxlan_dev *vxlan,\n+\t\t\t\t const struct vxlan_config *cfg,\n \t\t\t\t int addr_family,\n \t\t\t\t __be16 dst_port, int dst_ifindex, __be32 vni,\n \t\t\t\t struct dst_entry *dst,\n@@ -2348,22 +2376,22 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,\n \t/* Bypass encapsulation if the destination is local */\n \tif (rt_flags \u0026 RTCF_LOCAL \u0026\u0026\n \t    !(rt_flags \u0026 (RTCF_BROADCAST | RTCF_MULTICAST)) \u0026\u0026\n-\t    vxlan-\u003ecfg.flags \u0026 VXLAN_F_LOCALBYPASS) {\n+\t    cfg-\u003eflags \u0026 VXLAN_F_LOCALBYPASS) {\n \t\tstruct vxlan_dev *dst_vxlan;\n \n \t\tdst_release(dst);\n \t\tdst_vxlan = vxlan_find_vni(vxlan-\u003enet, dst_ifindex, vni,\n \t\t\t\t\t   addr_family, dst_port,\n-\t\t\t\t\t   vxlan-\u003ecfg.flags);\n+\t\t\t\t\t   cfg-\u003eflags);\n \t\tif (!dst_vxlan) {\n \t\t\tDEV_STATS_INC(dev, tx_errors);\n-\t\t\tvxlan_vnifilter_count(vxlan, vni, NULL,\n+\t\t\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL,\n \t\t\t\t\t      VXLAN_VNI_STATS_TX_ERRORS, 0);\n \t\t\tkfree_skb_reason(skb, SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND);\n \n \t\t\treturn -ENOENT;\n \t\t}\n-\t\tvxlan_encap_bypass(skb, vxlan, dst_vxlan, vni, true);\n+\t\tvxlan_encap_bypass(skb, vxlan, dst_vxlan, cfg, vni, true);\n \t\treturn 1;\n \t}\n \n@@ -2371,30 +2399,35 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,\n }\n \n void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n+\t\t    const struct vxlan_config *cfg,\n \t\t    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc)\n {\n+\tunsigned int pkt_len = skb-\u003elen;\n+\tstruct vxlan_metadata _md = {};\n+\t__be16 src_port = 0, dst_port;\n+\tstruct dst_entry *ndst = NULL;\n+\tenum skb_drop_reason reason;\n \tstruct dst_cache *dst_cache;\n+\tconst struct iphdr *old_iph;\n \tstruct ip_tunnel_info *info;\n \tstruct ip_tunnel_key *pkey;\n+\tstruct vxlan_metadata *md;\n \tstruct ip_tunnel_key key;\n-\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n-\tconst struct iphdr *old_iph;\n-\tstruct vxlan_metadata _md = {};\n-\tstruct vxlan_metadata *md = \u0026_md;\n-\tunsigned int pkt_len = skb-\u003elen;\n-\t__be16 src_port = 0, dst_port;\n-\tstruct dst_entry *ndst = NULL;\n+\tstruct vxlan_dev *vxlan;\n+\tu32 flags = cfg-\u003eflags;\n+\tbool udp_sum = false;\n+\tbool no_eth_encap;\n \tint addr_family;\n+\tbool use_cache;\n+\t__be32 vni = 0;\n \t__u8 tos, ttl;\n \tint ifindex;\n \tint err = 0;\n-\tu32 flags = vxlan-\u003ecfg.flags;\n-\tbool use_cache;\n-\tbool udp_sum = false;\n-\tbool xnet = !net_eq(vxlan-\u003enet, dev_net(vxlan-\u003edev));\n-\tenum skb_drop_reason reason;\n-\tbool no_eth_encap;\n-\t__be32 vni = 0;\n+\tbool xnet;\n+\n+\tvxlan = netdev_priv(dev);\n+\txnet = !net_eq(vxlan-\u003enet, dev_net(vxlan-\u003edev));\n+\tmd = \u0026_md;\n \n \tno_eth_encap = flags \u0026 VXLAN_F_GPE \u0026\u0026 skb-\u003eprotocol != htons(ETH_P_TEB);\n \treason = skb_vlan_inet_prepare(skb, no_eth_encap);\n@@ -2414,23 +2447,23 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\tif (vxlan_addr_any(\u0026rdst-\u003eremote_ip)) {\n \t\t\tif (did_rsc) {\n \t\t\t\t/* short-circuited back to local bridge */\n-\t\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan,\n+\t\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, cfg,\n \t\t\t\t\t\t   default_vni, true);\n \t\t\t\treturn;\n \t\t\t}\n \t\t\tgoto drop;\n \t\t}\n \n-\t\taddr_family = vxlan-\u003ecfg.saddr.sa.sa_family;\n-\t\tdst_port = rdst-\u003eremote_port ? rdst-\u003eremote_port : vxlan-\u003ecfg.dst_port;\n+\t\taddr_family = cfg-\u003esaddr.sa.sa_family;\n+\t\tdst_port = rdst-\u003eremote_port ? rdst-\u003eremote_port : cfg-\u003edst_port;\n \t\tvni = (rdst-\u003eremote_vni) ? : default_vni;\n \t\tifindex = rdst-\u003eremote_ifindex;\n \n \t\tif (addr_family == AF_INET) {\n-\t\t\tkey.u.ipv4.src = vxlan-\u003ecfg.saddr.sin.sin_addr.s_addr;\n+\t\t\tkey.u.ipv4.src = cfg-\u003esaddr.sin.sin_addr.s_addr;\n \t\t\tkey.u.ipv4.dst = rdst-\u003eremote_ip.sin.sin_addr.s_addr;\n \t\t} else {\n-\t\t\tkey.u.ipv6.src = vxlan-\u003ecfg.saddr.sin6.sin6_addr;\n+\t\t\tkey.u.ipv6.src = cfg-\u003esaddr.sin6.sin6_addr;\n \t\t\tkey.u.ipv6.dst = rdst-\u003eremote_ip.sin6.sin6_addr;\n \t\t}\n \n@@ -2439,11 +2472,11 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\tif (flags \u0026 VXLAN_F_TTL_INHERIT) {\n \t\t\tttl = ip_tunnel_get_ttl(old_iph, skb);\n \t\t} else {\n-\t\t\tttl = vxlan-\u003ecfg.ttl;\n+\t\t\tttl = cfg-\u003ettl;\n \t\t\tif (!ttl \u0026\u0026 vxlan_addr_multicast(\u0026rdst-\u003eremote_ip))\n \t\t\t\tttl = 1;\n \t\t}\n-\t\ttos = vxlan-\u003ecfg.tos;\n+\t\ttos = cfg-\u003etos;\n \t\tif (tos == 1)\n \t\t\ttos = ip_tunnel_get_dsfield(old_iph, skb);\n \t\tif (tos \u0026\u0026 !info)\n@@ -2454,9 +2487,9 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\telse\n \t\t\tudp_sum = !(flags \u0026 VXLAN_F_UDP_ZERO_CSUM6_TX);\n #if IS_ENABLED(CONFIG_IPV6)\n-\t\tswitch (vxlan-\u003ecfg.label_policy) {\n+\t\tswitch (cfg-\u003elabel_policy) {\n \t\tcase VXLAN_LABEL_FIXED:\n-\t\t\tkey.label = vxlan-\u003ecfg.label;\n+\t\t\tkey.label = cfg-\u003elabel;\n \t\t\tbreak;\n \t\tcase VXLAN_LABEL_INHERIT:\n \t\t\tkey.label = ip_tunnel_get_flowlabel(old_iph, skb);\n@@ -2474,7 +2507,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\t}\n \t\tpkey = \u0026info-\u003ekey;\n \t\taddr_family = ip_tunnel_info_af(info);\n-\t\tdst_port = info-\u003ekey.tp_dst ? : vxlan-\u003ecfg.dst_port;\n+\t\tdst_port = info-\u003ekey.tp_dst ? : cfg-\u003edst_port;\n \t\tvni = tunnel_id_to_key32(info-\u003ekey.tun_id);\n \t\tifindex = 0;\n \t\tdst_cache = \u0026info-\u003edst_cache;\n@@ -2487,8 +2520,8 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\ttos = info-\u003ekey.tos;\n \t\tudp_sum = test_bit(IP_TUNNEL_CSUM_BIT, info-\u003ekey.tun_flags);\n \t}\n-\tsrc_port = udp_flow_src_port(dev_net(dev), skb, vxlan-\u003ecfg.port_min,\n-\t\t\t\t     vxlan-\u003ecfg.port_max, true);\n+\tsrc_port = udp_flow_src_port(dev_net(dev), skb, cfg-\u003eport_min,\n+\t\t\t\t     cfg-\u003eport_max, true);\n \n \trcu_read_lock();\n \tif (addr_family == AF_INET) {\n@@ -2521,15 +2554,15 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \n \t\tif (!info) {\n \t\t\t/* Bypass encapsulation if the destination is local */\n-\t\t\terr = encap_bypass_if_local(skb, dev, vxlan, AF_INET,\n+\t\t\terr = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET,\n \t\t\t\t\t\t    dst_port, ifindex, vni,\n \t\t\t\t\t\t    \u0026rt-\u003edst, rt-\u003ert_flags);\n \t\t\tif (err)\n \t\t\t\tgoto out_unlock;\n \n-\t\t\tif (vxlan-\u003ecfg.df == VXLAN_DF_SET) {\n+\t\t\tif (cfg-\u003edf == VXLAN_DF_SET) {\n \t\t\t\tdf = htons(IP_DF);\n-\t\t\t} else if (vxlan-\u003ecfg.df == VXLAN_DF_INHERIT) {\n+\t\t\t} else if (cfg-\u003edf == VXLAN_DF_INHERIT) {\n \t\t\t\tstruct ethhdr *eth = eth_hdr(skb);\n \n \t\t\t\tif (ntohs(eth-\u003eh_proto) == ETH_P_IPV6 ||\n@@ -2558,7 +2591,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\t\t\tunclone-\u003ekey.u.ipv4.src = pkey-\u003eu.ipv4.dst;\n \t\t\t\tunclone-\u003ekey.u.ipv4.dst = saddr;\n \t\t\t}\n-\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, vni, false);\n+\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);\n \t\t\tdst_release(ndst);\n \t\t\tgoto out_unlock;\n \t\t}\n@@ -2608,7 +2641,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\tif (!info) {\n \t\t\tu32 rt6i_flags = dst_rt6_info(ndst)-\u003ert6i_flags;\n \n-\t\t\terr = encap_bypass_if_local(skb, dev, vxlan, AF_INET6,\n+\t\t\terr = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET6,\n \t\t\t\t\t\t    dst_port, ifindex, vni,\n \t\t\t\t\t\t    ndst, rt6i_flags);\n \t\t\tif (err)\n@@ -2632,7 +2665,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\t\t\tunclone-\u003ekey.u.ipv6.dst = saddr;\n \t\t\t}\n \n-\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, vni, false);\n+\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);\n \t\t\tdst_release(ndst);\n \t\t\tgoto out_unlock;\n \t\t}\n@@ -2653,14 +2686,14 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\t\t\t     ip6cb_flags);\n #endif\n \t}\n-\tvxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);\n+\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);\n out_unlock:\n \trcu_read_unlock();\n \treturn;\n \n drop:\n \tdev_dstats_tx_dropped(dev);\n-\tvxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);\n+\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);\n \tkfree_skb_reason(skb, reason);\n \treturn;\n \n@@ -2672,11 +2705,12 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n \t\tDEV_STATS_INC(dev, tx_carrier_errors);\n \tdst_release(ndst);\n \tDEV_STATS_INC(dev, tx_errors);\n-\tvxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);\n+\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);\n \tkfree_skb_reason(skb, reason);\n }\n \n static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,\n+\t\t\t  const struct vxlan_config *cfg,\n \t\t\t  struct vxlan_fdb *f, __be32 vni, bool did_rsc)\n {\n \tstruct vxlan_rdst nh_rdst;\n@@ -2693,7 +2727,7 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,\n \tdo_xmit = vxlan_fdb_nh_path_select(nh, hash, \u0026nh_rdst);\n \n \tif (likely(do_xmit))\n-\t\tvxlan_xmit_one(skb, dev, vni, \u0026nh_rdst, did_rsc);\n+\t\tvxlan_xmit_one(skb, dev, cfg, vni, \u0026nh_rdst, did_rsc);\n \telse\n \t\tgoto drop;\n \n@@ -2701,15 +2735,15 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,\n \n drop:\n \tdev_dstats_tx_dropped(dev);\n-\tvxlan_vnifilter_count(netdev_priv(dev), vni, NULL,\n+\tvxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,\n \t\t\t      VXLAN_VNI_STATS_TX_DROPS, 0);\n \tdev_kfree_skb(skb);\n }\n \n static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,\n-\t\t\t\t   u32 nhid, __be32 vni)\n+\t\t\t\t   u32 nhid, __be32 vni,\n+\t\t\t\t   const struct vxlan_config *cfg)\n {\n-\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct vxlan_rdst nh_rdst;\n \tstruct nexthop *nh;\n \tbool do_xmit;\n@@ -2727,11 +2761,11 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,\n \tdo_xmit = vxlan_fdb_nh_path_select(nh, hash, \u0026nh_rdst);\n \trcu_read_unlock();\n \n-\tif (vxlan-\u003ecfg.saddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)\n+\tif (cfg-\u003esaddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)\n \t\tgoto drop;\n \n \tif (likely(do_xmit))\n-\t\tvxlan_xmit_one(skb, dev, vni, \u0026nh_rdst, false);\n+\t\tvxlan_xmit_one(skb, dev, cfg, vni, \u0026nh_rdst, false);\n \telse\n \t\tgoto drop;\n \n@@ -2739,7 +2773,7 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,\n \n drop:\n \tdev_dstats_tx_dropped(dev);\n-\tvxlan_vnifilter_count(netdev_priv(dev), vni, NULL,\n+\tvxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,\n \t\t\t      VXLAN_VNI_STATS_TX_DROPS, 0);\n \tdev_kfree_skb(skb);\n \treturn NETDEV_TX_OK;\n@@ -2756,34 +2790,47 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct vxlan_rdst *rdst, *fdst = NULL;\n \tconst struct ip_tunnel_info *info;\n+\tconst struct vxlan_config *cfg;\n+\t__be32 default_vni;\n \tstruct vxlan_fdb *f;\n \tstruct ethhdr *eth;\n \t__be32 vni = 0;\n-\tu32 nhid = 0;\n \tbool did_rsc;\n+\tu32 nhid = 0;\n+\tu32 flags;\n+\n+\trcu_read_lock();\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\tflags = cfg-\u003eflags;\n+\tdefault_vni = cfg-\u003evni;\n \n \tinfo = skb_tunnel_info(skb);\n \n \tskb_reset_mac_header(skb);\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA) {\n+\tif (flags \u0026 VXLAN_F_COLLECT_METADATA) {\n \t\tif (info \u0026\u0026 info-\u003emode \u0026 IP_TUNNEL_INFO_BRIDGE \u0026\u0026\n \t\t    info-\u003emode \u0026 IP_TUNNEL_INFO_TX) {\n \t\t\tvni = tunnel_id_to_key32(info-\u003ekey.tun_id);\n \t\t\tnhid = info-\u003ekey.nhid;\n \t\t} else {\n \t\t\tif (info \u0026\u0026 info-\u003emode \u0026 IP_TUNNEL_INFO_TX)\n-\t\t\t\tvxlan_xmit_one(skb, dev, vni, NULL, false);\n+\t\t\t\tvxlan_xmit_one(skb, dev, cfg, vni, NULL, false);\n \t\t\telse\n \t\t\t\tkfree_skb_reason(skb, SKB_DROP_REASON_TUNNEL_TXINFO);\n+\t\t\trcu_read_unlock();\n \t\t\treturn NETDEV_TX_OK;\n \t\t}\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_PROXY) {\n+\tif (flags \u0026 VXLAN_F_PROXY) {\n \t\teth = eth_hdr(skb);\n-\t\tif (ntohs(eth-\u003eh_proto) == ETH_P_ARP)\n-\t\t\treturn arp_reduce(dev, skb, vni);\n+\t\tif (ntohs(eth-\u003eh_proto) == ETH_P_ARP) {\n+\t\t\tnetdev_tx_t res = arp_reduce(dev, skb, cfg, vni);\n+\n+\t\t\trcu_read_unlock();\n+\t\t\treturn res;\n+\t\t}\n #if IS_ENABLED(CONFIG_IPV6)\n \t\telse if (ntohs(eth-\u003eh_proto) == ETH_P_IPV6 \u0026\u0026\n \t\t\t pskb_network_may_pull(skb, sizeof(struct ipv6hdr) +\n@@ -2792,53 +2839,57 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)\n \t\t\tstruct nd_msg *m = (struct nd_msg *)(ipv6_hdr(skb) + 1);\n \n \t\t\tif (m-\u003eicmph.icmp6_code == 0 \u0026\u0026\n-\t\t\t    m-\u003eicmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION)\n-\t\t\t\treturn neigh_reduce(dev, skb, vni);\n+\t\t\t    m-\u003eicmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION) {\n+\t\t\t\tnetdev_tx_t res = neigh_reduce(dev, skb, cfg, vni);\n+\n+\t\t\t\trcu_read_unlock();\n+\t\t\t\treturn res;\n+\t\t\t}\n \t\t}\n #endif\n \t}\n \n-\tif (nhid)\n-\t\treturn vxlan_xmit_nhid(skb, dev, nhid, vni);\n+\tif (nhid) {\n+\t\tnetdev_tx_t res = vxlan_xmit_nhid(skb, dev, nhid, vni, cfg);\n+\n+\t\trcu_read_unlock();\n+\t\treturn res;\n+\t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_MDB) {\n+\tif (test_bit(VXLAN_DEV_F_MDB, \u0026vxlan-\u003eflags)) {\n \t\tstruct vxlan_mdb_entry *mdb_entry;\n \n-\t\trcu_read_lock();\n-\t\tmdb_entry = vxlan_mdb_entry_skb_get(vxlan, skb, vni);\n+\t\tmdb_entry = vxlan_mdb_entry_skb_get(vxlan, cfg, skb, vni);\n \t\tif (mdb_entry) {\n-\t\t\tnetdev_tx_t ret;\n+\t\t\tnetdev_tx_t ret = vxlan_mdb_xmit(vxlan, cfg, mdb_entry, skb);\n \n-\t\t\tret = vxlan_mdb_xmit(vxlan, mdb_entry, skb);\n \t\t\trcu_read_unlock();\n \t\t\treturn ret;\n \t\t}\n-\t\trcu_read_unlock();\n \t}\n \n \teth = eth_hdr(skb);\n-\trcu_read_lock();\n-\tf = vxlan_find_mac_tx(vxlan, eth-\u003eh_dest, vni);\n+\tf = vxlan_find_mac_tx(vxlan, cfg, eth-\u003eh_dest, vni);\n \tdid_rsc = false;\n \n-\tif (f \u0026\u0026 (f-\u003eflags \u0026 NTF_ROUTER) \u0026\u0026 (vxlan-\u003ecfg.flags \u0026 VXLAN_F_RSC) \u0026\u0026\n+\tif (f \u0026\u0026 (f-\u003eflags \u0026 NTF_ROUTER) \u0026\u0026 (flags \u0026 VXLAN_F_RSC) \u0026\u0026\n \t    (ntohs(eth-\u003eh_proto) == ETH_P_IP ||\n \t     ntohs(eth-\u003eh_proto) == ETH_P_IPV6)) {\n-\t\tdid_rsc = route_shortcircuit(dev, skb);\n+\t\tdid_rsc = route_shortcircuit(dev, skb, cfg);\n \t\teth = eth_hdr(skb);\n \t\tif (did_rsc)\n-\t\t\tf = vxlan_find_mac_tx(vxlan, eth-\u003eh_dest, vni);\n+\t\t\tf = vxlan_find_mac_tx(vxlan, cfg, eth-\u003eh_dest, vni);\n \t}\n \n \tif (f == NULL) {\n-\t\tf = vxlan_find_mac_tx(vxlan, all_zeros_mac, vni);\n+\t\tf = vxlan_find_mac_tx(vxlan, cfg, all_zeros_mac, vni);\n \t\tif (f == NULL) {\n-\t\t\tif ((vxlan-\u003ecfg.flags \u0026 VXLAN_F_L2MISS) \u0026\u0026\n+\t\t\tif ((flags \u0026 VXLAN_F_L2MISS) \u0026\u0026\n \t\t\t    !is_multicast_ether_addr(eth-\u003eh_dest))\n \t\t\t\tvxlan_fdb_miss(vxlan, eth-\u003eh_dest);\n \n \t\t\tdev_dstats_tx_dropped(dev);\n-\t\t\tvxlan_vnifilter_count(vxlan, vni, NULL,\n+\t\t\tvxlan_vnifilter_count(vxlan, cfg, vni, NULL,\n \t\t\t\t\t      VXLAN_VNI_STATS_TX_DROPS, 0);\n \t\t\tkfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);\n \t\t\tgoto out;\n@@ -2846,8 +2897,8 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)\n \t}\n \n \tif (rcu_access_pointer(f-\u003enh)) {\n-\t\tvxlan_xmit_nh(skb, dev, f,\n-\t\t\t      (vni ? : vxlan-\u003edefault_dst.remote_vni), did_rsc);\n+\t\tvxlan_xmit_nh(skb, dev, cfg, f,\n+\t\t\t      (vni ? : default_vni), did_rsc);\n \t} else {\n \t\tlist_for_each_entry_rcu(rdst, \u0026f-\u003eremotes, list) {\n \t\t\tstruct sk_buff *skb1;\n@@ -2858,10 +2909,10 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)\n \t\t\t}\n \t\t\tskb1 = skb_clone(skb, GFP_ATOMIC);\n \t\t\tif (skb1)\n-\t\t\t\tvxlan_xmit_one(skb1, dev, vni, rdst, did_rsc);\n+\t\t\t\tvxlan_xmit_one(skb1, dev, cfg, vni, rdst, did_rsc);\n \t\t}\n \t\tif (fdst)\n-\t\t\tvxlan_xmit_one(skb, dev, vni, fdst, did_rsc);\n+\t\t\tvxlan_xmit_one(skb, dev, cfg, vni, fdst, did_rsc);\n \t\telse\n \t\t\tkfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);\n \t}\n@@ -2876,12 +2927,15 @@ static void vxlan_cleanup(struct timer_list *t)\n {\n \tstruct vxlan_dev *vxlan = timer_container_of(vxlan, t, age_timer);\n \tunsigned long next_timer = jiffies + FDB_AGE_INTERVAL;\n+\tconst struct vxlan_config *cfg;\n \tstruct vxlan_fdb *f;\n \n \tif (!netif_running(vxlan-\u003edev))\n \t\treturn;\n \n \trcu_read_lock();\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\n \thlist_for_each_entry_rcu(f, \u0026vxlan-\u003efdb_list, fdb_node) {\n \t\tunsigned long timeout;\n \n@@ -2891,7 +2945,7 @@ static void vxlan_cleanup(struct timer_list *t)\n \t\tif (f-\u003eflags \u0026 NTF_EXT_LEARNED)\n \t\t\tcontinue;\n \n-\t\ttimeout = READ_ONCE(f-\u003eupdated) + vxlan-\u003ecfg.age_interval * HZ;\n+\t\ttimeout = READ_ONCE(f-\u003eupdated) + cfg-\u003eage_interval * HZ;\n \t\tif (time_before_eq(timeout, jiffies)) {\n \t\t\tspin_lock(\u0026vxlan-\u003ehash_lock);\n \t\t\tif (!hlist_unhashed(\u0026f-\u003efdb_node)) {\n@@ -2923,10 +2977,14 @@ static void vxlan_vs_del_dev(struct vxlan_dev *vxlan)\n static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,\n \t\t\t     struct vxlan_dev_node *node)\n {\n-\t__be32 vni = vxlan-\u003edefault_dst.remote_vni;\n+\tconst struct vxlan_config *cfg;\n+\t__be32 vni;\n \n \tASSERT_RTNL();\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tvni = cfg-\u003evni;\n+\n \tnode-\u003evxlan = vxlan;\n \thlist_add_head_rcu(\u0026node-\u003ehlist, vni_head(vs, vni));\n }\n@@ -2935,13 +2993,16 @@ static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,\n static int vxlan_init(struct net_device *dev)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tconst struct vxlan_config *cfg;\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n \terr = rhashtable_init(\u0026vxlan-\u003efdb_hash_tbl, \u0026vxlan_fdb_rht_params);\n \tif (err)\n \t\treturn err;\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER) {\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER) {\n \t\terr = vxlan_vnigroup_init(vxlan);\n \t\tif (err)\n \t\t\tgoto err_rhashtable_destroy;\n@@ -2961,7 +3022,7 @@ static int vxlan_init(struct net_device *dev)\n err_gro_cells_destroy:\n \tgro_cells_destroy(\u0026vxlan-\u003egro_cells);\n err_vnigroup_uninit:\n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)\n \t\tvxlan_vnigroup_uninit(vxlan);\n err_rhashtable_destroy:\n \trhashtable_destroy(\u0026vxlan-\u003efdb_hash_tbl);\n@@ -2971,10 +3032,13 @@ static int vxlan_init(struct net_device *dev)\n static void vxlan_uninit(struct net_device *dev)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tconst struct vxlan_config *cfg;\n+\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \tvxlan_mdb_fini(vxlan);\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)\n \t\tvxlan_vnigroup_uninit(vxlan);\n \n \tgro_cells_destroy(\u0026vxlan-\u003egro_cells);\n@@ -2986,6 +3050,7 @@ static void vxlan_uninit(struct net_device *dev)\n static int vxlan_open(struct net_device *dev)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tconst struct vxlan_config *cfg;\n \tint ret;\n \n \tret = vxlan_sock_add(vxlan);\n@@ -2998,7 +3063,8 @@ static int vxlan_open(struct net_device *dev)\n \t\treturn ret;\n \t}\n \n-\tif (vxlan-\u003ecfg.age_interval)\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tif (cfg-\u003eage_interval)\n \t\tmod_timer(\u0026vxlan-\u003eage_timer, jiffies + FDB_AGE_INTERVAL);\n \n \treturn ret;\n@@ -3020,8 +3086,10 @@ struct vxlan_fdb_flush_desc {\n static bool vxlan_fdb_is_default_entry(const struct vxlan_fdb *f,\n \t\t\t\t       const struct vxlan_dev *vxlan)\n {\n+\tconst struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n+\n \treturn is_zero_ether_addr(f-\u003ekey.eth_addr) \u0026\u0026\n-\t       f-\u003ekey.vni == vxlan-\u003ecfg.vni;\n+\t       f-\u003ekey.vni == cfg-\u003evni;\n }\n \n static bool vxlan_fdb_nhid_matches(const struct vxlan_fdb *f, u32 nhid)\n@@ -3238,15 +3306,18 @@ static void vxlan_set_multicast_list(struct net_device *dev)\n static int vxlan_change_mtu(struct net_device *dev, int new_mtu)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n-\tstruct net_device *lowerdev = __dev_get_by_index(vxlan-\u003enet,\n-\t\t\t\t\t\t\t dst-\u003eremote_ifindex);\n+\tconst struct vxlan_config *cfg;\n+\tstruct net_device *lowerdev;\n+\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n+\tlowerdev = __dev_get_by_index(vxlan-\u003enet, cfg-\u003eremote_ifindex);\n \n \t/* This check is different than dev-\u003emax_mtu, because it looks at\n \t * the lowerdev-\u003emtu, rather than the static dev-\u003emax_mtu\n \t */\n \tif (lowerdev) {\n-\t\tint max_mtu = lowerdev-\u003emtu - vxlan_headroom(vxlan-\u003ecfg.flags);\n+\t\tint max_mtu = lowerdev-\u003emtu - vxlan_headroom(cfg-\u003eflags);\n \t\tif (new_mtu \u003e max_mtu)\n \t\t\treturn -EINVAL;\n \t}\n@@ -3259,11 +3330,14 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct ip_tunnel_info *info = skb_tunnel_info(skb);\n+\tconst struct vxlan_config *cfg;\n \t__be16 sport, dport;\n \n-\tsport = udp_flow_src_port(dev_net(dev), skb, vxlan-\u003ecfg.port_min,\n-\t\t\t\t  vxlan-\u003ecfg.port_max, true);\n-\tdport = info-\u003ekey.tp_dst ? : vxlan-\u003ecfg.dst_port;\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\n+\tsport = udp_flow_src_port(dev_net(dev), skb, cfg-\u003eport_min,\n+\t\t\t\t  cfg-\u003eport_max, true);\n+\tdport = info-\u003ekey.tp_dst ? : cfg-\u003edst_port;\n \n \tif (ip_tunnel_info_af(info) == AF_INET) {\n \t\tstruct vxlan_sock *sock4 = rcu_dereference(vxlan-\u003evn4_sock);\n@@ -3373,6 +3447,15 @@ static void vxlan_offload_rx_ports(struct net_device *dev, bool push)\n \t}\n }\n \n+static void vxlan_free_dev(struct net_device *dev)\n+{\n+\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tstruct vxlan_config *cfg = rcu_dereference_protected(vxlan-\u003ecfg, 1);\n+\n+\tRCU_INIT_POINTER(vxlan-\u003ecfg, NULL);\n+\tkfree(cfg);\n+}\n+\n /* Initialize the device structure. */\n static void vxlan_setup(struct net_device *dev)\n {\n@@ -3381,6 +3464,8 @@ static void vxlan_setup(struct net_device *dev)\n \teth_hw_addr_random(dev);\n \tether_setup(dev);\n \n+\tdev-\u003epriv_destructor = vxlan_free_dev;\n+\n \tdev-\u003eneeds_free_netdev = true;\n \tSET_NETDEV_DEVTYPE(dev, \u0026vxlan_type);\n \n@@ -3555,9 +3640,11 @@ static int vxlan_get_link_ksettings(struct net_device *dev,\n \t\t\t\t    struct ethtool_link_ksettings *cmd)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n-\tstruct net_device *lowerdev = __dev_get_by_index(vxlan-\u003enet,\n-\t\t\t\t\t\t\t dst-\u003eremote_ifindex);\n+\tconst struct vxlan_config *cfg;\n+\tstruct net_device *lowerdev;\n+\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tlowerdev = __dev_get_by_index(vxlan-\u003enet, cfg-\u003eremote_ifindex);\n \n \tif (!lowerdev) {\n \t\tcmd-\u003ebase.duplex = DUPLEX_UNKNOWN;\n@@ -3663,21 +3750,22 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, bool ipv6,\n \n static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)\n {\n-\tbool metadata = vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA;\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tbool metadata = cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA;\n \tstruct vxlan_sock *vs = NULL;\n \tstruct vxlan_dev_node *node;\n \tint l3mdev_index = 0;\n \n \tASSERT_RTNL();\n \n-\tif (vxlan-\u003ecfg.remote_ifindex)\n+\tif (cfg-\u003eremote_ifindex)\n \t\tl3mdev_index = l3mdev_master_upper_ifindex_by_index(\n-\t\t\tvxlan-\u003enet, vxlan-\u003ecfg.remote_ifindex);\n+\t\t\tvxlan-\u003enet, cfg-\u003eremote_ifindex);\n \n-\tif (!vxlan-\u003ecfg.no_share) {\n+\tif (!cfg-\u003eno_share) {\n \t\trcu_read_lock();\n \t\tvs = vxlan_find_sock(vxlan-\u003enet, ipv6 ? AF_INET6 : AF_INET,\n-\t\t\t\t     vxlan-\u003ecfg.dst_port, vxlan-\u003ecfg.flags,\n+\t\t\t\t     cfg-\u003edst_port, cfg-\u003eflags,\n \t\t\t\t     l3mdev_index);\n \t\tif (vs \u0026\u0026 !refcount_inc_not_zero(\u0026vs-\u003erefcnt)) {\n \t\t\trcu_read_unlock();\n@@ -3687,7 +3775,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)\n \t}\n \tif (!vs)\n \t\tvs = vxlan_socket_create(vxlan-\u003enet, ipv6,\n-\t\t\t\t\t vxlan-\u003ecfg.dst_port, vxlan-\u003ecfg.flags,\n+\t\t\t\t\t cfg-\u003edst_port, cfg-\u003eflags,\n \t\t\t\t\t l3mdev_index);\n \tif (IS_ERR(vs))\n \t\treturn PTR_ERR(vs);\n@@ -3702,7 +3790,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)\n \t\tnode = \u0026vxlan-\u003ehlist4;\n \t}\n \n-\tif (metadata \u0026\u0026 (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER))\n+\tif (metadata \u0026\u0026 (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER))\n \t\tvxlan_vs_add_vnigrp(vxlan, vs, ipv6);\n \telse\n \t\tvxlan_vs_add_dev(vs, vxlan, node);\n@@ -3712,11 +3800,14 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)\n \n static int vxlan_sock_add(struct vxlan_dev *vxlan)\n {\n-\tbool metadata = vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA;\n-\tbool ipv6 = vxlan-\u003ecfg.flags \u0026 VXLAN_F_IPV6 || metadata;\n-\tbool ipv4 = !ipv6 || metadata;\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tbool metadata, ipv6, ipv4;\n \tint ret = 0;\n \n+\tmetadata = cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA;\n+\tipv6 = (cfg-\u003eflags \u0026 VXLAN_F_IPV6) || metadata;\n+\tipv4 = !ipv6 || metadata;\n+\n \tRCU_INIT_POINTER(vxlan-\u003evn4_sock, NULL);\n #if IS_ENABLED(CONFIG_IPV6)\n \tRCU_INIT_POINTER(vxlan-\u003evn6_sock, NULL);\n@@ -3734,28 +3825,33 @@ static int vxlan_sock_add(struct vxlan_dev *vxlan)\n }\n \n int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,\n-\t\t     struct vxlan_config *conf, __be32 vni)\n+\t\t     const struct vxlan_config *conf, __be32 vni)\n {\n \tstruct vxlan_net *vn = net_generic(src_net, vxlan_net_id);\n \tstruct vxlan_dev *tmp;\n \n \tlist_for_each_entry(tmp, \u0026vn-\u003evxlan_list, next) {\n+\t\tconst struct vxlan_config *tmp_cfg;\n+\n \t\tif (tmp == vxlan)\n \t\t\tcontinue;\n-\t\tif (tmp-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER) {\n+\n+\t\ttmp_cfg = rtnl_dereference(tmp-\u003ecfg);\n+\n+\t\tif (tmp_cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER) {\n \t\t\tif (!vxlan_vnifilter_lookup(tmp, vni))\n \t\t\t\tcontinue;\n-\t\t} else if (tmp-\u003ecfg.vni != vni) {\n+\t\t} else if (tmp_cfg-\u003evni != vni) {\n \t\t\tcontinue;\n \t\t}\n-\t\tif (tmp-\u003ecfg.dst_port != conf-\u003edst_port)\n+\t\tif (tmp_cfg-\u003edst_port != conf-\u003edst_port)\n \t\t\tcontinue;\n-\t\tif ((tmp-\u003ecfg.flags \u0026 (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=\n+\t\tif ((tmp_cfg-\u003eflags \u0026 (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=\n \t\t    (conf-\u003eflags \u0026 (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)))\n \t\t\tcontinue;\n \n \t\tif ((conf-\u003eflags \u0026 VXLAN_F_IPV6_LINKLOCAL) \u0026\u0026\n-\t\t    tmp-\u003ecfg.remote_ifindex != conf-\u003eremote_ifindex)\n+\t\t    tmp_cfg-\u003eremote_ifindex != conf-\u003eremote_ifindex)\n \t\t\tcontinue;\n \n \t\treturn -EEXIST;\n@@ -3886,6 +3982,13 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,\n \t\t\treturn -EINVAL;\n \t\t}\n \n+\t\tif ((conf-\u003eflags \u0026 VXLAN_F_VNIFILTER) \u0026\u0026 old \u0026\u0026\n+\t\t    vxlan_vnifilter_has_multicast(old)) {\n+\t\t\tNL_SET_ERR_MSG(extack,\n+\t\t\t\t       \"Local interface required for multicast remote group\");\n+\t\t\treturn -EINVAL;\n+\t\t}\n+\n #if IS_ENABLED(CONFIG_IPV6)\n \t\tif (conf-\u003eflags \u0026 VXLAN_F_IPV6_LINKLOCAL) {\n \t\t\tNL_SET_ERR_MSG(extack,\n@@ -3917,16 +4020,16 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,\n }\n \n static void vxlan_config_apply(struct net_device *dev,\n-\t\t\t       struct vxlan_config *conf,\n+\t\t\t       struct vxlan_config *new_cfg,\n \t\t\t       struct net_device *lowerdev,\n \t\t\t       struct net *src_net,\n-\t\t\t       bool changelink)\n+\t\t\t       bool changelink, bool lowerdev_changed)\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n \tunsigned short needed_headroom = ETH_HLEN;\n+\tstruct vxlan_config *old_cfg;\n \tint max_mtu = ETH_MAX_MTU;\n-\tu32 flags = conf-\u003eflags;\n+\tu32 flags = new_cfg-\u003eflags;\n \n \tif (!changelink) {\n \t\tif (flags \u0026 VXLAN_F_GPE)\n@@ -3934,19 +4037,14 @@ static void vxlan_config_apply(struct net_device *dev,\n \t\telse\n \t\t\tvxlan_ether_setup(dev);\n \n-\t\tif (conf-\u003emtu)\n-\t\t\tdev-\u003emtu = conf-\u003emtu;\n+\t\tif (new_cfg-\u003emtu)\n+\t\t\tWRITE_ONCE(dev-\u003emtu, new_cfg-\u003emtu);\n \n \t\tvxlan-\u003enet = src_net;\n \t}\n \n-\tdst-\u003eremote_vni = conf-\u003evni;\n-\n-\tmemcpy(\u0026dst-\u003eremote_ip, \u0026conf-\u003eremote_ip, sizeof(conf-\u003eremote_ip));\n-\n+\tdev-\u003eneeded_tailroom = 0;\n \tif (lowerdev) {\n-\t\tdst-\u003eremote_ifindex = conf-\u003eremote_ifindex;\n-\n \t\tnetif_inherit_tso_max(dev, lowerdev);\n \n \t\tneeded_headroom = lowerdev-\u003ehard_header_len;\n@@ -3954,23 +4052,27 @@ static void vxlan_config_apply(struct net_device *dev,\n \n \t\tdev-\u003eneeded_tailroom = lowerdev-\u003eneeded_tailroom;\n \n-\t\tmax_mtu = lowerdev-\u003emtu - vxlan_headroom(flags);\n+\t\tmax_mtu = READ_ONCE(lowerdev-\u003emtu) - vxlan_headroom(flags);\n \t\tif (max_mtu \u003c ETH_MIN_MTU)\n \t\t\tmax_mtu = ETH_MIN_MTU;\n \n-\t\tif (!changelink \u0026\u0026 !conf-\u003emtu)\n-\t\t\tdev-\u003emtu = max_mtu;\n+\t\tif (!changelink \u0026\u0026 !new_cfg-\u003emtu)\n+\t\t\tWRITE_ONCE(dev-\u003emtu, max_mtu);\n \t}\n \n-\tif (dev-\u003emtu \u003e max_mtu)\n-\t\tdev-\u003emtu = max_mtu;\n+\t/* A changelink leaving the lower device alone must not shrink the mtu */\n+\tif (lowerdev_changed \u0026\u0026 READ_ONCE(dev-\u003emtu) \u003e max_mtu)\n+\t\tWRITE_ONCE(dev-\u003emtu, max_mtu);\n \n \tif (flags \u0026 VXLAN_F_COLLECT_METADATA)\n \t\tflags |= VXLAN_F_IPV6;\n \tneeded_headroom += vxlan_headroom(flags);\n \tdev-\u003eneeded_headroom = needed_headroom;\n \n-\tmemcpy(\u0026vxlan-\u003ecfg, conf, sizeof(*conf));\n+\told_cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\trcu_assign_pointer(vxlan-\u003ecfg, new_cfg);\n+\tif (old_cfg)\n+\t\tkfree_rcu(old_cfg, rcu);\n }\n \n static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,\n@@ -3979,13 +4081,18 @@ static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,\n {\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct net_device *lowerdev;\n+\tstruct vxlan_config *new_cfg;\n \tint ret;\n \n \tret = vxlan_config_validate(src_net, conf, \u0026lowerdev, vxlan, extack);\n \tif (ret)\n \t\treturn ret;\n \n-\tvxlan_config_apply(dev, conf, lowerdev, src_net, false);\n+\tnew_cfg = kmemdup(conf, sizeof(*conf), GFP_KERNEL);\n+\tif (!new_cfg)\n+\t\treturn -ENOMEM;\n+\n+\tvxlan_config_apply(dev, new_cfg, lowerdev, src_net, false, true);\n \n \treturn 0;\n }\n@@ -3997,22 +4104,25 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,\n \tstruct vxlan_net *vn = net_generic(net, vxlan_net_id);\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct net_device *remote_dev = NULL;\n-\tstruct vxlan_rdst *dst;\n+\tconst struct vxlan_config *cfg;\n \tint err;\n \n-\tdst = \u0026vxlan-\u003edefault_dst;\n \terr = vxlan_dev_configure(net, dev, conf, extack);\n \tif (err)\n \t\treturn err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n \tdev-\u003eethtool_ops = \u0026vxlan_ethtool_ops;\n \n \terr = register_netdevice(dev);\n-\tif (err)\n+\tif (err) {\n+\t\tvxlan_free_dev(dev);\n \t\treturn err;\n+\t}\n \n-\tif (dst-\u003eremote_ifindex) {\n-\t\tremote_dev = __dev_get_by_index(net, dst-\u003eremote_ifindex);\n+\tif (cfg-\u003eremote_ifindex) {\n+\t\tremote_dev = __dev_get_by_index(net, cfg-\u003eremote_ifindex);\n \t\tif (!remote_dev) {\n \t\t\terr = -ENODEV;\n \t\t\tgoto unregister;\n@@ -4022,7 +4132,7 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,\n \t\tif (err)\n \t\t\tgoto unregister;\n \n-\t\tdst-\u003eremote_dev = remote_dev;\n+\t\tvxlan-\u003elowerdev = remote_dev;\n \t}\n \n \terr = rtnl_configure_link(dev, NULL, 0, NULL);\n@@ -4030,16 +4140,18 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,\n \t\tgoto unlink;\n \n \t/* create an fdb entry for a valid default destination */\n-\tif (!vxlan_addr_any(\u0026dst-\u003eremote_ip)) {\n+\tif (!vxlan_addr_any(\u0026cfg-\u003eremote_ip)) {\n+\t\tunion vxlan_addr rip = cfg-\u003eremote_ip;\n+\n \t\tspin_lock_bh(\u0026vxlan-\u003ehash_lock);\n \t\terr = vxlan_fdb_update(vxlan, all_zeros_mac,\n-\t\t\t\t       \u0026dst-\u003eremote_ip,\n+\t\t\t\t       \u0026rip,\n \t\t\t\t       NUD_REACHABLE | NUD_PERMANENT,\n \t\t\t\t       NLM_F_EXCL | NLM_F_CREATE,\n-\t\t\t\t       vxlan-\u003ecfg.dst_port,\n-\t\t\t\t       dst-\u003eremote_vni,\n-\t\t\t\t       dst-\u003eremote_vni,\n-\t\t\t\t       dst-\u003eremote_ifindex,\n+\t\t\t\t       cfg-\u003edst_port,\n+\t\t\t\t       cfg-\u003evni,\n+\t\t\t\t       cfg-\u003evni,\n+\t\t\t\t       cfg-\u003eremote_ifindex,\n \t\t\t\t       NTF_SELF, 0, true, extack);\n \t\tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n \t\tif (err)\n@@ -4100,8 +4212,12 @@ static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],\n \tmemset(conf, 0, sizeof(*conf));\n \n \t/* if changelink operation, start with old existing cfg */\n-\tif (changelink)\n-\t\tmemcpy(conf, \u0026vxlan-\u003ecfg, sizeof(*conf));\n+\tif (changelink) {\n+\t\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n+\t\tif (cfg)\n+\t\t\tmemcpy(conf, cfg, sizeof(*conf));\n+\t}\n \n \tif (data[IFLA_VXLAN_ID]) {\n \t\t__be32 vni = cpu_to_be32(nla_get_u32(data[IFLA_VXLAN_ID]));\n@@ -4447,17 +4563,19 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],\n \t\t\t    struct nlattr *data[],\n \t\t\t    struct netlink_ext_ack *extack)\n {\n+\tbool lowerdev_changed, rem_ip_changed, change_igmp;\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n-\tbool rem_ip_changed, change_igmp;\n+\tconst struct vxlan_config *cfg;\n+\tstruct vxlan_config *new_cfg;\n \tstruct net_device *lowerdev;\n \tstruct vxlan_config conf;\n-\tstruct vxlan_rdst *dst;\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n \tif (!rtnl_dev_link_net_capable(dev, vxlan-\u003enet))\n \t\treturn -EPERM;\n \n-\tdst = \u0026vxlan-\u003edefault_dst;\n \terr = vxlan_nl2conf(tb, data, dev, \u0026conf, true, extack);\n \tif (err)\n \t\treturn err;\n@@ -4467,76 +4585,108 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],\n \tif (err)\n \t\treturn err;\n \n-\tif (dst-\u003eremote_dev == lowerdev)\n-\t\tlowerdev = NULL;\n+\tnew_cfg = kmemdup(\u0026conf, sizeof(conf), GFP_KERNEL);\n+\tif (!new_cfg)\n+\t\treturn -ENOMEM;\n \n-\terr = netdev_adjacent_change_prepare(dst-\u003eremote_dev, lowerdev, dev,\n-\t\t\t\t\t     extack);\n-\tif (err)\n-\t\treturn err;\n+\tlowerdev_changed = vxlan-\u003elowerdev != lowerdev;\n+\tif (lowerdev_changed) {\n+\t\terr = netdev_adjacent_change_prepare(vxlan-\u003elowerdev, lowerdev,\n+\t\t\t\t\t\t     dev, extack);\n+\t\tif (err) {\n+\t\t\tkfree(new_cfg);\n+\t\t\treturn err;\n+\t\t}\n+\t}\n \n-\trem_ip_changed = !vxlan_addr_equal(\u0026conf.remote_ip, \u0026dst-\u003eremote_ip);\n+\trem_ip_changed = !vxlan_addr_equal(\u0026conf.remote_ip, \u0026cfg-\u003eremote_ip);\n \tchange_igmp = vxlan-\u003edev-\u003eflags \u0026 IFF_UP \u0026\u0026\n \t\t      (rem_ip_changed ||\n-\t\t       dst-\u003eremote_ifindex != conf.remote_ifindex);\n+\t\t       cfg-\u003eremote_ifindex != conf.remote_ifindex);\n \n \t/* handle default dst entry */\n-\tif (rem_ip_changed) {\n+\tif (rem_ip_changed || cfg-\u003eremote_ifindex != conf.remote_ifindex) {\n \t\tspin_lock_bh(\u0026vxlan-\u003ehash_lock);\n \t\tif (!vxlan_addr_any(\u0026conf.remote_ip)) {\n \t\t\terr = vxlan_fdb_update(vxlan, all_zeros_mac,\n \t\t\t\t\t       \u0026conf.remote_ip,\n \t\t\t\t\t       NUD_REACHABLE | NUD_PERMANENT,\n \t\t\t\t\t       NLM_F_APPEND | NLM_F_CREATE,\n-\t\t\t\t\t       vxlan-\u003ecfg.dst_port,\n+\t\t\t\t\t       cfg-\u003edst_port,\n \t\t\t\t\t       conf.vni, conf.vni,\n \t\t\t\t\t       conf.remote_ifindex,\n \t\t\t\t\t       NTF_SELF, 0, true, extack);\n \t\t\tif (err) {\n \t\t\t\tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n-\t\t\t\tnetdev_adjacent_change_abort(dst-\u003eremote_dev,\n-\t\t\t\t\t\t\t     lowerdev, dev);\n+\t\t\t\tif (lowerdev_changed)\n+\t\t\t\t\tnetdev_adjacent_change_abort(vxlan-\u003elowerdev,\n+\t\t\t\t\t\t\t\t     lowerdev, dev);\n+\t\t\t\tkfree(new_cfg);\n \t\t\t\treturn err;\n \t\t\t}\n \t\t}\n-\t\tif (!vxlan_addr_any(\u0026dst-\u003eremote_ip))\n+\t\tif (!vxlan_addr_any(\u0026cfg-\u003eremote_ip))\n \t\t\t__vxlan_fdb_delete(vxlan, all_zeros_mac,\n-\t\t\t\t\t   dst-\u003eremote_ip,\n-\t\t\t\t\t   vxlan-\u003ecfg.dst_port,\n-\t\t\t\t\t   dst-\u003eremote_vni,\n-\t\t\t\t\t   dst-\u003eremote_vni,\n-\t\t\t\t\t   dst-\u003eremote_ifindex,\n+\t\t\t\t\t   cfg-\u003eremote_ip,\n+\t\t\t\t\t   cfg-\u003edst_port,\n+\t\t\t\t\t   cfg-\u003evni,\n+\t\t\t\t\t   cfg-\u003evni,\n+\t\t\t\t\t   cfg-\u003eremote_ifindex,\n \t\t\t\t\t   true);\n \t\tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n \n-\t\t/* If vni filtering device, also update fdb entries of\n-\t\t * all vnis that were using default remote ip\n+\t\t/* If vni filtering device, also update default fdb entries of\n+\t\t * all vnis\n \t\t */\n-\t\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER) {\n-\t\t\terr = vxlan_vnilist_update_group(vxlan, \u0026dst-\u003eremote_ip,\n-\t\t\t\t\t\t\t \u0026conf.remote_ip, extack);\n+\t\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER) {\n+\t\t\terr = vxlan_vnilist_update_group(vxlan, \u0026cfg-\u003eremote_ip,\n+\t\t\t\t\t\t\t \u0026conf.remote_ip,\n+\t\t\t\t\t\t\t cfg-\u003eremote_ifindex,\n+\t\t\t\t\t\t\t conf.remote_ifindex,\n+\t\t\t\t\t\t\t extack);\n \t\t\tif (err) {\n-\t\t\t\tnetdev_adjacent_change_abort(dst-\u003eremote_dev,\n-\t\t\t\t\t\t\t     lowerdev, dev);\n+\t\t\t\tvxlan_update_default_fdb_entry(vxlan, conf.vni,\n+\t\t\t\t\t\t\t       \u0026conf.remote_ip,\n+\t\t\t\t\t\t\t       \u0026cfg-\u003eremote_ip,\n+\t\t\t\t\t\t\t       conf.remote_ifindex,\n+\t\t\t\t\t\t\t       cfg-\u003eremote_ifindex,\n+\t\t\t\t\t\t\t       NULL);\n+\t\t\t\tif (lowerdev_changed)\n+\t\t\t\t\tnetdev_adjacent_change_abort(vxlan-\u003elowerdev,\n+\t\t\t\t\t\t\t\t     lowerdev, dev);\n+\t\t\t\tkfree(new_cfg);\n \t\t\t\treturn err;\n \t\t\t}\n \t\t}\n \t}\n \n-\tif (change_igmp \u0026\u0026 vxlan_addr_multicast(\u0026dst-\u003eremote_ip))\n+\tif (change_igmp \u0026\u0026\n+\t    (vxlan_addr_multicast(\u0026cfg-\u003eremote_ip) ||\n+\t     (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)))\n \t\terr = vxlan_multicast_leave(vxlan);\n \n-\tif (netif_running(dev) \u0026\u0026 conf.age_interval != vxlan-\u003ecfg.age_interval)\n+\tif (netif_running(dev) \u0026\u0026 conf.age_interval != cfg-\u003eage_interval)\n \t\tmod_timer(\u0026vxlan-\u003eage_timer, jiffies);\n \n-\tnetdev_adjacent_change_commit(dst-\u003eremote_dev, lowerdev, dev);\n-\tif (lowerdev \u0026\u0026 lowerdev != dst-\u003eremote_dev)\n-\t\tdst-\u003eremote_dev = lowerdev;\n-\tvxlan_config_apply(dev, \u0026conf, lowerdev, vxlan-\u003enet, true);\n+\tif (lowerdev_changed) {\n+\t\tif (lowerdev)\n+\t\t\tnetdev_adjacent_change_commit(vxlan-\u003elowerdev, lowerdev,\n+\t\t\t\t\t\t      dev);\n+\t\telse\n+\t\t\tnetdev_upper_dev_unlink(vxlan-\u003elowerdev, dev);\n+\t\tvxlan-\u003elowerdev = lowerdev;\n+\t}\n+\tvxlan_config_apply(dev, new_cfg, lowerdev, vxlan-\u003enet, true,\n+\t\t\t   lowerdev_changed);\n+\n+\tif (change_igmp \u0026\u0026\n+\t    (vxlan_addr_multicast(\u0026new_cfg-\u003eremote_ip) ||\n+\t     (new_cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER))) {\n+\t\tint join_err = vxlan_multicast_join(vxlan);\n \n-\tif (!err \u0026\u0026 change_igmp \u0026\u0026\n-\t    vxlan_addr_multicast(\u0026dst-\u003eremote_ip))\n-\t\terr = vxlan_multicast_join(vxlan);\n+\t\tif (join_err)\n+\t\t\terr = join_err;\n+\t}\n \n \treturn err;\n }\n@@ -4550,8 +4700,8 @@ static void vxlan_dellink(struct net_device *dev, struct list_head *head)\n \n \tlist_del(\u0026vxlan-\u003enext);\n \tunregister_netdevice_queue(dev, head);\n-\tif (vxlan-\u003edefault_dst.remote_dev)\n-\t\tnetdev_upper_dev_unlink(vxlan-\u003edefault_dst.remote_dev, dev);\n+\tif (vxlan-\u003elowerdev)\n+\t\tnetdev_upper_dev_unlink(vxlan-\u003elowerdev, dev);\n }\n \n static size_t vxlan_get_size(const struct net_device *dev)\n@@ -4595,110 +4745,114 @@ static size_t vxlan_get_size(const struct net_device *dev)\n static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)\n {\n \tconst struct vxlan_dev *vxlan = netdev_priv(dev);\n-\tconst struct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n-\tstruct ifla_vxlan_port_range ports = {\n-\t\t.low =  htons(vxlan-\u003ecfg.port_min),\n-\t\t.high = htons(vxlan-\u003ecfg.port_max),\n-\t};\n+\tstruct ifla_vxlan_port_range ports;\n+\tconst struct vxlan_config *cfg;\n+\n+\trcu_read_lock();\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n \n-\tif (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(dst-\u003eremote_vni)))\n+\tif (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(cfg-\u003evni)))\n \t\tgoto nla_put_failure;\n \n-\tif (!vxlan_addr_any(\u0026dst-\u003eremote_ip)) {\n-\t\tif (dst-\u003eremote_ip.sa.sa_family == AF_INET) {\n+\tif (!vxlan_addr_any(\u0026cfg-\u003eremote_ip)) {\n+\t\tif (cfg-\u003eremote_ip.sa.sa_family == AF_INET) {\n \t\t\tif (nla_put_in_addr(skb, IFLA_VXLAN_GROUP,\n-\t\t\t\t\t    dst-\u003eremote_ip.sin.sin_addr.s_addr))\n+\t\t\t\t\t    cfg-\u003eremote_ip.sin.sin_addr.s_addr))\n \t\t\t\tgoto nla_put_failure;\n #if IS_ENABLED(CONFIG_IPV6)\n \t\t} else {\n \t\t\tif (nla_put_in6_addr(skb, IFLA_VXLAN_GROUP6,\n-\t\t\t\t\t     \u0026dst-\u003eremote_ip.sin6.sin6_addr))\n+\t\t\t\t\t     \u0026cfg-\u003eremote_ip.sin6.sin6_addr))\n \t\t\t\tgoto nla_put_failure;\n #endif\n \t\t}\n \t}\n \n-\tif (dst-\u003eremote_ifindex \u0026\u0026 nla_put_u32(skb, IFLA_VXLAN_LINK, dst-\u003eremote_ifindex))\n+\tif (cfg-\u003eremote_ifindex \u0026\u0026 nla_put_u32(skb, IFLA_VXLAN_LINK, cfg-\u003eremote_ifindex))\n \t\tgoto nla_put_failure;\n \n-\tif (!vxlan_addr_any(\u0026vxlan-\u003ecfg.saddr)) {\n-\t\tif (vxlan-\u003ecfg.saddr.sa.sa_family == AF_INET) {\n+\tif (!vxlan_addr_any(\u0026cfg-\u003esaddr)) {\n+\t\tif (cfg-\u003esaddr.sa.sa_family == AF_INET) {\n \t\t\tif (nla_put_in_addr(skb, IFLA_VXLAN_LOCAL,\n-\t\t\t\t\t    vxlan-\u003ecfg.saddr.sin.sin_addr.s_addr))\n+\t\t\t\t\t    cfg-\u003esaddr.sin.sin_addr.s_addr))\n \t\t\t\tgoto nla_put_failure;\n #if IS_ENABLED(CONFIG_IPV6)\n \t\t} else {\n \t\t\tif (nla_put_in6_addr(skb, IFLA_VXLAN_LOCAL6,\n-\t\t\t\t\t     \u0026vxlan-\u003ecfg.saddr.sin6.sin6_addr))\n+\t\t\t\t\t     \u0026cfg-\u003esaddr.sin6.sin6_addr))\n \t\t\t\tgoto nla_put_failure;\n #endif\n \t\t}\n \t}\n \n-\tif (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan-\u003ecfg.ttl) ||\n+\tif (nla_put_u8(skb, IFLA_VXLAN_TTL, cfg-\u003ettl) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_TTL_INHERIT,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_TTL_INHERIT)) ||\n-\t    nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan-\u003ecfg.tos) ||\n-\t    nla_put_u8(skb, IFLA_VXLAN_DF, vxlan-\u003ecfg.df) ||\n-\t    nla_put_be32(skb, IFLA_VXLAN_LABEL, vxlan-\u003ecfg.label) ||\n-\t    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, vxlan-\u003ecfg.label_policy) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_TTL_INHERIT)) ||\n+\t    nla_put_u8(skb, IFLA_VXLAN_TOS, cfg-\u003etos) ||\n+\t    nla_put_u8(skb, IFLA_VXLAN_DF, cfg-\u003edf) ||\n+\t    nla_put_be32(skb, IFLA_VXLAN_LABEL, cfg-\u003elabel) ||\n+\t    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, cfg-\u003elabel_policy) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_LEARNING,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_LEARN)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_LEARN)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_PROXY,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_PROXY)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_PROXY)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_RSC,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_RSC)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_RSC)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_L2MISS,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_L2MISS)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_L2MISS)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_L3MISS,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_L3MISS)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_L3MISS)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_COLLECT_METADATA,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA)) ||\n-\t    nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan-\u003ecfg.age_interval) ||\n-\t    nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan-\u003ecfg.addrmax) ||\n-\t    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan-\u003ecfg.dst_port) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA)) ||\n+\t    nla_put_u32(skb, IFLA_VXLAN_AGEING, cfg-\u003eage_interval) ||\n+\t    nla_put_u32(skb, IFLA_VXLAN_LIMIT, cfg-\u003eaddrmax) ||\n+\t    nla_put_be16(skb, IFLA_VXLAN_PORT, cfg-\u003edst_port) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_UDP_CSUM,\n-\t\t       !(vxlan-\u003ecfg.flags \u0026 VXLAN_F_UDP_ZERO_CSUM_TX)) ||\n+\t\t       !(cfg-\u003eflags \u0026 VXLAN_F_UDP_ZERO_CSUM_TX)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_TX,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_UDP_ZERO_CSUM6_TX)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_UDP_ZERO_CSUM6_TX)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_RX,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_UDP_ZERO_CSUM6_RX)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_UDP_ZERO_CSUM6_RX)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_TX,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_REMCSUM_TX)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_REMCSUM_TX)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_RX,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_REMCSUM_RX)) ||\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_REMCSUM_RX)) ||\n \t    nla_put_u8(skb, IFLA_VXLAN_LOCALBYPASS,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_LOCALBYPASS)))\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_LOCALBYPASS)))\n \t\tgoto nla_put_failure;\n \n+\tports.low = htons(cfg-\u003eport_min);\n+\tports.high = htons(cfg-\u003eport_max);\n \tif (nla_put(skb, IFLA_VXLAN_PORT_RANGE, sizeof(ports), \u0026ports))\n \t\tgoto nla_put_failure;\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_GBP \u0026\u0026\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_GBP \u0026\u0026\n \t    nla_put_flag(skb, IFLA_VXLAN_GBP))\n \t\tgoto nla_put_failure;\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_GPE \u0026\u0026\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_GPE \u0026\u0026\n \t    nla_put_flag(skb, IFLA_VXLAN_GPE))\n \t\tgoto nla_put_failure;\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_REMCSUM_NOPARTIAL \u0026\u0026\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_REMCSUM_NOPARTIAL \u0026\u0026\n \t    nla_put_flag(skb, IFLA_VXLAN_REMCSUM_NOPARTIAL))\n \t\tgoto nla_put_failure;\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER \u0026\u0026\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER \u0026\u0026\n \t    nla_put_u8(skb, IFLA_VXLAN_VNIFILTER,\n-\t\t       !!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)))\n+\t\t       !!(cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)))\n \t\tgoto nla_put_failure;\n \n \tif (nla_put(skb, IFLA_VXLAN_RESERVED_BITS,\n-\t\t    sizeof(vxlan-\u003ecfg.reserved_bits),\n-\t\t    \u0026vxlan-\u003ecfg.reserved_bits))\n+\t\t    sizeof(cfg-\u003ereserved_bits),\n+\t\t    \u0026cfg-\u003ereserved_bits))\n \t\tgoto nla_put_failure;\n \n+\trcu_read_unlock();\n \treturn 0;\n \n nla_put_failure:\n+\trcu_read_unlock();\n \treturn -EMSGSIZE;\n }\n \n@@ -4731,7 +4885,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,\n \tLIST_HEAD(list_kill);\n \n \tlist_for_each_entry_safe(vxlan, next, \u0026vn-\u003evxlan_list, next) {\n-\t\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n+\t\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \t\t/* In case we created vxlan device with carrier\n \t\t * and we loose the carrier due to module unload\n@@ -4739,7 +4893,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,\n \t\t * cases, it's not necessary and remote_ifindex\n \t\t * is 0 here, so no matches.\n \t\t */\n-\t\tif (dst-\u003eremote_ifindex == dev-\u003eifindex)\n+\t\tif (cfg-\u003eremote_ifindex == dev-\u003eifindex)\n \t\t\tvxlan_dellink(vxlan-\u003edev, \u0026list_kill);\n \t}\n \n@@ -4752,6 +4906,10 @@ static int vxlan_netdevice_event(struct notifier_block *unused,\n \tstruct net_device *dev = netdev_notifier_info_to_dev(ptr);\n \tstruct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);\n \n+\tif ((event == NETDEV_REGISTER || event == NETDEV_UNREGISTER) \u0026\u0026\n+\t    netif_is_vxlan(dev))\n+\t\tvxlan_vnifilter_seq_inc(dev_net(dev));\n+\n \tif (event == NETDEV_UNREGISTER)\n \t\tvxlan_handle_lowerdev_unregister(vn, dev);\n \telse if (event == NETDEV_UDP_TUNNEL_PUSH_INFO)\ndiff --git a/drivers/net/vxlan/vxlan_mdb.c b/drivers/net/vxlan/vxlan_mdb.c\nindex 841f42ffecb9a..c1a0551990555 100644\n--- a/drivers/net/vxlan/vxlan_mdb.c\n+++ b/drivers/net/vxlan/vxlan_mdb.c\n@@ -165,6 +165,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,\n \t\t\t\t     const struct vxlan_mdb_entry *mdb_entry,\n \t\t\t\t     const struct vxlan_mdb_remote *remote)\n {\n+\tconst struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n \tstruct vxlan_rdst *rd = rtnl_dereference(remote-\u003erd);\n \tstruct br_mdb_entry e;\n \tstruct nlattr *nest;\n@@ -189,12 +190,12 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,\n \t    vxlan_nla_put_addr(skb, MDBA_MDB_EATTR_DST, \u0026rd-\u003eremote_ip))\n \t\tgoto nest_err;\n \n-\tif (rd-\u003eremote_port \u0026\u0026 rd-\u003eremote_port != vxlan-\u003ecfg.dst_port \u0026\u0026\n+\tif (rd-\u003eremote_port \u0026\u0026 rd-\u003eremote_port != cfg-\u003edst_port \u0026\u0026\n \t    nla_put_u16(skb, MDBA_MDB_EATTR_DST_PORT,\n \t\t\tbe16_to_cpu(rd-\u003eremote_port)))\n \t\tgoto nest_err;\n \n-\tif (rd-\u003eremote_vni != vxlan-\u003edefault_dst.remote_vni \u0026\u0026\n+\tif (rd-\u003eremote_vni != cfg-\u003evni \u0026\u0026\n \t    nla_put_u32(skb, MDBA_MDB_EATTR_VNI, be32_to_cpu(rd-\u003eremote_vni)))\n \t\tgoto nest_err;\n \n@@ -202,7 +203,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,\n \t    nla_put_u32(skb, MDBA_MDB_EATTR_IFINDEX, rd-\u003eremote_ifindex))\n \t\tgoto nest_err;\n \n-\tif ((vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026\n+\tif ((cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026\n \t    mdb_entry-\u003ekey.vni \u0026\u0026 nla_put_u32(skb, MDBA_MDB_EATTR_SRC_VNI,\n \t\t\t\t\t      be32_to_cpu(mdb_entry-\u003ekey.vni)))\n \t\tgoto nest_err;\n@@ -613,16 +614,19 @@ static int vxlan_mdb_config_init(struct vxlan_mdb_config *cfg,\n {\n \tstruct br_mdb_entry *entry = nla_data(tb[MDBA_SET_ENTRY]);\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n+\tconst struct vxlan_config *vcfg;\n+\n+\tvcfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \tmemset(cfg, 0, sizeof(*cfg));\n \tcfg-\u003evxlan = vxlan;\n-\tcfg-\u003egroup.vni = vxlan-\u003edefault_dst.remote_vni;\n+\tcfg-\u003egroup.vni = vcfg-\u003evni;\n \tINIT_LIST_HEAD(\u0026cfg-\u003esrc_list);\n \tcfg-\u003enlflags = nlmsg_flags;\n \tcfg-\u003efilter_mode = MCAST_EXCLUDE;\n \tcfg-\u003ert_protocol = RTPROT_STATIC;\n-\tcfg-\u003eremote_vni = vxlan-\u003edefault_dst.remote_vni;\n-\tcfg-\u003eremote_port = vxlan-\u003ecfg.dst_port;\n+\tcfg-\u003eremote_vni = vcfg-\u003evni;\n+\tcfg-\u003eremote_port = vcfg-\u003edst_port;\n \n \tif (entry-\u003eifindex != dev-\u003eifindex) {\n \t\tNL_SET_ERR_MSG_MOD(extack, \"Port net device must be the VXLAN net device\");\n@@ -957,11 +961,12 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,\n \t\t\t    const struct vxlan_mdb_entry *mdb_entry,\n \t\t\t    const struct vxlan_mdb_remote *remote)\n {\n+\tconst struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan-\u003ecfg);\n \tconst struct vxlan_mdb_entry_key *group = \u0026mdb_entry-\u003ekey;\n \tstruct vxlan_rdst *rd = rtnl_dereference(remote-\u003erd);\n \tsize_t nlmsg_size;\n \n-\t\t     /* MDBA_MDB_ENTRY_INFO */\n+\t/* MDBA_MDB_ENTRY_INFO */\n \tnlmsg_size = nla_total_size(sizeof(struct br_mdb_entry)) +\n \t\t     /* MDBA_MDB_EATTR_TIMER */\n \t\t     nla_total_size(sizeof(u32));\n@@ -978,16 +983,16 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,\n \t/* MDBA_MDB_EATTR_DST */\n \tnlmsg_size += nla_total_size(vxlan_addr_size(\u0026rd-\u003eremote_ip));\n \t/* MDBA_MDB_EATTR_DST_PORT */\n-\tif (rd-\u003eremote_port \u0026\u0026 rd-\u003eremote_port != vxlan-\u003ecfg.dst_port)\n+\tif (rd-\u003eremote_port \u0026\u0026 rd-\u003eremote_port != cfg-\u003edst_port)\n \t\tnlmsg_size += nla_total_size(sizeof(u16));\n \t/* MDBA_MDB_EATTR_VNI */\n-\tif (rd-\u003eremote_vni != vxlan-\u003edefault_dst.remote_vni)\n+\tif (rd-\u003eremote_vni != cfg-\u003evni)\n \t\tnlmsg_size += nla_total_size(sizeof(u32));\n \t/* MDBA_MDB_EATTR_IFINDEX */\n \tif (rd-\u003eremote_ifindex)\n \t\tnlmsg_size += nla_total_size(sizeof(u32));\n \t/* MDBA_MDB_EATTR_SRC_VNI */\n-\tif ((vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026 group-\u003evni)\n+\tif ((cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA) \u0026\u0026 group-\u003evni)\n \t\tnlmsg_size += nla_total_size(sizeof(u32));\n \n \treturn nlmsg_size;\n@@ -1216,7 +1221,7 @@ vxlan_mdb_entry_get(struct vxlan_dev *vxlan,\n \t\tgoto err_free_entry;\n \n \tif (hlist_is_singular_node(\u0026mdb_entry-\u003emdb_node, \u0026vxlan-\u003emdb_list))\n-\t\tvxlan-\u003ecfg.flags |= VXLAN_F_MDB;\n+\t\tset_bit(VXLAN_DEV_F_MDB, \u0026vxlan-\u003eflags);\n \n \treturn mdb_entry;\n \n@@ -1233,7 +1238,7 @@ static void vxlan_mdb_entry_put(struct vxlan_dev *vxlan,\n \t\treturn;\n \n \tif (hlist_is_singular_node(\u0026mdb_entry-\u003emdb_node, \u0026vxlan-\u003emdb_list))\n-\t\tvxlan-\u003ecfg.flags \u0026= ~VXLAN_F_MDB;\n+\t\tclear_bit(VXLAN_DEV_F_MDB, \u0026vxlan-\u003eflags);\n \n \trhashtable_remove_fast(\u0026vxlan-\u003emdb_tbl, \u0026mdb_entry-\u003erhnode,\n \t\t\t       vxlan_mdb_rht_params);\n@@ -1483,11 +1488,13 @@ static int vxlan_mdb_get_parse(struct net_device *dev, struct nlattr *tb[],\n {\n \tstruct br_mdb_entry *entry = nla_data(tb[MDBA_GET_ENTRY]);\n \tstruct nlattr *mdbe_attrs[MDBE_ATTR_MAX + 1];\n+\tconst struct vxlan_config *cfg;\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tint err;\n \n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \tmemset(group, 0, sizeof(*group));\n-\tgroup-\u003evni = vxlan-\u003edefault_dst.remote_vni;\n+\tgroup-\u003evni = cfg-\u003evni;\n \n \tif (!tb[MDBA_GET_ENTRY_ATTRS]) {\n \t\tvxlan_mdb_group_set(group, entry, NULL);\n@@ -1621,6 +1628,7 @@ int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,\n }\n \n struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,\n+\t\t\t\t\t\tconst struct vxlan_config *cfg,\n \t\t\t\t\t\tstruct sk_buff *skb,\n \t\t\t\t\t\t__be32 src_vni)\n {\n@@ -1634,8 +1642,8 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,\n \t/* When not in collect metadata mode, 'src_vni' is zero, but MDB\n \t * entries are stored with the VNI of the VXLAN device.\n \t */\n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_COLLECT_METADATA))\n-\t\tsrc_vni = vxlan-\u003edefault_dst.remote_vni;\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_COLLECT_METADATA))\n+\t\tsrc_vni = cfg-\u003evni;\n \n \tmemset(\u0026group, 0, sizeof(group));\n \tgroup.vni = src_vni;\n@@ -1700,6 +1708,7 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,\n }\n \n netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,\n+\t\t\t   const struct vxlan_config *cfg,\n \t\t\t   const struct vxlan_mdb_entry *mdb_entry,\n \t\t\t   struct sk_buff *skb)\n {\n@@ -1721,12 +1730,12 @@ netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,\n \n \t\tskb1 = skb_clone(skb, GFP_ATOMIC);\n \t\tif (skb1)\n-\t\t\tvxlan_xmit_one(skb1, vxlan-\u003edev, src_vni,\n+\t\t\tvxlan_xmit_one(skb1, vxlan-\u003edev, cfg, src_vni,\n \t\t\t\t       rcu_dereference(remote-\u003erd), false);\n \t}\n \n \tif (fremote)\n-\t\tvxlan_xmit_one(skb, vxlan-\u003edev, src_vni,\n+\t\tvxlan_xmit_one(skb, vxlan-\u003edev, cfg, src_vni,\n \t\t\t       rcu_dereference(fremote-\u003erd), false);\n \telse\n \t\tkfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);\n@@ -1757,7 +1766,7 @@ void vxlan_mdb_fini(struct vxlan_dev *vxlan)\n \tstruct vxlan_mdb_flush_desc desc = {};\n \n \tvxlan_mdb_flush(vxlan, \u0026desc);\n-\tWARN_ON_ONCE(vxlan-\u003ecfg.flags \u0026 VXLAN_F_MDB);\n+\tWARN_ON_ONCE(test_bit(VXLAN_DEV_F_MDB, \u0026vxlan-\u003eflags));\n \trhashtable_free_and_destroy(\u0026vxlan-\u003emdb_tbl, vxlan_mdb_check_empty,\n \t\t\t\t    NULL);\n }\ndiff --git a/drivers/net/vxlan/vxlan_multicast.c b/drivers/net/vxlan/vxlan_multicast.c\nindex 3b75b48dc726d..a5ef6f66f00d4 100644\n--- a/drivers/net/vxlan/vxlan_multicast.c\n+++ b/drivers/net/vxlan/vxlan_multicast.c\n@@ -14,11 +14,12 @@\n /* Update multicast group membership when first VNI on\n  * multicast address is brought up\n  */\n-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,\n \t\t    int rifindex)\n {\n-\tunion vxlan_addr *ip = (rip ? : \u0026vxlan-\u003edefault_dst.remote_ip);\n-\tint ifindex = (rifindex ? : vxlan-\u003edefault_dst.remote_ifindex);\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tconst union vxlan_addr *ip = (rip ? : \u0026cfg-\u003eremote_ip);\n+\tint ifindex = (rifindex ? : cfg-\u003eremote_ifindex);\n \tint ret = -EINVAL;\n \tstruct sock *sk;\n \n@@ -47,11 +48,12 @@ int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n \treturn ret;\n }\n \n-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,\n \t\t     int rifindex)\n {\n-\tunion vxlan_addr *ip = (rip ? : \u0026vxlan-\u003edefault_dst.remote_ip);\n-\tint ifindex = (rifindex ? : vxlan-\u003edefault_dst.remote_ifindex);\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tconst union vxlan_addr *ip = (rip ? : \u0026cfg-\u003eremote_ip);\n+\tint ifindex = (rifindex ? : cfg-\u003eremote_ifindex);\n \tint ret = -EINVAL;\n \tstruct sock *sk;\n \n@@ -80,8 +82,8 @@ int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n \treturn ret;\n }\n \n-static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,\n-\t\t\t\t   union vxlan_addr *rip, int rifindex)\n+static bool vxlan_group_used_match(const union vxlan_addr *ip, int ifindex,\n+\t\t\t\t   const union vxlan_addr *rip, int rifindex)\n {\n \tif (!vxlan_addr_multicast(rip))\n \t\treturn false;\n@@ -96,14 +98,16 @@ static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,\n }\n \n static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,\n-\t\t\t\t\t  union vxlan_addr *ip, int ifindex)\n+\t\t\t\t\t  const struct vxlan_config *cfg,\n+\t\t\t\t\t  const union vxlan_addr *ip,\n+\t\t\t\t\t  int ifindex)\n {\n \tstruct vxlan_vni_group *vg = rtnl_dereference(vxlan-\u003evnigrp);\n \tstruct vxlan_vni_node *v, *tmp;\n \n \tif (vxlan_group_used_match(ip, ifindex,\n-\t\t\t\t   \u0026vxlan-\u003edefault_dst.remote_ip,\n-\t\t\t\t   vxlan-\u003edefault_dst.remote_ifindex))\n+\t\t\t\t   \u0026cfg-\u003eremote_ip,\n+\t\t\t\t   cfg-\u003eremote_ifindex))\n \t\treturn true;\n \n \tlist_for_each_entry_safe(v, tmp, \u0026vg-\u003evni_list, vlist) {\n@@ -112,7 +116,7 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,\n \n \t\tif (vxlan_group_used_match(ip, ifindex,\n \t\t\t\t\t   \u0026v-\u003eremote_ip,\n-\t\t\t\t\t   vxlan-\u003edefault_dst.remote_ifindex))\n+\t\t\t\t\t   cfg-\u003eremote_ifindex))\n \t\t\treturn true;\n \t}\n \n@@ -121,16 +125,17 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,\n \n /* See if multicast group is already in use by other ID */\n bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,\n-\t\t      __be32 vni, union vxlan_addr *rip, int rifindex)\n+\t\t      __be32 vni, const union vxlan_addr *rip, int rifindex)\n {\n-\tunion vxlan_addr *ip = (rip ? : \u0026dev-\u003edefault_dst.remote_ip);\n-\tint ifindex = (rifindex ? : dev-\u003edefault_dst.remote_ifindex);\n+\tconst struct vxlan_config *dev_cfg = rtnl_dereference(dev-\u003ecfg);\n+\tconst union vxlan_addr *ip = (rip ? : \u0026dev_cfg-\u003eremote_ip);\n+\tint ifindex = (rifindex ? : dev_cfg-\u003eremote_ifindex);\n \tstruct vxlan_dev *vxlan;\n \tstruct vxlan_sock *sock4;\n #if IS_ENABLED(CONFIG_IPV6)\n \tstruct vxlan_sock *sock6;\n #endif\n-\tunsigned short family = dev-\u003edefault_dst.remote_ip.sa.sa_family;\n+\tunsigned short family = dev_cfg-\u003eremote_ip.sa.sa_family;\n \n \tsock4 = rtnl_dereference(dev-\u003evn4_sock);\n \n@@ -147,6 +152,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,\n #endif\n \n \tlist_for_each_entry(vxlan, \u0026vn-\u003evxlan_list, next) {\n+\t\tconst struct vxlan_config *cfg;\n+\n \t\tif (!netif_running(vxlan-\u003edev) || vxlan == dev)\n \t\t\tcontinue;\n \n@@ -158,13 +165,15 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,\n \t\t    rtnl_dereference(vxlan-\u003evn6_sock) != sock6)\n \t\t\tcontinue;\n #endif\n-\t\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER) {\n-\t\t\tif (!vxlan_group_used_by_vnifilter(vxlan, ip, ifindex))\n+\t\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\n+\t\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER) {\n+\t\t\tif (!vxlan_group_used_by_vnifilter(vxlan, cfg, ip, ifindex))\n \t\t\t\tcontinue;\n \t\t} else {\n \t\t\tif (!vxlan_group_used_match(ip, ifindex,\n-\t\t\t\t\t\t    \u0026vxlan-\u003edefault_dst.remote_ip,\n-\t\t\t\t\t\t    vxlan-\u003edefault_dst.remote_ifindex))\n+\t\t\t\t\t\t    \u0026cfg-\u003eremote_ip,\n+\t\t\t\t\t\t    cfg-\u003eremote_ifindex))\n \t\t\t\tcontinue;\n \t\t}\n \n@@ -174,7 +183,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,\n \treturn false;\n }\n \n-static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)\n+static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan,\n+\t\t\t\t       const struct vxlan_config *cfg)\n {\n \tstruct vxlan_vni_group *vg = rtnl_dereference(vxlan-\u003evnigrp);\n \tstruct vxlan_vni_node *v, *tmp, *vgood = NULL;\n@@ -185,7 +195,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)\n \t\t\tcontinue;\n \t\t/* skip if address is same as default address */\n \t\tif (vxlan_addr_equal(\u0026v-\u003eremote_ip,\n-\t\t\t\t     \u0026vxlan-\u003edefault_dst.remote_ip))\n+\t\t\t\t     \u0026cfg-\u003eremote_ip))\n \t\t\tcontinue;\n \t\tret = vxlan_igmp_join(vxlan, \u0026v-\u003eremote_ip, 0);\n \t\tif (ret == -EADDRINUSE)\n@@ -200,7 +210,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)\n \t\t\tif (!vxlan_addr_multicast(\u0026v-\u003eremote_ip))\n \t\t\t\tcontinue;\n \t\t\tif (vxlan_addr_equal(\u0026v-\u003eremote_ip,\n-\t\t\t\t\t     \u0026vxlan-\u003edefault_dst.remote_ip))\n+\t\t\t\t\t     \u0026cfg-\u003eremote_ip))\n \t\t\t\tcontinue;\n \t\t\tvxlan_igmp_leave(vxlan, \u0026v-\u003eremote_ip, 0);\n \t\t\tif (v == vgood)\n@@ -211,7 +221,8 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)\n \treturn ret;\n }\n \n-static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)\n+static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan,\n+\t\t\t\t\tconst struct vxlan_config *cfg)\n {\n \tstruct vxlan_net *vn = net_generic(vxlan-\u003enet, vxlan_net_id);\n \tstruct vxlan_vni_group *vg = rtnl_dereference(vxlan-\u003evnigrp);\n@@ -219,10 +230,17 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)\n \tint last_err = 0, ret;\n \n \tlist_for_each_entry_safe(v, tmp, \u0026vg-\u003evni_list, vlist) {\n-\t\tif (vxlan_addr_multicast(\u0026v-\u003eremote_ip) \u0026\u0026\n-\t\t    !vxlan_group_used(vn, vxlan, v-\u003evni, \u0026v-\u003eremote_ip,\n+\t\tif (!vxlan_addr_multicast(\u0026v-\u003eremote_ip))\n+\t\t\tcontinue;\n+\t\t/* skip if address is same as default address */\n+\t\tif (vxlan_addr_equal(\u0026v-\u003eremote_ip,\n+\t\t\t\t     \u0026cfg-\u003eremote_ip))\n+\t\t\tcontinue;\n+\t\tif (!vxlan_group_used(vn, vxlan, v-\u003evni, \u0026v-\u003eremote_ip,\n \t\t\t\t      0)) {\n \t\t\tret = vxlan_igmp_leave(vxlan, \u0026v-\u003eremote_ip, 0);\n+\t\t\tif (ret == -EADDRNOTAVAIL)\n+\t\t\t\tret = 0;\n \t\t\tif (ret)\n \t\t\t\tlast_err = ret;\n \t\t}\n@@ -233,38 +251,42 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)\n \n int vxlan_multicast_join(struct vxlan_dev *vxlan)\n {\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tint ret = 0;\n \n-\tif (vxlan_addr_multicast(\u0026vxlan-\u003edefault_dst.remote_ip)) {\n-\t\tret = vxlan_igmp_join(vxlan, \u0026vxlan-\u003edefault_dst.remote_ip,\n-\t\t\t\t      vxlan-\u003edefault_dst.remote_ifindex);\n+\tif (vxlan_addr_multicast(\u0026cfg-\u003eremote_ip)) {\n+\t\tret = vxlan_igmp_join(vxlan, \u0026cfg-\u003eremote_ip,\n+\t\t\t\t      cfg-\u003eremote_ifindex);\n \t\tif (ret == -EADDRINUSE)\n \t\t\tret = 0;\n \t\tif (ret)\n \t\t\treturn ret;\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)\n-\t\treturn vxlan_multicast_join_vnigrp(vxlan);\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)\n+\t\treturn vxlan_multicast_join_vnigrp(vxlan, cfg);\n \n \treturn 0;\n }\n \n int vxlan_multicast_leave(struct vxlan_dev *vxlan)\n {\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tstruct vxlan_net *vn = net_generic(vxlan-\u003enet, vxlan_net_id);\n \tint ret = 0;\n \n-\tif (vxlan_addr_multicast(\u0026vxlan-\u003edefault_dst.remote_ip) \u0026\u0026\n+\tif (vxlan_addr_multicast(\u0026cfg-\u003eremote_ip) \u0026\u0026\n \t    !vxlan_group_used(vn, vxlan, 0, NULL, 0)) {\n-\t\tret = vxlan_igmp_leave(vxlan, \u0026vxlan-\u003edefault_dst.remote_ip,\n-\t\t\t\t       vxlan-\u003edefault_dst.remote_ifindex);\n+\t\tret = vxlan_igmp_leave(vxlan, \u0026cfg-\u003eremote_ip,\n+\t\t\t\t       cfg-\u003eremote_ifindex);\n+\t\tif (ret == -EADDRNOTAVAIL)\n+\t\t\tret = 0;\n \t\tif (ret)\n \t\t\treturn ret;\n \t}\n \n-\tif (vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER)\n-\t\treturn vxlan_multicast_leave_vnigrp(vxlan);\n+\tif (cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)\n+\t\treturn vxlan_multicast_leave_vnigrp(vxlan, cfg);\n \n \treturn 0;\n }\ndiff --git a/drivers/net/vxlan/vxlan_private.h b/drivers/net/vxlan/vxlan_private.h\nindex b1eec22163608..f0ee85f83732a 100644\n--- a/drivers/net/vxlan/vxlan_private.h\n+++ b/drivers/net/vxlan/vxlan_private.h\n@@ -22,6 +22,8 @@ struct vxlan_net {\n \t/* sock_list is protected by rtnl lock */\n \tstruct hlist_head sock_list[PORT_HASH_SIZE];\n \tstruct notifier_block nexthop_notifier_block;\n+\t/* Generation counter for RTM_GETTUNNEL dumps */\n+\tatomic_t vnifilter_seq;\n };\n \n struct vxlan_fdb_key {\n@@ -177,6 +179,13 @@ vxlan_vnifilter_lookup(struct vxlan_dev *vxlan, __be32 vni)\n \t\t\t\t      vxlan_vni_rht_params);\n }\n \n+static inline void vxlan_vnifilter_seq_inc(const struct net *net)\n+{\n+\tstruct vxlan_net *vn = net_generic(net, vxlan_net_id);\n+\n+\tatomic_inc(\u0026vn-\u003evnifilter_seq);\n+}\n+\n /* vxlan_core.c */\n int vxlan_fdb_create(struct vxlan_dev *vxlan,\n \t\t     const u8 *mac, union vxlan_addr *ip,\n@@ -195,9 +204,10 @@ int vxlan_fdb_update(struct vxlan_dev *vxlan,\n \t\t     __u32 ifindex, __u16 ndm_flags, u32 nhid,\n \t\t     bool swdev_notify, struct netlink_ext_ack *extack);\n void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,\n+\t\t    const struct vxlan_config *cfg,\n \t\t    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc);\n int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,\n-\t\t     struct vxlan_config *conf, __be32 vni);\n+\t\t     const struct vxlan_config *conf, __be32 vni);\n \n /* vxlan_vnifilter.c */\n int vxlan_vnigroup_init(struct vxlan_dev *vxlan);\n@@ -205,7 +215,8 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan);\n \n int vxlan_vnifilter_init(void);\n void vxlan_vnifilter_uninit(void);\n-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,\n+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,\n+\t\t\t   const struct vxlan_config *cfg, __be32 vni,\n \t\t\t   struct vxlan_vni_node *vninode,\n \t\t\t   int type, unsigned int len);\n \n@@ -213,9 +224,16 @@ void vxlan_vs_add_vnigrp(struct vxlan_dev *vxlan,\n \t\t\t struct vxlan_sock *vs,\n \t\t\t bool ipv6);\n void vxlan_vs_del_vnigrp(struct vxlan_dev *vxlan);\n+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan);\n+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,\n+\t\t\t\t   const union vxlan_addr *old_remote_ip,\n+\t\t\t\t   const union vxlan_addr *remote_ip,\n+\t\t\t\t   u32 old_ifindex, u32 new_ifindex,\n+\t\t\t\t   struct netlink_ext_ack *extack);\n int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,\n-\t\t\t       union vxlan_addr *old_remote_ip,\n-\t\t\t       union vxlan_addr *new_remote_ip,\n+\t\t\t       const union vxlan_addr *old_remote_ip,\n+\t\t\t       const union vxlan_addr *new_remote_ip,\n+\t\t\t       u32 old_ifindex, u32 new_ifindex,\n \t\t\t       struct netlink_ext_ack *extack);\n \n \n@@ -223,10 +241,10 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,\n int vxlan_multicast_join(struct vxlan_dev *vxlan);\n int vxlan_multicast_leave(struct vxlan_dev *vxlan);\n bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,\n-\t\t      __be32 vni, union vxlan_addr *rip, int rifindex);\n-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n+\t\t      __be32 vni, const union vxlan_addr *rip, int rifindex);\n+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,\n \t\t    int rifindex);\n-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,\n+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,\n \t\t     int rifindex);\n \n /* vxlan_mdb.c */\n@@ -241,9 +259,11 @@ int vxlan_mdb_del_bulk(struct net_device *dev, struct nlattr *tb[],\n int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,\n \t\t  u32 seq, struct netlink_ext_ack *extack);\n struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,\n+\t\t\t\t\t\tconst struct vxlan_config *cfg,\n \t\t\t\t\t\tstruct sk_buff *skb,\n \t\t\t\t\t\t__be32 src_vni);\n netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,\n+\t\t\t   const struct vxlan_config *cfg,\n \t\t\t   const struct vxlan_mdb_entry *mdb_entry,\n \t\t\t   struct sk_buff *skb);\n int vxlan_mdb_init(struct vxlan_dev *vxlan);\ndiff --git a/drivers/net/vxlan/vxlan_vnifilter.c b/drivers/net/vxlan/vxlan_vnifilter.c\nindex dd94085e08865..41095096b698d 100644\n--- a/drivers/net/vxlan/vxlan_vnifilter.c\n+++ b/drivers/net/vxlan/vxlan_vnifilter.c\n@@ -171,13 +171,14 @@ static void vxlan_vnifilter_stats_add(struct vxlan_vni_node *vninode,\n \tu64_stats_update_end(\u0026pstats-\u003esyncp);\n }\n \n-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,\n+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,\n+\t\t\t   const struct vxlan_config *cfg, __be32 vni,\n \t\t\t   struct vxlan_vni_node *vninode,\n \t\t\t   int type, unsigned int len)\n {\n \tstruct vxlan_vni_node *vnode;\n \n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER))\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER))\n \t\treturn;\n \n \tif (vninode) {\n@@ -333,22 +334,36 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,\n \t\t\t\t    struct sk_buff *skb,\n \t\t\t\t    struct netlink_callback *cb)\n {\n-\tstruct vxlan_vni_node *tmp, *v, *vbegin = NULL, *vend = NULL;\n+\tstruct vxlan_vni_node *v, *vbegin = NULL, *vend = NULL;\n \tstruct vxlan_dev *vxlan = netdev_priv(dev);\n \tstruct tunnel_msg *new_tmsg, *tmsg;\n-\tint idx = 0, s_idx = cb-\u003eargs[1];\n+\tconst struct vxlan_config *cfg;\n \tstruct vxlan_vni_group *vg;\n \tstruct nlmsghdr *nlh;\n+\tint idx = 0, s_idx;\n \tbool dump_stats;\n \tint err = 0;\n \n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER))\n+\tif (cb-\u003eargs[2] != dev-\u003eifindex) {\n+\t\tcb-\u003eargs[1] = 0;\n+\t\tcb-\u003eargs[2] = dev-\u003eifindex;\n+\t}\n+\ts_idx = cb-\u003eargs[1];\n+\n+\tcfg = rcu_dereference(vxlan-\u003ecfg);\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER)) {\n+\t\tcb-\u003eargs[1] = 0;\n+\t\tcb-\u003eargs[2] = 0;\n \t\treturn -EINVAL;\n+\t}\n \n \t/* RCU needed because of the vni locking rules (rcu || rtnl) */\n \tvg = rcu_dereference(vxlan-\u003evnigrp);\n-\tif (!vg || !vg-\u003enum_vnis)\n+\tif (!vg || !vg-\u003enum_vnis) {\n+\t\tcb-\u003eargs[1] = 0;\n+\t\tcb-\u003eargs[2] = 0;\n \t\treturn 0;\n+\t}\n \n \ttmsg = nlmsg_data(cb-\u003enlh);\n \tdump_stats = !!(tmsg-\u003eflags \u0026 TUNNEL_MSG_FLAG_STATS);\n@@ -362,7 +377,7 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,\n \tnew_tmsg-\u003efamily = PF_BRIDGE;\n \tnew_tmsg-\u003eifindex = dev-\u003eifindex;\n \n-\tlist_for_each_entry_safe(v, tmp, \u0026vg-\u003evni_list, vlist) {\n+\tlist_for_each_entry_rcu(v, \u0026vg-\u003evni_list, vlist) {\n \t\tif (idx \u003c s_idx) {\n \t\t\tidx++;\n \t\t\tcontinue;\n@@ -394,12 +409,26 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,\n \t}\n \n \tcb-\u003eargs[1] = err ? idx : 0;\n+\tcb-\u003eargs[2] = err ? dev-\u003eifindex : 0;\n \n \tnlmsg_end(skb, nlh);\n \n+\tnl_dump_check_consistent(cb, nlh);\n+\n \treturn err;\n }\n \n+static u32 vxlan_vnifilter_base_seq(const struct net *net)\n+{\n+\tconst struct vxlan_net *vn = net_generic(net, vxlan_net_id);\n+\tu32 res = atomic_read(\u0026vn-\u003evnifilter_seq);\n+\n+\t/* Must not return 0 (see nl_dump_check_consistent()) */\n+\tif (!res)\n+\t\tres = 0x80000000;\n+\treturn res;\n+}\n+\n static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb)\n {\n \tint idx = 0, err = 0, s_idx = cb-\u003eargs[0];\n@@ -419,6 +448,9 @@ static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb\n \t}\n \n \trcu_read_lock();\n+\n+\tcb-\u003eseq = vxlan_vnifilter_base_seq(net);\n+\n \tif (tmsg-\u003eifindex) {\n \t\tdev = dev_get_by_index_rcu(net, tmsg-\u003eifindex);\n \t\tif (!dev) {\n@@ -470,26 +502,50 @@ static const struct nla_policy vni_filter_policy[VXLAN_VNIFILTER_MAX + 1] = {\n \t[VXLAN_VNIFILTER_ENTRY] = { .type = NLA_NESTED },\n };\n \n-static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,\n-\t\t\t\t\t  union vxlan_addr *old_remote_ip,\n-\t\t\t\t\t  union vxlan_addr *remote_ip,\n-\t\t\t\t\t  struct netlink_ext_ack *extack)\n+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan)\n {\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n+\tstruct vxlan_vni_group *vg = rtnl_dereference(vxlan-\u003evnigrp);\n+\tstruct vxlan_vni_node *v;\n+\n+\tif (!vg)\n+\t\treturn false;\n+\n+\tlist_for_each_entry(v, \u0026vg-\u003evni_list, vlist) {\n+\t\tif (vxlan_addr_multicast(\u0026v-\u003eremote_ip))\n+\t\t\treturn true;\n+\t}\n+\n+\treturn false;\n+}\n+\n+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,\n+\t\t\t\t   const union vxlan_addr *old_remote_ip,\n+\t\t\t\t   const union vxlan_addr *remote_ip,\n+\t\t\t\t   u32 old_ifindex, u32 new_ifindex,\n+\t\t\t\t   struct netlink_ext_ack *extack)\n+{\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tint err = 0;\n \n+\tif (old_remote_ip \u0026\u0026 remote_ip \u0026\u0026\n+\t    vxlan_addr_equal(old_remote_ip, remote_ip) \u0026\u0026\n+\t    old_ifindex == new_ifindex)\n+\t\treturn 0;\n+\n \tspin_lock_bh(\u0026vxlan-\u003ehash_lock);\n \tif (remote_ip \u0026\u0026 !vxlan_addr_any(remote_ip)) {\n+\t\tunion vxlan_addr rip = *remote_ip;\n+\n \t\terr = vxlan_fdb_update(vxlan, all_zeros_mac,\n-\t\t\t\t       remote_ip,\n+\t\t\t\t       \u0026rip,\n \t\t\t\t       NUD_REACHABLE | NUD_PERMANENT,\n \t\t\t\t       NLM_F_APPEND | NLM_F_CREATE,\n-\t\t\t\t       vxlan-\u003ecfg.dst_port,\n+\t\t\t\t       cfg-\u003edst_port,\n \t\t\t\t       vni,\n \t\t\t\t       vni,\n-\t\t\t\t       dst-\u003eremote_ifindex,\n+\t\t\t\t       new_ifindex,\n \t\t\t\t       NTF_SELF, 0, true, extack);\n-\t\tif (err) {\n+\t\tif (err \u0026\u0026 extack) {\n \t\t\tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n \t\t\treturn err;\n \t\t}\n@@ -498,9 +554,9 @@ static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,\n \tif (old_remote_ip \u0026\u0026 !vxlan_addr_any(old_remote_ip)) {\n \t\t__vxlan_fdb_delete(vxlan, all_zeros_mac,\n \t\t\t\t   *old_remote_ip,\n-\t\t\t\t   vxlan-\u003ecfg.dst_port,\n+\t\t\t\t   cfg-\u003edst_port,\n \t\t\t\t   vni, vni,\n-\t\t\t\t   dst-\u003eremote_ifindex,\n+\t\t\t\t   old_ifindex,\n \t\t\t\t   true);\n \t}\n \tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n@@ -515,8 +571,8 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,\n \t\t\t\t  struct netlink_ext_ack *extack)\n {\n \tstruct vxlan_net *vn = net_generic(vxlan-\u003enet, vxlan_net_id);\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n-\tunion vxlan_addr *newrip = NULL, *oldrip = NULL;\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tconst union vxlan_addr *newrip = NULL, *oldrip = NULL;\n \tunion vxlan_addr old_remote_ip;\n \tint ret = 0;\n \n@@ -528,15 +584,16 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,\n \tif (group \u0026\u0026 !vxlan_addr_any(group)) {\n \t\tnewrip = group;\n \t} else {\n-\t\tif (!vxlan_addr_any(\u0026dst-\u003eremote_ip))\n-\t\t\tnewrip = \u0026dst-\u003eremote_ip;\n+\t\tif (!vxlan_addr_any(\u0026cfg-\u003eremote_ip))\n+\t\t\tnewrip = \u0026cfg-\u003eremote_ip;\n \t}\n \n-\t/* if old rip exists, and no newrip,\n-\t * explicitly delete old rip\n-\t */\n-\tif (!newrip \u0026\u0026 !vxlan_addr_any(\u0026old_remote_ip))\n-\t\toldrip = \u0026old_remote_ip;\n+\tif (!create) {\n+\t\tif (!vxlan_addr_any(\u0026old_remote_ip))\n+\t\t\toldrip = \u0026old_remote_ip;\n+\t\telse if (!vxlan_addr_any(\u0026cfg-\u003eremote_ip))\n+\t\t\toldrip = \u0026cfg-\u003eremote_ip;\n+\t}\n \n \tif (!newrip \u0026\u0026 !oldrip)\n \t\treturn 0;\n@@ -546,20 +603,27 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,\n \n \tret = vxlan_update_default_fdb_entry(vxlan, vninode-\u003evni,\n \t\t\t\t\t     oldrip, newrip,\n+\t\t\t\t\t     cfg-\u003eremote_ifindex,\n+\t\t\t\t\t     cfg-\u003eremote_ifindex,\n \t\t\t\t\t     extack);\n \tif (ret)\n \t\tgoto out;\n \n-\tif (group)\n+\tif (group) {\n \t\tmemcpy(\u0026vninode-\u003eremote_ip, group, sizeof(vninode-\u003eremote_ip));\n+\t\tif (!create)\n+\t\t\tvxlan_vnifilter_seq_inc(dev_net(vxlan-\u003edev));\n+\t}\n \n \tif (vxlan-\u003edev-\u003eflags \u0026 IFF_UP) {\n \t\tif (vxlan_addr_multicast(\u0026old_remote_ip) \u0026\u0026\n \t\t    !vxlan_group_used(vn, vxlan, vninode-\u003evni,\n \t\t\t\t      \u0026old_remote_ip,\n-\t\t\t\t      vxlan-\u003edefault_dst.remote_ifindex)) {\n+\t\t\t\t      cfg-\u003eremote_ifindex)) {\n \t\t\tret = vxlan_igmp_leave(vxlan, \u0026old_remote_ip,\n \t\t\t\t\t       0);\n+\t\t\tif (ret == -EADDRNOTAVAIL)\n+\t\t\t\tret = 0;\n \t\t\tif (ret)\n \t\t\t\tgoto out;\n \t\t}\n@@ -581,10 +645,12 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,\n }\n \n int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,\n-\t\t\t       union vxlan_addr *old_remote_ip,\n-\t\t\t       union vxlan_addr *new_remote_ip,\n+\t\t\t       const union vxlan_addr *old_remote_ip,\n+\t\t\t       const union vxlan_addr *new_remote_ip,\n+\t\t\t       u32 old_ifindex, u32 new_ifindex,\n \t\t\t       struct netlink_ext_ack *extack)\n {\n+\tconst union vxlan_addr *oldrip, *newrip;\n \tstruct list_head *headp, *hpos;\n \tstruct vxlan_vni_group *vg;\n \tstruct vxlan_vni_node *vent;\n@@ -595,37 +661,66 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,\n \theadp = \u0026vg-\u003evni_list;\n \tlist_for_each_prev(hpos, headp) {\n \t\tvent = list_entry(hpos, struct vxlan_vni_node, vlist);\n+\n \t\tif (vxlan_addr_any(\u0026vent-\u003eremote_ip)) {\n-\t\t\tret = vxlan_update_default_fdb_entry(vxlan, vent-\u003evni,\n-\t\t\t\t\t\t\t     old_remote_ip,\n-\t\t\t\t\t\t\t     new_remote_ip,\n-\t\t\t\t\t\t\t     extack);\n-\t\t\tif (ret)\n-\t\t\t\treturn ret;\n+\t\t\toldrip = old_remote_ip;\n+\t\t\tnewrip = new_remote_ip;\n+\t\t} else {\n+\t\t\t/* A vni with its own group keeps it, but its fdb entry\n+\t\t\t * is still keyed on the device remote_ifindex.\n+\t\t\t */\n+\t\t\toldrip = \u0026vent-\u003eremote_ip;\n+\t\t\tnewrip = \u0026vent-\u003eremote_ip;\n \t\t}\n+\n+\t\tret = vxlan_update_default_fdb_entry(vxlan, vent-\u003evni,\n+\t\t\t\t\t\t     oldrip, newrip,\n+\t\t\t\t\t\t     old_ifindex, new_ifindex,\n+\t\t\t\t\t\t     extack);\n+\t\tif (ret)\n+\t\t\tgoto err_unwind;\n \t}\n \n \treturn 0;\n+\n+err_unwind:\n+\tlist_for_each_continue(hpos, headp) {\n+\t\tvent = list_entry(hpos, struct vxlan_vni_node, vlist);\n+\n+\t\tif (vxlan_addr_any(\u0026vent-\u003eremote_ip)) {\n+\t\t\toldrip = old_remote_ip;\n+\t\t\tnewrip = new_remote_ip;\n+\t\t} else {\n+\t\t\toldrip = \u0026vent-\u003eremote_ip;\n+\t\t\tnewrip = \u0026vent-\u003eremote_ip;\n+\t\t}\n+\n+\t\tvxlan_update_default_fdb_entry(vxlan, vent-\u003evni,\n+\t\t\t\t\t       newrip, oldrip,\n+\t\t\t\t\t       new_ifindex, old_ifindex,\n+\t\t\t\t\t       NULL);\n+\t}\n+\treturn ret;\n }\n \n static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,\n \t\t\t\t   struct vxlan_vni_node *vninode)\n {\n \tstruct vxlan_net *vn = net_generic(vxlan-\u003enet, vxlan_net_id);\n-\tstruct vxlan_rdst *dst = \u0026vxlan-\u003edefault_dst;\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \n \t/* if per vni remote_ip not present, delete the\n \t * default dst remote_ip previously added for this vni\n \t */\n \tif (!vxlan_addr_any(\u0026vninode-\u003eremote_ip) ||\n-\t    !vxlan_addr_any(\u0026dst-\u003eremote_ip)) {\n+\t    !vxlan_addr_any(\u0026cfg-\u003eremote_ip)) {\n \t\tspin_lock_bh(\u0026vxlan-\u003ehash_lock);\n \t\t__vxlan_fdb_delete(vxlan, all_zeros_mac,\n \t\t\t\t   (vxlan_addr_any(\u0026vninode-\u003eremote_ip) ?\n-\t\t\t\t   dst-\u003eremote_ip : vninode-\u003eremote_ip),\n-\t\t\t\t   vxlan-\u003ecfg.dst_port,\n+\t\t\t\t   cfg-\u003eremote_ip : vninode-\u003eremote_ip),\n+\t\t\t\t   cfg-\u003edst_port,\n \t\t\t\t   vninode-\u003evni, vninode-\u003evni,\n-\t\t\t\t   dst-\u003eremote_ifindex,\n+\t\t\t\t   cfg-\u003eremote_ifindex,\n \t\t\t\t   true);\n \t\tspin_unlock_bh(\u0026vxlan-\u003ehash_lock);\n \t}\n@@ -634,7 +729,7 @@ static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,\n \t\tif (vxlan_addr_multicast(\u0026vninode-\u003eremote_ip) \u0026\u0026\n \t\t    !vxlan_group_used(vn, vxlan, vninode-\u003evni,\n \t\t\t\t      \u0026vninode-\u003eremote_ip,\n-\t\t\t\t      dst-\u003eremote_ifindex)) {\n+\t\t\t\t      cfg-\u003eremote_ifindex)) {\n \t\t\tvxlan_igmp_leave(vxlan, \u0026vninode-\u003eremote_ip, 0);\n \t\t}\n \t}\n@@ -665,7 +760,8 @@ static int vxlan_vni_update(struct vxlan_dev *vxlan,\n \treturn 0;\n }\n \n-static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,\n+static void __vxlan_vni_add_list(struct vxlan_dev *vxlan,\n+\t\t\t\t struct vxlan_vni_group *vg,\n \t\t\t\t struct vxlan_vni_node *v)\n {\n \tstruct list_head *headp, *hpos;\n@@ -681,13 +777,16 @@ static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,\n \t}\n \tlist_add_rcu(\u0026v-\u003evlist, hpos);\n \tvg-\u003enum_vnis++;\n+\tvxlan_vnifilter_seq_inc(dev_net(vxlan-\u003edev));\n }\n \n-static void __vxlan_vni_del_list(struct vxlan_vni_group *vg,\n+static void __vxlan_vni_del_list(struct vxlan_dev *vxlan,\n+\t\t\t\t struct vxlan_vni_group *vg,\n \t\t\t\t struct vxlan_vni_node *v)\n {\n \tlist_del_rcu(\u0026v-\u003evlist);\n \tvg-\u003enum_vnis--;\n+\tvxlan_vnifilter_seq_inc(dev_net(vxlan-\u003edev));\n }\n \n static struct vxlan_vni_node *vxlan_vni_alloc(struct vxlan_dev *vxlan,\n@@ -723,6 +822,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,\n \t\t\t u32 vni, union vxlan_addr *group,\n \t\t\t struct netlink_ext_ack *extack)\n {\n+\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n \tstruct vxlan_vni_node *vninode;\n \t__be32 v = cpu_to_be32(vni);\n \tbool changed = false;\n@@ -731,7 +831,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,\n \tif (vxlan_vnifilter_lookup(vxlan, v))\n \t\treturn vxlan_vni_update(vxlan, vg, v, group, \u0026changed, extack);\n \n-\terr = vxlan_vni_in_use(vxlan-\u003enet, vxlan, \u0026vxlan-\u003ecfg, v);\n+\terr = vxlan_vni_in_use(vxlan-\u003enet, vxlan, cfg, v);\n \tif (err) {\n \t\tNL_SET_ERR_MSG(extack, \"VNI in use\");\n \t\treturn err;\n@@ -749,7 +849,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,\n \t\treturn err;\n \t}\n \n-\t__vxlan_vni_add_list(vg, vninode);\n+\t__vxlan_vni_add_list(vxlan, vg, vninode);\n \n \tif (vxlan-\u003edev-\u003eflags \u0026 IFF_UP)\n \t\tvxlan_vs_add_del_vninode(vxlan, vninode, false);\n@@ -795,7 +895,7 @@ static int vxlan_vni_del(struct vxlan_dev *vxlan,\n \tif (err)\n \t\tgoto out;\n \n-\t__vxlan_vni_del_list(vg, vninode);\n+\t__vxlan_vni_del_list(vxlan, vg, vninode);\n \n \tvxlan_vnifilter_notify(vxlan, vninode, RTM_DELTUNNEL);\n \n@@ -844,6 +944,7 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,\n \t\t\t\t    int cmd, struct netlink_ext_ack *extack)\n {\n \tstruct nlattr *vattrs[VXLAN_VNIFILTER_ENTRY_MAX + 1];\n+\tconst struct vxlan_config *cfg;\n \tu32 vni_start = 0, vni_end = 0;\n \tunion vxlan_addr group;\n \tint err;\n@@ -881,7 +982,8 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,\n \t\tmemset(\u0026group, 0, sizeof(group));\n \t}\n \n-\tif (vxlan_addr_multicast(\u0026group) \u0026\u0026 !vxlan-\u003edefault_dst.remote_ifindex) {\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n+\tif (vxlan_addr_multicast(\u0026group) \u0026\u0026 !cfg-\u003eremote_ifindex) {\n \t\tNL_SET_ERR_MSG(extack,\n \t\t\t       \"Local interface required for multicast remote group\");\n \n@@ -909,7 +1011,7 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan)\n #if IS_ENABLED(CONFIG_IPV6)\n \t\thlist_del_init_rcu(\u0026v-\u003ehlist6.hlist);\n #endif\n-\t\t__vxlan_vni_del_list(vg, v);\n+\t\t__vxlan_vni_del_list(vxlan, vg, v);\n \t\tvxlan_vnifilter_notify(vxlan, v, RTM_DELTUNNEL);\n \t\tcall_rcu(\u0026v-\u003ercu, vxlan_vni_node_rcu_free);\n \t}\n@@ -940,6 +1042,7 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,\n \t\t\t\t   struct netlink_ext_ack *extack)\n {\n \tstruct net *net = sock_net(skb-\u003esk);\n+\tconst struct vxlan_config *cfg;\n \tstruct tunnel_msg *tmsg;\n \tstruct vxlan_dev *vxlan;\n \tstruct net_device *dev;\n@@ -964,8 +1067,9 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,\n \t}\n \n \tvxlan = netdev_priv(dev);\n+\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n \n-\tif (!(vxlan-\u003ecfg.flags \u0026 VXLAN_F_VNIFILTER))\n+\tif (!(cfg-\u003eflags \u0026 VXLAN_F_VNIFILTER))\n \t\treturn -EOPNOTSUPP;\n \n \tnlmsg_for_each_attr_type(attr, VXLAN_VNIFILTER_ENTRY, nlh,\ndiff --git a/include/net/vxlan.h b/include/net/vxlan.h\nindex 7b82075055237..c3c9f2ccc3d1b 100644\n--- a/include/net/vxlan.h\n+++ b/include/net/vxlan.h\n@@ -204,7 +204,6 @@ struct vxlan_rdst {\n \tu8\t\t\t offloaded:1;\n \t__be32\t\t\t remote_vni;\n \tu32\t\t\t remote_ifindex;\n-\tstruct net_device\t *remote_dev;\n \tstruct list_head\t list;\n \tstruct rcu_head\t\t rcu;\n \tstruct dst_cache\t dst_cache;\n@@ -229,6 +228,7 @@ struct vxlan_config {\n \tbool\t\t\t\tno_share;\n \tenum ifla_vxlan_df\t\tdf;\n \tstruct vxlanhdr\t\t\treserved_bits;\n+\tstruct rcu_head\t\t\trcu;\n };\n \n enum {\n@@ -294,14 +294,15 @@ struct vxlan_dev {\n #endif\n \tstruct net_device *dev;\n \tstruct net\t  *net;\t\t/* netns for packet i/o */\n-\tstruct vxlan_rdst default_dst;\t/* default destination */\n+\tstruct net_device *lowerdev;\n \n \tstruct timer_list age_timer;\n \tspinlock_t\t  hash_lock;\n \tunsigned int\t  addrcnt;\n \tstruct gro_cells  gro_cells;\n+\tunsigned long\t  flags;\n \n-\tstruct vxlan_config\tcfg;\n+\tstruct vxlan_config __rcu\t*cfg;\n \n \tstruct vxlan_vni_group  __rcu *vnigrp;\n \n@@ -313,6 +314,10 @@ struct vxlan_dev {\n \tunsigned int mdb_seq;\n };\n \n+enum vxlan_dev_flags {\n+\tVXLAN_DEV_F_MDB,\n+};\n+\n #define VXLAN_F_LEARN\t\t\t0x01\n #define VXLAN_F_PROXY\t\t\t0x02\n #define VXLAN_F_RSC\t\t\t0x04\n@@ -331,7 +336,6 @@ struct vxlan_dev {\n #define VXLAN_F_IPV6_LINKLOCAL\t\t0x8000\n #define VXLAN_F_TTL_INHERIT\t\t0x10000\n #define VXLAN_F_VNIFILTER               0x20000\n-#define VXLAN_F_MDB\t\t\t0x40000\n #define VXLAN_F_LOCALBYPASS\t\t0x80000\n #define VXLAN_F_MC_ROUTE\t\t0x100000\n \n"
}

2/1 2026/09/25 02:07 agent patch-evaluator
0m Model:
core

Args:
null

Results:
{
  "EnableConfigs": [
    "VXLAN",
    "INET",
    "IPV6"
  ],
  "FocusSymbols": [
    "vxlan_changelink",
    "vxlan_config_apply",
    "vxlan_dev_configure",
    "vxlan_xmit",
    "vxlan_rcv",
    "vxlan_vnilist_update_group",
    "vxlan_update_default_fdb_entry",
    "vxlan_vnifilter_dump",
    "vxlan_multicast_join",
    "vxlan_multicast_leave",
    "vxlan_free_dev"
  ],
  "Reasoning": "The patch refactors the VXLAN driver by converting the embedded vxlan_config in struct vxlan_dev to a dynamically allocated RCU-protected pointer (vxlan-\u003ecfg). It updates device configuration, changelink, transmission (vxlan_xmit), reception (vxlan_rcv), multicast, and VNI-filter operations to access configuration under RCU, and adds memory management (kmemdup, kfree_rcu, and priv_destructor vxlan_free_dev) and rollback handling. These reachable networking changes introduce subtle lifetime and concurrency considerations that warrant fuzzing via netlink and packet I/O.",
  "WorthFuzzing": true
}

Instruction:
You are an expert Linux kernel maintainer and security engineer.
Your job is to review a provided patch series and evaluate whether it warrants fuzzing with syzkaller.

IMPORTANT: The changes have ALREADY been applied and committed as the HEAD commit in
your workspace. Do NOT rely on internal assumptions. You must actively use your code access
tools to inspect the actual source code, callers, and surrounding context.

================================================================================
1. CORE TRIAGE PHILOSOPHY
================================================================================
The goal of patch fuzzing is to discover crashes, regressions, exposed latent bugs,
and newly triggered assertions introduced by the patch series.

- REACHABILITY IS THE PRIMARY GATE:
  Fuzzing can only discover bugs in code that can actually execute in standard virtualized
  environments (GCE or QEMU, utilizing software-emulated devices like USB gadgets, netdev, tun/tap).
  If the modified code is structurally unreachable (see Section 2), it MUST NOT be fuzzed,
  regardless of whether it adds assertions or complex logic.

- DO NOT BLINDLY TRUST "NO FUNCTIONAL CHANGE" (NFCI) OR "REFACTORING" CLAIMS:
  Patch authors routinely label changes as "cleanups", "refactorings", or state
  "No functional change intended". Do NOT take these claims at face value.
  Code refactorings that rearrange logic, introduce helper functions, or alter state management
  in core subsystems frequently introduce subtle semantic shifts or uncover latent kernel bugs.
  If reachable executable code is modified or refactored, it MUST be fuzzed.

- NEW OR MODIFIED ASSERTIONS IN REACHABLE CODE MUST BE FUZZED:
  When a patch introduces or modifies runtime checks or assertions (e.g., WARN_ON*, VM_WARN_ON*,
  BUG_ON*, lockdep_assert*) in reachable code paths, it enforces new or stricter invariants.
  Even if the author believes the invariant always holds, fuzzing is essential to verify whether
  an unusual sequence of operations can violate it.

================================================================================
2. WHEN TO RETURN WorthFuzzing=false (NEGATIVE CRITERIA)
================================================================================
Return WorthFuzzing=false ONLY IF all modified code falls strictly into one or more of these categories:

- Non-kernel and non-executable changes:
  * Modifications to Documentation/, comments, or spelling fixes.
  * User-space directories, self-tests, samples, or scripts (e.g., tools/, samples/, scripts/, usr/)
    that do not affect the compiled kernel image (vmlinux) or kernel modules.
  * Purely decorative logging (e.g., message strings in pr_err, printk, dev_info) or tracepoints
    that do not alter control flow or data structures.
  * Build system or Kconfig changes that do not alter compiled C logic.
- Structurally unreachable hardware:
  * Vendor-specific PCIe switches, SmartNICs, or GPU drivers (e.g., mlxsw, pds_core, qed,
    ionic, amdgpu) requiring physical ASIC/PCIe cards not emulated in standard QEMU.
- Unreachable execution paths:
  * Driver teardown callbacks (.remove, .shutdown, pci_unregister_driver) executed only during
    physical PCI hot-unplug or manual sysfs driver unbinding.
  * Code paths exclusive to architectures other than the target architecture.

================================================================================
3. WHEN TO RETURN WorthFuzzing=true (POSITIVE CRITERIA)
================================================================================
Return WorthFuzzing=true whenever the patch touches reachable executable code, including:
- Core Subsystems:
  * Any logic modifications in memory management (mm/), synchronization/locking (kernel/locking/),
    BPF, scheduler, core networking, VFS, or syscall handling.
- Refactorings and Code Cleanups:
  * Any restructuring of reachable data structures, helper abstractions, or algorithm flows.
- Runtime Assertions and Defensive Checks:
  * Any introduction or alteration of assertions (WARN_ON*, VM_WARN_ON*, BUG_ON*, etc.) in reachable paths.
- Reachable Drivers and Protocols:
  * Drivers accessible via virtual buses (virtio, USB gadget, loopback, netlink, binder, sockets, etc.).

================================================================================
4. EXTRACTING FocusSymbols (PREVENTING DILUTION)
================================================================================
When WorthFuzzing=true, you must extract specific kernel functions into FocusSymbols to guide the fuzzer:

- AVOID UBIQUITOUS LIFECYCLE HOT-PATHS:
  Do NOT list generic, ubiquitous functions called by almost every program in the corpus
  (including, but not limited to: general memory allocators and deallocators, page fault
  and trap handlers, or core synchronization primitives; this is not an exhaustive list).
  Listing ubiquitous functions causes the fuzzer to classify thousands of unrelated tests as "focused",
  which severely dilutes fuzzing effort away from the actual changes.

- TARGET SPECIFIC FEATURE LOGIC AND ENTRYPOINTS:
  List functions that specifically implement the logic being added or altered, or direct API entrypoints
  for the subsystem feature under review.

- HANDLING STATIC INLINE FUNCTIONS IN HEADERS (.h):
  Compiler-inlined static functions (such as static inlines in mm/*.h or include/linux/*.h) lack
  distinct symbol addresses in vmlinux and cannot be targeted directly by symbol coverage filters.
  If the changes are primarily in static inline helpers, identify non-static, feature-specific caller
  functions in .c files that exercise them (avoiding ubiquitous lifecycle wrappers).

================================================================================
5. IDENTIFYING EnableConfigs
================================================================================
Identify any specific CONFIG_ options required to properly compile and reach the modified code:
- Inspect Kconfig files and #ifdef guards; do not make assumptions.
- Check "depends on" lines in Kconfig to include any non-standard parent subsystem configs needed.
- Strip any 'CONFIG_' prefix (e.g., return "NET_IPV4" instead of "CONFIG_NET_IPV4").
Prefer calling several tools at the same time to save round-trips.


Use set-results tool to provide results of the analysis.
It must be called exactly once before the final reply.
Ignore results of this tool.

Prompt:
Target architecture: amd64

For your convenience, here is the diff of the changes:
commit c6e7846ceee7f17a881c3bd9be86fc5863645aba
Author: syz-cluster <triage@syzkaller.com>
Date:   Fri Sep 25 02:07:19 2026 +0000

    syz-cluster: applied patch under review

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
index 7a18a469961db..467fbe43b89e9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
@@ -241,9 +241,16 @@ static bool mlx5e_tc_tun_encap_info_equal_vxlan(struct mlx5e_encap_key *a,
 static int mlx5e_tc_tun_get_remote_ifindex(struct net_device *mirred_dev)
 {
 	const struct vxlan_dev *vxlan = netdev_priv(mirred_dev);
-	const struct vxlan_rdst *dst = &vxlan->default_dst;
+	const struct vxlan_config *cfg;
+	int ifindex = 0;
 
-	return dst->remote_ifindex;
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+	if (cfg)
+		ifindex = cfg->remote_ifindex;
+	rcu_read_unlock();
+
+	return ifindex;
 }
 
 struct mlx5e_tc_tunnel vxlan_tunnel = {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
index 52c2fe3644d4b..21e7b499f3879 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
@@ -59,8 +59,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,
 					   const struct mlxsw_sp_nve_params *params,
 					   struct netlink_ext_ack *extack)
 {
-	struct vxlan_dev *vxlan = netdev_priv(params->dev);
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
+	struct vxlan_dev *vxlan;
+
+	vxlan = netdev_priv(params->dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	if (vxlan_addr_multicast(&cfg->remote_ip)) {
 		NL_SET_ERR_MSG_MOD(extack, "VxLAN: Multicast destination IP is not supported");
@@ -92,6 +95,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,
 		return false;
 	}
 
+	if (test_bit(VXLAN_DEV_F_MDB, &vxlan->flags)) {
+		NL_SET_ERR_MSG_MOD(extack, "VxLAN: MDB entries are not supported");
+		return false;
+	}
+
 	switch (cfg->saddr.sa.sa_family) {
 	case AF_INET:
 		if (!mlxsw_sp_nve_vxlan_ipv4_flags_check(cfg, extack))
@@ -148,8 +156,11 @@ static void mlxsw_sp_nve_vxlan_config(const struct mlxsw_sp_nve *nve,
 				      const struct mlxsw_sp_nve_params *params,
 				      struct mlxsw_sp_nve_config *config)
 {
-	struct vxlan_dev *vxlan = netdev_priv(params->dev);
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
+	struct vxlan_dev *vxlan;
+
+	vxlan = netdev_priv(params->dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	config->type = MLXSW_SP_NVE_TYPE_VXLAN;
 	config->ttl = cfg->ttl;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index fe45e533a4b2e..de60b698bea98 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -2513,15 +2513,17 @@ mlxsw_sp_bridge_vlan_aware_vxlan_join(struct mlxsw_sp_bridge_device *bridge_devi
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device->dev);
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	struct mlxsw_sp_nve_params params = {
-		.type = MLXSW_SP_NVE_TYPE_VXLAN,
-		.vni = vxlan->cfg.vni,
-		.dev = vxlan_dev,
-		.ethertype = ethertype,
-	};
+	struct mlxsw_sp_nve_params params;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	params.type = MLXSW_SP_NVE_TYPE_VXLAN;
+	params.vni = cfg->vni;
+	params.dev = vxlan_dev;
+	params.ethertype = ethertype;
+
 	/* If the VLAN is 0, we need to find the VLAN that is configured as
 	 * PVID and egress untagged on the bridge port of the VxLAN device.
 	 * It is possible no such VLAN exists
@@ -2704,15 +2706,17 @@ mlxsw_sp_bridge_8021d_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device->dev);
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	struct mlxsw_sp_nve_params params = {
-		.type = MLXSW_SP_NVE_TYPE_VXLAN,
-		.vni = vxlan->cfg.vni,
-		.dev = vxlan_dev,
-		.ethertype = ETH_P_8021Q,
-	};
+	struct mlxsw_sp_nve_params params;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	params.type = MLXSW_SP_NVE_TYPE_VXLAN;
+	params.vni = cfg->vni;
+	params.dev = vxlan_dev;
+	params.ethertype = ETH_P_8021Q;
+
 	fid = mlxsw_sp_fid_8021d_get(mlxsw_sp, bridge_device->dev->ifindex);
 	if (IS_ERR(fid)) {
 		NL_SET_ERR_MSG_MOD(extack, "Failed to create 802.1D FID");
@@ -2933,10 +2937,13 @@ static void __mlxsw_sp_bridge_vxlan_leave(struct mlxsw_sp *mlxsw_sp,
 					  const struct net_device *vxlan_dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	/* If the VxLAN device is down, then the FID does not have a VNI */
-	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vxlan->cfg.vni);
+	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, cfg->vni);
 	if (!fid)
 		return;
 
@@ -3029,11 +3036,13 @@ static void mlxsw_sp_fdb_vxlan_call_notifiers(struct net_device *dev,
 	struct switchdev_notifier_vxlan_fdb_info info;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	enum switchdev_notifier_type type;
+	const struct vxlan_config *cfg;
 
+	cfg = rtnl_dereference(vxlan->cfg);
 	type = adding ? SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE :
 			SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE;
 	mlxsw_sp_switchdev_addr_vxlan_convert(proto, addr, &info.remote_ip);
-	info.remote_port = vxlan->cfg.dst_port;
+	info.remote_port = cfg->dst_port;
 	info.remote_vni = vni;
 	info.remote_ifindex = 0;
 	ether_addr_copy(info.eth_addr, mac);
@@ -3236,8 +3245,10 @@ __mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp *mlxsw_sp,
 
 	if (adding && netif_is_vxlan(dev)) {
 		struct vxlan_dev *vxlan = netdev_priv(dev);
+		const struct vxlan_config *cfg;
 
-		if (!(vxlan->cfg.flags & VXLAN_F_LEARN))
+		cfg = rtnl_dereference(vxlan->cfg);
+		if (!(cfg->flags & VXLAN_F_LEARN))
 			return -EINVAL;
 	}
 
@@ -3722,9 +3733,11 @@ mlxsw_sp_switchdev_vxlan_work_prepare(struct mlxsw_sp_switchdev_event_work *
 {
 	struct vxlan_dev *vxlan = netdev_priv(switchdev_work->dev);
 	struct switchdev_notifier_vxlan_fdb_info *vxlan_fdb_info;
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
 	struct netlink_ext_ack *extack;
 
+	cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	extack = switchdev_notifier_info_to_extack(info);
 	vxlan_fdb_info = container_of(info,
 				      struct switchdev_notifier_vxlan_fdb_info,
@@ -3851,11 +3864,15 @@ mlxsw_sp_switchdev_vxlan_vlan_add(struct mlxsw_sp *mlxsw_sp,
 				  struct netlink_ext_ack *extack)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	__be32 vni = vxlan->cfg.vni;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	u16 old_vid;
+	__be32 vni;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	/* We cannot have the same VLAN as PVID and egress untagged on multiple
 	 * VxLAN devices. Note that we get this notification before the VLAN is
 	 * actually added to the bridge's database, so it is not possible for
@@ -3935,12 +3952,16 @@ mlxsw_sp_switchdev_vxlan_vlan_del(struct mlxsw_sp *mlxsw_sp,
 				  const struct net_device *vxlan_dev, u16 vid)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	__be32 vni = vxlan->cfg.vni;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
+	__be32 vni;
 
 	if (!netif_running(vxlan_dev))
 		return;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vni);
 	if (!fid)
 		return;
diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 347245cc1de4e..0e5ab0055ff8c 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -110,20 +110,23 @@ static struct vxlan_dev *vxlan_vs_find_vni(struct vxlan_sock *vs,
 		vni = 0;
 
 	hlist_for_each_entry_rcu(node, vni_head(vs, vni), hlist) {
+		const struct vxlan_config *cfg;
+
 		if (!node->vxlan)
 			continue;
+
+		cfg = rcu_dereference(node->vxlan->cfg);
+
 		vnode = NULL;
-		if (node->vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
 			vnode = vxlan_vnifilter_lookup(node->vxlan, vni);
 			if (!vnode)
 				continue;
-		} else if (node->vxlan->default_dst.remote_vni != vni) {
+		} else if (cfg->vni != vni) {
 			continue;
 		}
 
 		if (IS_ENABLED(CONFIG_IPV6)) {
-			const struct vxlan_config *cfg = &node->vxlan->cfg;
-
 			if ((cfg->flags & VXLAN_F_IPV6_LINKLOCAL) &&
 			    cfg->remote_ifindex != ifindex)
 				continue;
@@ -157,6 +160,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			  u32 portid, u32 seq, int type, unsigned int flags,
 			  const struct vxlan_rdst *rdst)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	unsigned long now = jiffies;
 	struct nda_cacheinfo ci;
 	bool send_ip, send_eth;
@@ -216,10 +220,10 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			goto nla_put_failure;
 
 		if (rdst->remote_port &&
-		    rdst->remote_port != vxlan->cfg.dst_port &&
+		    rdst->remote_port != cfg->dst_port &&
 		    nla_put_be16(skb, NDA_PORT, rdst->remote_port))
 			goto nla_put_failure;
-		if (rdst->remote_vni != vxlan->default_dst.remote_vni &&
+		if (rdst->remote_vni != cfg->vni &&
 		    nla_put_u32(skb, NDA_VNI, be32_to_cpu(rdst->remote_vni)))
 			goto nla_put_failure;
 		if (rdst->remote_ifindex &&
@@ -227,7 +231,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			goto nla_put_failure;
 	}
 
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) && fdb->key.vni &&
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) && fdb->key.vni &&
 	    nla_put_u32(skb, NDA_SRC_VNI,
 			be32_to_cpu(fdb->key.vni)))
 		goto nla_put_failure;
@@ -377,14 +381,15 @@ static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])
 
 /* Look up Ethernet address in forwarding table */
 static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,
+					    const struct vxlan_config *cfg,
 					    const u8 *mac, __be32 vni)
 {
 	struct vxlan_fdb_key key;
 
 	memset(&key, 0, sizeof(key));
 	memcpy(key.eth_addr, mac, sizeof(key.eth_addr));
-	if (!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA))
-		key.vni = vxlan->default_dst.remote_vni;
+	if (!(cfg->flags & VXLAN_F_COLLECT_METADATA))
+		key.vni = cfg->vni;
 	else
 		key.vni = vni;
 
@@ -393,11 +398,12 @@ static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,
 }
 
 static struct vxlan_fdb *vxlan_find_mac_tx(struct vxlan_dev *vxlan,
+					   const struct vxlan_config *cfg,
 					   const u8 *mac, __be32 vni)
 {
 	struct vxlan_fdb *f;
 
-	f = vxlan_find_mac_rcu(vxlan, mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, mac, vni);
 	if (f) {
 		unsigned long now = jiffies;
 
@@ -416,7 +422,7 @@ static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan,
 	lockdep_assert_held_once(&vxlan->hash_lock);
 
 	rcu_read_lock();
-	f = vxlan_find_mac_rcu(vxlan, mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan->cfg), mac, vni);
 	rcu_read_unlock();
 
 	return f;
@@ -457,7 +463,7 @@ int vxlan_fdb_find_uc(struct net_device *dev, const u8 *mac, __be32 vni,
 
 	rcu_read_lock();
 
-	f = vxlan_find_mac_rcu(vxlan, eth_addr, vni);
+	f = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan->cfg), eth_addr, vni);
 	if (f)
 		rdst = first_remote_rcu(f);
 	if (!rdst) {
@@ -798,6 +804,7 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
 			       u32 nhid, struct netlink_ext_ack *extack)
 {
 	struct nexthop *old_nh = rtnl_dereference(fdb->nh);
+	const struct vxlan_config *cfg;
 	struct nexthop *nh;
 	int err = -EINVAL;
 
@@ -826,7 +833,8 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
 	}
 
 	/* check nexthop group family */
-	switch (vxlan->default_dst.remote_ip.sa.sa_family) {
+	cfg = rtnl_dereference(vxlan->cfg);
+	switch (cfg->remote_ip.sa.sa_family) {
 	case AF_INET:
 		if (!nexthop_has_v4(nh)) {
 			err = -EAFNOSUPPORT;
@@ -863,12 +871,13 @@ int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		     u32 nhid, struct vxlan_fdb **fdb,
 		     struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	struct vxlan_rdst *rd = NULL;
 	struct vxlan_fdb *f;
 	int rc;
 
-	if (vxlan->cfg.addrmax &&
-	    vxlan->addrcnt >= vxlan->cfg.addrmax)
+	if (cfg->addrmax &&
+	    vxlan->addrcnt >= cfg->addrmax)
 		return -ENOSPC;
 
 	netdev_dbg(vxlan->dev, "add %pM -> %pIS\n", mac, ip);
@@ -1154,6 +1163,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 			   __be32 *vni, u32 *ifindex, u32 *nhid,
 			   struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct net *net = dev_net(vxlan->dev);
 	int err;
 
@@ -1170,7 +1180,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 			return err;
 		}
 	} else {
-		union vxlan_addr *remote = &vxlan->default_dst.remote_ip;
+		const union vxlan_addr *remote = &cfg->remote_ip;
 
 		if (remote->sa.sa_family == AF_INET) {
 			ip->sin.sin_addr.s_addr = htonl(INADDR_ANY);
@@ -1190,7 +1200,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*port = nla_get_be16(tb[NDA_PORT]);
 	} else {
-		*port = vxlan->cfg.dst_port;
+		*port = cfg->dst_port;
 	}
 
 	if (tb[NDA_VNI]) {
@@ -1200,7 +1210,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*vni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));
 	} else {
-		*vni = vxlan->default_dst.remote_vni;
+		*vni = cfg->vni;
 	}
 
 	if (tb[NDA_SRC_VNI]) {
@@ -1210,7 +1220,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*src_vni = cpu_to_be32(nla_get_u32(tb[NDA_SRC_VNI]));
 	} else {
-		*src_vni = vxlan->default_dst.remote_vni;
+		*src_vni = cfg->vni;
 	}
 
 	if (tb[NDA_IFINDEX]) {
@@ -1241,6 +1251,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 			 const unsigned char *addr, u16 vid, u16 flags,
 			 bool *notified, struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	/* struct net *net = dev_net(vxlan->dev); */
 	union vxlan_addr ip;
@@ -1268,7 +1279,8 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 		return -EINVAL;
 	}
 
-	if (vxlan->default_dst.remote_ip.sa.sa_family != ip.sa.sa_family)
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (cfg->remote_ip.sa.sa_family != ip.sa.sa_family)
 		return -EAFNOSUPPORT;
 
 	spin_lock_bh(&vxlan->hash_lock);
@@ -1405,18 +1417,21 @@ static int vxlan_fdb_get(struct sk_buff *skb,
 			 struct netlink_ext_ack *extack)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	struct vxlan_fdb *f;
 	__be32 vni;
 	int err;
 
+	cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	if (tb[NDA_VNI])
 		vni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));
 	else
-		vni = vxlan->default_dst.remote_vni;
+		vni = cfg->vni;
 
 	rcu_read_lock();
 
-	f = vxlan_find_mac_rcu(vxlan, addr, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, addr, vni);
 	if (!f) {
 		NL_SET_ERR_MSG(extack, "Fdb entry not found");
 		err = -ENOENT;
@@ -1434,6 +1449,7 @@ static int vxlan_fdb_get(struct sk_buff *skb,
  * and Tunnel endpoint.
  */
 static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
+					const struct vxlan_config *cfg,
 					union vxlan_addr *src_ip,
 					const u8 *src_mac, u32 src_ifindex,
 					__be32 vni)
@@ -1452,7 +1468,7 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
 		ifindex = src_ifindex;
 #endif
 
-	f = vxlan_find_mac_rcu(vxlan, src_mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, src_mac, vni);
 	if (likely(f)) {
 		struct vxlan_rdst *rdst = first_remote_rcu(f);
 		unsigned long now = jiffies;
@@ -1488,9 +1504,9 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
 			vxlan_fdb_update(vxlan, src_mac, src_ip,
 					 NUD_REACHABLE,
 					 NLM_F_EXCL|NLM_F_CREATE,
-					 vxlan->cfg.dst_port,
+					 cfg->dst_port,
 					 vni,
-					 vxlan->default_dst.remote_vni,
+					 cfg->vni,
 					 ifindex, NTF_SELF, 0, true, NULL);
 		spin_unlock(&vxlan->hash_lock);
 	}
@@ -1518,6 +1534,7 @@ static bool __vxlan_sock_release_prep(struct vxlan_sock *vs)
 
 static void vxlan_sock_release(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_sock *sock4 = rtnl_dereference(vxlan->vn4_sock);
 #if IS_ENABLED(CONFIG_IPV6)
 	struct vxlan_sock *sock6 = rtnl_dereference(vxlan->vn6_sock);
@@ -1527,7 +1544,7 @@ static void vxlan_sock_release(struct vxlan_dev *vxlan)
 
 	RCU_INIT_POINTER(vxlan->vn4_sock, NULL);
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vs_del_vnigrp(vxlan);
 	else
 		vxlan_vs_del_dev(vxlan);
@@ -1598,6 +1615,7 @@ static void vxlan_parse_gbp_hdr(struct sk_buff *skb, u32 vxflags,
 }
 
 static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,
+					  const struct vxlan_config *cfg,
 					  struct vxlan_sock *vs,
 					  struct sk_buff *skb, __be32 vni)
 {
@@ -1623,10 +1641,10 @@ static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,
 #endif
 	}
 
-	if (!(vxlan->cfg.flags & VXLAN_F_LEARN))
+	if (!(cfg->flags & VXLAN_F_LEARN))
 		return SKB_NOT_DROPPED_YET;
 
-	return vxlan_snoop(skb->dev, &saddr, eth_hdr(skb)->h_source,
+	return vxlan_snoop(skb->dev, cfg, &saddr, eth_hdr(skb)->h_source,
 			   ifindex, vni);
 }
 
@@ -1657,18 +1675,21 @@ static bool vxlan_ecn_decapsulate(struct vxlan_sock *vs, void *oiph,
 static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 {
 	struct vxlan_vni_node *vninode = NULL;
-	const struct vxlanhdr *vh;
-	struct vxlan_dev *vxlan;
-	struct vxlan_sock *vs;
-	struct vxlan_metadata _md;
-	struct vxlan_metadata *md = &_md;
 	__be16 protocol = htons(ETH_P_TEB);
+	const struct vxlan_config *cfg;
 	enum skb_drop_reason reason;
+	const struct vxlanhdr *vh;
+	struct vxlan_metadata *md;
+	struct vxlan_metadata _md;
+	struct vxlan_dev *vxlan;
 	bool raw_proto = false;
-	void *oiph;
+	struct vxlan_sock *vs;
 	__be32 vni = 0;
+	void *oiph;
 	int nh;
 
+	md = &_md;
+
 	/* Need UDP and VXLAN header to be present */
 	reason = pskb_may_pull_reason(skb, VXLAN_HLEN);
 	if (reason)
@@ -1696,8 +1717,10 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		goto drop;
 	}
 
-	if (vh->vx_flags & vxlan->cfg.reserved_bits.vx_flags ||
-	    vh->vx_vni & vxlan->cfg.reserved_bits.vx_vni) {
+	cfg = rcu_dereference(vxlan->cfg);
+
+	if (vh->vx_flags & cfg->reserved_bits.vx_flags ||
+	    vh->vx_vni & cfg->reserved_bits.vx_vni) {
 		/* If the header uses bits besides those enabled by the
 		 * netdevice configuration, treat this as a malformed packet.
 		 * This behavior diverges from VXLAN RFC (RFC7348) which
@@ -1709,12 +1732,12 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		reason = SKB_DROP_REASON_VXLAN_INVALID_HDR;
 		DEV_STATS_INC(vxlan->dev, rx_frame_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_GPE) {
+	if (cfg->flags & VXLAN_F_GPE) {
 		if (!vxlan_parse_gpe_proto(vh, &protocol))
 			goto drop;
 		raw_proto = true;
@@ -1726,8 +1749,8 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		goto drop;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_REMCSUM_RX) {
-		reason = vxlan_remcsum(skb, vxlan->cfg.flags);
+	if (cfg->flags & VXLAN_F_REMCSUM_RX) {
+		reason = vxlan_remcsum(skb, cfg->flags);
 		if (unlikely(reason))
 			goto drop;
 	}
@@ -1752,14 +1775,14 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		memset(md, 0, sizeof(*md));
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_GBP)
-		vxlan_parse_gbp_hdr(skb, vxlan->cfg.flags, md);
+	if (cfg->flags & VXLAN_F_GBP)
+		vxlan_parse_gbp_hdr(skb, cfg->flags, md);
 	/* Note that GBP and GPE can never be active together. This is
 	 * ensured in vxlan_dev_configure.
 	 */
 
 	if (!raw_proto) {
-		reason = vxlan_set_mac(vxlan, vs, skb, vni);
+		reason = vxlan_set_mac(vxlan, cfg, vs, skb, vni);
 		if (reason)
 			goto drop;
 	} else {
@@ -1780,7 +1803,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	if (reason) {
 		DEV_STATS_INC(vxlan->dev, rx_length_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
@@ -1792,7 +1815,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		reason = SKB_DROP_REASON_IP_TUNNEL_ECN;
 		DEV_STATS_INC(vxlan->dev, rx_frame_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
@@ -1802,14 +1825,15 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	if (unlikely(!(vxlan->dev->flags & IFF_UP))) {
 		rcu_read_unlock();
 		dev_dstats_rx_dropped(vxlan->dev);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_DROPS, 0);
 		reason = SKB_DROP_REASON_DEV_READY;
 		goto drop;
 	}
 
 	dev_dstats_rx_add(vxlan->dev, skb->len);
-	vxlan_vnifilter_count(vxlan, vni, vninode, VXLAN_VNI_STATS_RX, skb->len);
+	vxlan_vnifilter_count(vxlan, cfg, vni, vninode, VXLAN_VNI_STATS_RX,
+			      skb->len);
 	gro_cells_receive(&vxlan->gro_cells, skb);
 
 	rcu_read_unlock();
@@ -1850,7 +1874,8 @@ static int vxlan_err_lookup(struct sock *sk, struct sk_buff *skb)
 	return 0;
 }
 
-static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
+static int arp_reduce(struct net_device *dev, struct sk_buff *skb,
+		      const struct vxlan_config *cfg, __be32 vni)
 {
 	struct neigh_table *tbl = arp_table(dev_net(dev));
 	struct vxlan_dev *vxlan = netdev_priv(dev);
@@ -1864,7 +1889,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 	if (!pskb_network_may_pull(skb, arp_hdr_len(dev))) {
 		dev_dstats_tx_dropped(dev);
-		vxlan_vnifilter_count(vxlan, vni, NULL,
+		vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 				      VXLAN_VNI_STATS_TX_DROPS, 0);
 		goto out;
 	}
@@ -1905,7 +1930,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 		neigh_ha_snapshot(ha, n, n->dev);
 
 		rcu_read_lock();
-		f = vxlan_find_mac_tx(vxlan, ha, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, ha, vni);
 		if (f)
 			rdst = first_remote_rcu(f);
 		if (rdst && vxlan_addr_any(&rdst->remote_ip)) {
@@ -1931,11 +1956,11 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 		if (netif_rx(reply) == NET_RX_DROP) {
 			dev_dstats_rx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_RX_DROPS, 0);
 		}
 
-	} else if (vxlan->cfg.flags & VXLAN_F_L3MISS) {
+	} else if (cfg->flags & VXLAN_F_L3MISS) {
 		union vxlan_addr ipa = {
 			.sin.sin_addr.s_addr = tip,
 			.sin.sin_family = AF_INET,
@@ -2043,7 +2068,8 @@ static struct sk_buff *vxlan_na_create(struct sk_buff *request,
 	return reply;
 }
 
-static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
+static int neigh_reduce(struct net_device *dev, struct sk_buff *skb,
+			const struct vxlan_config *cfg, __be32 vni)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	const struct in6_addr *daddr;
@@ -2077,7 +2103,7 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 		}
 
 		neigh_ha_snapshot(ha, n, n->dev);
-		f = vxlan_find_mac_tx(vxlan, ha, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, ha, vni);
 		if (f)
 			rdst = first_remote_rcu(f);
 		if (rdst && vxlan_addr_any(&rdst->remote_ip)) {
@@ -2096,10 +2122,10 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 		if (netif_rx(reply) == NET_RX_DROP) {
 			dev_dstats_rx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_RX_DROPS, 0);
 		}
-	} else if (vxlan->cfg.flags & VXLAN_F_L3MISS) {
+	} else if (cfg->flags & VXLAN_F_L3MISS) {
 		union vxlan_addr ipa = {
 			.sin6.sin6_addr = msg->target,
 			.sin6.sin6_family = AF_INET6,
@@ -2115,9 +2141,9 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 }
 #endif
 
-static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
+static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb,
+			       const struct vxlan_config *cfg)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct neigh_table *tbl;
 	struct neighbour *n;
 
@@ -2136,7 +2162,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
 		tbl = arp_table(dev_net(dev));
 		pip = ip_hdr(skb);
 		n = neigh_lookup(tbl, &pip->daddr, dev);
-		if (!n && (vxlan->cfg.flags & VXLAN_F_L3MISS)) {
+		if (!n && (cfg->flags & VXLAN_F_L3MISS)) {
 			union vxlan_addr ipa = {
 				.sin.sin_addr.s_addr = pip->daddr,
 				.sin.sin_family = AF_INET,
@@ -2164,7 +2190,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
 		tbl = nd_table(dev_net(dev));
 		pip6 = ipv6_hdr(skb);
 		n = neigh_lookup(tbl, &pip6->daddr, dev);
-		if (!n && (vxlan->cfg.flags & VXLAN_F_L3MISS)) {
+		if (!n && (cfg->flags & VXLAN_F_L3MISS)) {
 			union vxlan_addr ipa = {
 				.sin6.sin6_addr = pip6->daddr,
 				.sin6.sin6_family = AF_INET6,
@@ -2282,20 +2308,28 @@ static int vxlan_build_skb(struct sk_buff *skb, struct dst_entry *dst,
 
 /* Bypass encapsulation if the destination is local */
 static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
-			       struct vxlan_dev *dst_vxlan, __be32 vni,
-			       bool snoop)
+			       struct vxlan_dev *dst_vxlan,
+			       const struct vxlan_config *src_cfg,
+			       __be32 vni, bool snoop)
 {
+	const struct vxlan_config *dst_cfg;
 	union vxlan_addr loopback;
-	union vxlan_addr *remote_ip = &dst_vxlan->default_dst.remote_ip;
 	unsigned int len = skb->len;
-	struct net_device *dev;
+	struct net_device *dev = dst_vxlan->dev;
 
 	skb->pkt_type = PACKET_HOST;
 	skb->encapsulation = 0;
-	skb->dev = dst_vxlan->dev;
+	skb->dev = dev;
 	__skb_pull(skb, skb_network_offset(skb));
 
-	if (remote_ip->sa.sa_family == AF_INET) {
+	rcu_read_lock();
+	dst_cfg = rcu_dereference(dst_vxlan->cfg);
+	if (unlikely(!(dev->flags & IFF_UP))) {
+		kfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);
+		goto drop;
+	}
+
+	if (dst_cfg->remote_ip.sa.sa_family == AF_INET) {
 		loopback.sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
 		loopback.sa.sa_family =  AF_INET;
 #if IS_ENABLED(CONFIG_IPV6)
@@ -2305,27 +2339,20 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
 #endif
 	}
 
-	rcu_read_lock();
-	dev = skb->dev;
-	if (unlikely(!(dev->flags & IFF_UP))) {
-		kfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);
-		goto drop;
-	}
-
-	if ((dst_vxlan->cfg.flags & VXLAN_F_LEARN) && snoop)
-		vxlan_snoop(dev, &loopback, eth_hdr(skb)->h_source, 0, vni);
+	if ((dst_cfg->flags & VXLAN_F_LEARN) && snoop)
+		vxlan_snoop(dev, dst_cfg, &loopback, eth_hdr(skb)->h_source, 0, vni);
 
 	dev_dstats_tx_add(src_vxlan->dev, len);
-	vxlan_vnifilter_count(src_vxlan, vni, NULL, VXLAN_VNI_STATS_TX, len);
+	vxlan_vnifilter_count(src_vxlan, src_cfg, vni, NULL, VXLAN_VNI_STATS_TX, len);
 
 	if (__netif_rx(skb) == NET_RX_SUCCESS) {
 		dev_dstats_rx_add(dst_vxlan->dev, len);
-		vxlan_vnifilter_count(dst_vxlan, vni, NULL, VXLAN_VNI_STATS_RX,
+		vxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL, VXLAN_VNI_STATS_RX,
 				      len);
 	} else {
 drop:
 		dev_dstats_rx_dropped(dev);
-		vxlan_vnifilter_count(dst_vxlan, vni, NULL,
+		vxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL,
 				      VXLAN_VNI_STATS_RX_DROPS, 0);
 	}
 	rcu_read_unlock();
@@ -2333,6 +2360,7 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
 
 static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 				 struct vxlan_dev *vxlan,
+				 const struct vxlan_config *cfg,
 				 int addr_family,
 				 __be16 dst_port, int dst_ifindex, __be32 vni,
 				 struct dst_entry *dst,
@@ -2348,22 +2376,22 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 	/* Bypass encapsulation if the destination is local */
 	if (rt_flags & RTCF_LOCAL &&
 	    !(rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) &&
-	    vxlan->cfg.flags & VXLAN_F_LOCALBYPASS) {
+	    cfg->flags & VXLAN_F_LOCALBYPASS) {
 		struct vxlan_dev *dst_vxlan;
 
 		dst_release(dst);
 		dst_vxlan = vxlan_find_vni(vxlan->net, dst_ifindex, vni,
 					   addr_family, dst_port,
-					   vxlan->cfg.flags);
+					   cfg->flags);
 		if (!dst_vxlan) {
 			DEV_STATS_INC(dev, tx_errors);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_TX_ERRORS, 0);
 			kfree_skb_reason(skb, SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND);
 
 			return -ENOENT;
 		}
-		vxlan_encap_bypass(skb, vxlan, dst_vxlan, vni, true);
+		vxlan_encap_bypass(skb, vxlan, dst_vxlan, cfg, vni, true);
 		return 1;
 	}
 
@@ -2371,30 +2399,35 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 }
 
 void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
+		    const struct vxlan_config *cfg,
 		    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc)
 {
+	unsigned int pkt_len = skb->len;
+	struct vxlan_metadata _md = {};
+	__be16 src_port = 0, dst_port;
+	struct dst_entry *ndst = NULL;
+	enum skb_drop_reason reason;
 	struct dst_cache *dst_cache;
+	const struct iphdr *old_iph;
 	struct ip_tunnel_info *info;
 	struct ip_tunnel_key *pkey;
+	struct vxlan_metadata *md;
 	struct ip_tunnel_key key;
-	struct vxlan_dev *vxlan = netdev_priv(dev);
-	const struct iphdr *old_iph;
-	struct vxlan_metadata _md = {};
-	struct vxlan_metadata *md = &_md;
-	unsigned int pkt_len = skb->len;
-	__be16 src_port = 0, dst_port;
-	struct dst_entry *ndst = NULL;
+	struct vxlan_dev *vxlan;
+	u32 flags = cfg->flags;
+	bool udp_sum = false;
+	bool no_eth_encap;
 	int addr_family;
+	bool use_cache;
+	__be32 vni = 0;
 	__u8 tos, ttl;
 	int ifindex;
 	int err = 0;
-	u32 flags = vxlan->cfg.flags;
-	bool use_cache;
-	bool udp_sum = false;
-	bool xnet = !net_eq(vxlan->net, dev_net(vxlan->dev));
-	enum skb_drop_reason reason;
-	bool no_eth_encap;
-	__be32 vni = 0;
+	bool xnet;
+
+	vxlan = netdev_priv(dev);
+	xnet = !net_eq(vxlan->net, dev_net(vxlan->dev));
+	md = &_md;
 
 	no_eth_encap = flags & VXLAN_F_GPE && skb->protocol != htons(ETH_P_TEB);
 	reason = skb_vlan_inet_prepare(skb, no_eth_encap);
@@ -2414,23 +2447,23 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (vxlan_addr_any(&rdst->remote_ip)) {
 			if (did_rsc) {
 				/* short-circuited back to local bridge */
-				vxlan_encap_bypass(skb, vxlan, vxlan,
+				vxlan_encap_bypass(skb, vxlan, vxlan, cfg,
 						   default_vni, true);
 				return;
 			}
 			goto drop;
 		}
 
-		addr_family = vxlan->cfg.saddr.sa.sa_family;
-		dst_port = rdst->remote_port ? rdst->remote_port : vxlan->cfg.dst_port;
+		addr_family = cfg->saddr.sa.sa_family;
+		dst_port = rdst->remote_port ? rdst->remote_port : cfg->dst_port;
 		vni = (rdst->remote_vni) ? : default_vni;
 		ifindex = rdst->remote_ifindex;
 
 		if (addr_family == AF_INET) {
-			key.u.ipv4.src = vxlan->cfg.saddr.sin.sin_addr.s_addr;
+			key.u.ipv4.src = cfg->saddr.sin.sin_addr.s_addr;
 			key.u.ipv4.dst = rdst->remote_ip.sin.sin_addr.s_addr;
 		} else {
-			key.u.ipv6.src = vxlan->cfg.saddr.sin6.sin6_addr;
+			key.u.ipv6.src = cfg->saddr.sin6.sin6_addr;
 			key.u.ipv6.dst = rdst->remote_ip.sin6.sin6_addr;
 		}
 
@@ -2439,11 +2472,11 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (flags & VXLAN_F_TTL_INHERIT) {
 			ttl = ip_tunnel_get_ttl(old_iph, skb);
 		} else {
-			ttl = vxlan->cfg.ttl;
+			ttl = cfg->ttl;
 			if (!ttl && vxlan_addr_multicast(&rdst->remote_ip))
 				ttl = 1;
 		}
-		tos = vxlan->cfg.tos;
+		tos = cfg->tos;
 		if (tos == 1)
 			tos = ip_tunnel_get_dsfield(old_iph, skb);
 		if (tos && !info)
@@ -2454,9 +2487,9 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		else
 			udp_sum = !(flags & VXLAN_F_UDP_ZERO_CSUM6_TX);
 #if IS_ENABLED(CONFIG_IPV6)
-		switch (vxlan->cfg.label_policy) {
+		switch (cfg->label_policy) {
 		case VXLAN_LABEL_FIXED:
-			key.label = vxlan->cfg.label;
+			key.label = cfg->label;
 			break;
 		case VXLAN_LABEL_INHERIT:
 			key.label = ip_tunnel_get_flowlabel(old_iph, skb);
@@ -2474,7 +2507,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		}
 		pkey = &info->key;
 		addr_family = ip_tunnel_info_af(info);
-		dst_port = info->key.tp_dst ? : vxlan->cfg.dst_port;
+		dst_port = info->key.tp_dst ? : cfg->dst_port;
 		vni = tunnel_id_to_key32(info->key.tun_id);
 		ifindex = 0;
 		dst_cache = &info->dst_cache;
@@ -2487,8 +2520,8 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		tos = info->key.tos;
 		udp_sum = test_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags);
 	}
-	src_port = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
-				     vxlan->cfg.port_max, true);
+	src_port = udp_flow_src_port(dev_net(dev), skb, cfg->port_min,
+				     cfg->port_max, true);
 
 	rcu_read_lock();
 	if (addr_family == AF_INET) {
@@ -2521,15 +2554,15 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 
 		if (!info) {
 			/* Bypass encapsulation if the destination is local */
-			err = encap_bypass_if_local(skb, dev, vxlan, AF_INET,
+			err = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET,
 						    dst_port, ifindex, vni,
 						    &rt->dst, rt->rt_flags);
 			if (err)
 				goto out_unlock;
 
-			if (vxlan->cfg.df == VXLAN_DF_SET) {
+			if (cfg->df == VXLAN_DF_SET) {
 				df = htons(IP_DF);
-			} else if (vxlan->cfg.df == VXLAN_DF_INHERIT) {
+			} else if (cfg->df == VXLAN_DF_INHERIT) {
 				struct ethhdr *eth = eth_hdr(skb);
 
 				if (ntohs(eth->h_proto) == ETH_P_IPV6 ||
@@ -2558,7 +2591,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				unclone->key.u.ipv4.src = pkey->u.ipv4.dst;
 				unclone->key.u.ipv4.dst = saddr;
 			}
-			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
+			vxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);
 			dst_release(ndst);
 			goto out_unlock;
 		}
@@ -2608,7 +2641,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (!info) {
 			u32 rt6i_flags = dst_rt6_info(ndst)->rt6i_flags;
 
-			err = encap_bypass_if_local(skb, dev, vxlan, AF_INET6,
+			err = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET6,
 						    dst_port, ifindex, vni,
 						    ndst, rt6i_flags);
 			if (err)
@@ -2632,7 +2665,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				unclone->key.u.ipv6.dst = saddr;
 			}
 
-			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
+			vxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);
 			dst_release(ndst);
 			goto out_unlock;
 		}
@@ -2653,14 +2686,14 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				     ip6cb_flags);
 #endif
 	}
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);
 out_unlock:
 	rcu_read_unlock();
 	return;
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);
 	kfree_skb_reason(skb, reason);
 	return;
 
@@ -2672,11 +2705,12 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		DEV_STATS_INC(dev, tx_carrier_errors);
 	dst_release(ndst);
 	DEV_STATS_INC(dev, tx_errors);
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);
 	kfree_skb_reason(skb, reason);
 }
 
 static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
+			  const struct vxlan_config *cfg,
 			  struct vxlan_fdb *f, __be32 vni, bool did_rsc)
 {
 	struct vxlan_rdst nh_rdst;
@@ -2693,7 +2727,7 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
 	do_xmit = vxlan_fdb_nh_path_select(nh, hash, &nh_rdst);
 
 	if (likely(do_xmit))
-		vxlan_xmit_one(skb, dev, vni, &nh_rdst, did_rsc);
+		vxlan_xmit_one(skb, dev, cfg, vni, &nh_rdst, did_rsc);
 	else
 		goto drop;
 
@@ -2701,15 +2735,15 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(netdev_priv(dev), vni, NULL,
+	vxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,
 			      VXLAN_VNI_STATS_TX_DROPS, 0);
 	dev_kfree_skb(skb);
 }
 
 static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
-				   u32 nhid, __be32 vni)
+				   u32 nhid, __be32 vni,
+				   const struct vxlan_config *cfg)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst nh_rdst;
 	struct nexthop *nh;
 	bool do_xmit;
@@ -2727,11 +2761,11 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
 	do_xmit = vxlan_fdb_nh_path_select(nh, hash, &nh_rdst);
 	rcu_read_unlock();
 
-	if (vxlan->cfg.saddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)
+	if (cfg->saddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)
 		goto drop;
 
 	if (likely(do_xmit))
-		vxlan_xmit_one(skb, dev, vni, &nh_rdst, false);
+		vxlan_xmit_one(skb, dev, cfg, vni, &nh_rdst, false);
 	else
 		goto drop;
 
@@ -2739,7 +2773,7 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(netdev_priv(dev), vni, NULL,
+	vxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,
 			      VXLAN_VNI_STATS_TX_DROPS, 0);
 	dev_kfree_skb(skb);
 	return NETDEV_TX_OK;
@@ -2756,34 +2790,47 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst *rdst, *fdst = NULL;
 	const struct ip_tunnel_info *info;
+	const struct vxlan_config *cfg;
+	__be32 default_vni;
 	struct vxlan_fdb *f;
 	struct ethhdr *eth;
 	__be32 vni = 0;
-	u32 nhid = 0;
 	bool did_rsc;
+	u32 nhid = 0;
+	u32 flags;
+
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+	flags = cfg->flags;
+	default_vni = cfg->vni;
 
 	info = skb_tunnel_info(skb);
 
 	skb_reset_mac_header(skb);
 
-	if (vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) {
+	if (flags & VXLAN_F_COLLECT_METADATA) {
 		if (info && info->mode & IP_TUNNEL_INFO_BRIDGE &&
 		    info->mode & IP_TUNNEL_INFO_TX) {
 			vni = tunnel_id_to_key32(info->key.tun_id);
 			nhid = info->key.nhid;
 		} else {
 			if (info && info->mode & IP_TUNNEL_INFO_TX)
-				vxlan_xmit_one(skb, dev, vni, NULL, false);
+				vxlan_xmit_one(skb, dev, cfg, vni, NULL, false);
 			else
 				kfree_skb_reason(skb, SKB_DROP_REASON_TUNNEL_TXINFO);
+			rcu_read_unlock();
 			return NETDEV_TX_OK;
 		}
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_PROXY) {
+	if (flags & VXLAN_F_PROXY) {
 		eth = eth_hdr(skb);
-		if (ntohs(eth->h_proto) == ETH_P_ARP)
-			return arp_reduce(dev, skb, vni);
+		if (ntohs(eth->h_proto) == ETH_P_ARP) {
+			netdev_tx_t res = arp_reduce(dev, skb, cfg, vni);
+
+			rcu_read_unlock();
+			return res;
+		}
 #if IS_ENABLED(CONFIG_IPV6)
 		else if (ntohs(eth->h_proto) == ETH_P_IPV6 &&
 			 pskb_network_may_pull(skb, sizeof(struct ipv6hdr) +
@@ -2792,53 +2839,57 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 			struct nd_msg *m = (struct nd_msg *)(ipv6_hdr(skb) + 1);
 
 			if (m->icmph.icmp6_code == 0 &&
-			    m->icmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION)
-				return neigh_reduce(dev, skb, vni);
+			    m->icmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION) {
+				netdev_tx_t res = neigh_reduce(dev, skb, cfg, vni);
+
+				rcu_read_unlock();
+				return res;
+			}
 		}
 #endif
 	}
 
-	if (nhid)
-		return vxlan_xmit_nhid(skb, dev, nhid, vni);
+	if (nhid) {
+		netdev_tx_t res = vxlan_xmit_nhid(skb, dev, nhid, vni, cfg);
+
+		rcu_read_unlock();
+		return res;
+	}
 
-	if (vxlan->cfg.flags & VXLAN_F_MDB) {
+	if (test_bit(VXLAN_DEV_F_MDB, &vxlan->flags)) {
 		struct vxlan_mdb_entry *mdb_entry;
 
-		rcu_read_lock();
-		mdb_entry = vxlan_mdb_entry_skb_get(vxlan, skb, vni);
+		mdb_entry = vxlan_mdb_entry_skb_get(vxlan, cfg, skb, vni);
 		if (mdb_entry) {
-			netdev_tx_t ret;
+			netdev_tx_t ret = vxlan_mdb_xmit(vxlan, cfg, mdb_entry, skb);
 
-			ret = vxlan_mdb_xmit(vxlan, mdb_entry, skb);
 			rcu_read_unlock();
 			return ret;
 		}
-		rcu_read_unlock();
 	}
 
 	eth = eth_hdr(skb);
-	rcu_read_lock();
-	f = vxlan_find_mac_tx(vxlan, eth->h_dest, vni);
+	f = vxlan_find_mac_tx(vxlan, cfg, eth->h_dest, vni);
 	did_rsc = false;
 
-	if (f && (f->flags & NTF_ROUTER) && (vxlan->cfg.flags & VXLAN_F_RSC) &&
+	if (f && (f->flags & NTF_ROUTER) && (flags & VXLAN_F_RSC) &&
 	    (ntohs(eth->h_proto) == ETH_P_IP ||
 	     ntohs(eth->h_proto) == ETH_P_IPV6)) {
-		did_rsc = route_shortcircuit(dev, skb);
+		did_rsc = route_shortcircuit(dev, skb, cfg);
 		eth = eth_hdr(skb);
 		if (did_rsc)
-			f = vxlan_find_mac_tx(vxlan, eth->h_dest, vni);
+			f = vxlan_find_mac_tx(vxlan, cfg, eth->h_dest, vni);
 	}
 
 	if (f == NULL) {
-		f = vxlan_find_mac_tx(vxlan, all_zeros_mac, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, all_zeros_mac, vni);
 		if (f == NULL) {
-			if ((vxlan->cfg.flags & VXLAN_F_L2MISS) &&
+			if ((flags & VXLAN_F_L2MISS) &&
 			    !is_multicast_ether_addr(eth->h_dest))
 				vxlan_fdb_miss(vxlan, eth->h_dest);
 
 			dev_dstats_tx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_TX_DROPS, 0);
 			kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
 			goto out;
@@ -2846,8 +2897,8 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	if (rcu_access_pointer(f->nh)) {
-		vxlan_xmit_nh(skb, dev, f,
-			      (vni ? : vxlan->default_dst.remote_vni), did_rsc);
+		vxlan_xmit_nh(skb, dev, cfg, f,
+			      (vni ? : default_vni), did_rsc);
 	} else {
 		list_for_each_entry_rcu(rdst, &f->remotes, list) {
 			struct sk_buff *skb1;
@@ -2858,10 +2909,10 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 			}
 			skb1 = skb_clone(skb, GFP_ATOMIC);
 			if (skb1)
-				vxlan_xmit_one(skb1, dev, vni, rdst, did_rsc);
+				vxlan_xmit_one(skb1, dev, cfg, vni, rdst, did_rsc);
 		}
 		if (fdst)
-			vxlan_xmit_one(skb, dev, vni, fdst, did_rsc);
+			vxlan_xmit_one(skb, dev, cfg, vni, fdst, did_rsc);
 		else
 			kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
 	}
@@ -2876,12 +2927,15 @@ static void vxlan_cleanup(struct timer_list *t)
 {
 	struct vxlan_dev *vxlan = timer_container_of(vxlan, t, age_timer);
 	unsigned long next_timer = jiffies + FDB_AGE_INTERVAL;
+	const struct vxlan_config *cfg;
 	struct vxlan_fdb *f;
 
 	if (!netif_running(vxlan->dev))
 		return;
 
 	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+
 	hlist_for_each_entry_rcu(f, &vxlan->fdb_list, fdb_node) {
 		unsigned long timeout;
 
@@ -2891,7 +2945,7 @@ static void vxlan_cleanup(struct timer_list *t)
 		if (f->flags & NTF_EXT_LEARNED)
 			continue;
 
-		timeout = READ_ONCE(f->updated) + vxlan->cfg.age_interval * HZ;
+		timeout = READ_ONCE(f->updated) + cfg->age_interval * HZ;
 		if (time_before_eq(timeout, jiffies)) {
 			spin_lock(&vxlan->hash_lock);
 			if (!hlist_unhashed(&f->fdb_node)) {
@@ -2923,10 +2977,14 @@ static void vxlan_vs_del_dev(struct vxlan_dev *vxlan)
 static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,
 			     struct vxlan_dev_node *node)
 {
-	__be32 vni = vxlan->default_dst.remote_vni;
+	const struct vxlan_config *cfg;
+	__be32 vni;
 
 	ASSERT_RTNL();
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	node->vxlan = vxlan;
 	hlist_add_head_rcu(&node->hlist, vni_head(vs, vni));
 }
@@ -2935,13 +2993,16 @@ static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,
 static int vxlan_init(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	err = rhashtable_init(&vxlan->fdb_hash_tbl, &vxlan_fdb_rht_params);
 	if (err)
 		return err;
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
+	if (cfg->flags & VXLAN_F_VNIFILTER) {
 		err = vxlan_vnigroup_init(vxlan);
 		if (err)
 			goto err_rhashtable_destroy;
@@ -2961,7 +3022,7 @@ static int vxlan_init(struct net_device *dev)
 err_gro_cells_destroy:
 	gro_cells_destroy(&vxlan->gro_cells);
 err_vnigroup_uninit:
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vnigroup_uninit(vxlan);
 err_rhashtable_destroy:
 	rhashtable_destroy(&vxlan->fdb_hash_tbl);
@@ -2971,10 +3032,13 @@ static int vxlan_init(struct net_device *dev)
 static void vxlan_uninit(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
+
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	vxlan_mdb_fini(vxlan);
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vnigroup_uninit(vxlan);
 
 	gro_cells_destroy(&vxlan->gro_cells);
@@ -2986,6 +3050,7 @@ static void vxlan_uninit(struct net_device *dev)
 static int vxlan_open(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	int ret;
 
 	ret = vxlan_sock_add(vxlan);
@@ -2998,7 +3063,8 @@ static int vxlan_open(struct net_device *dev)
 		return ret;
 	}
 
-	if (vxlan->cfg.age_interval)
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (cfg->age_interval)
 		mod_timer(&vxlan->age_timer, jiffies + FDB_AGE_INTERVAL);
 
 	return ret;
@@ -3020,8 +3086,10 @@ struct vxlan_fdb_flush_desc {
 static bool vxlan_fdb_is_default_entry(const struct vxlan_fdb *f,
 				       const struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	return is_zero_ether_addr(f->key.eth_addr) &&
-	       f->key.vni == vxlan->cfg.vni;
+	       f->key.vni == cfg->vni;
 }
 
 static bool vxlan_fdb_nhid_matches(const struct vxlan_fdb *f, u32 nhid)
@@ -3238,15 +3306,18 @@ static void vxlan_set_multicast_list(struct net_device *dev)
 static int vxlan_change_mtu(struct net_device *dev, int new_mtu)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct net_device *lowerdev = __dev_get_by_index(vxlan->net,
-							 dst->remote_ifindex);
+	const struct vxlan_config *cfg;
+	struct net_device *lowerdev;
+
+	cfg = rtnl_dereference(vxlan->cfg);
+
+	lowerdev = __dev_get_by_index(vxlan->net, cfg->remote_ifindex);
 
 	/* This check is different than dev->max_mtu, because it looks at
 	 * the lowerdev->mtu, rather than the static dev->max_mtu
 	 */
 	if (lowerdev) {
-		int max_mtu = lowerdev->mtu - vxlan_headroom(vxlan->cfg.flags);
+		int max_mtu = lowerdev->mtu - vxlan_headroom(cfg->flags);
 		if (new_mtu > max_mtu)
 			return -EINVAL;
 	}
@@ -3259,11 +3330,14 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct ip_tunnel_info *info = skb_tunnel_info(skb);
+	const struct vxlan_config *cfg;
 	__be16 sport, dport;
 
-	sport = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
-				  vxlan->cfg.port_max, true);
-	dport = info->key.tp_dst ? : vxlan->cfg.dst_port;
+	cfg = rcu_dereference(vxlan->cfg);
+
+	sport = udp_flow_src_port(dev_net(dev), skb, cfg->port_min,
+				  cfg->port_max, true);
+	dport = info->key.tp_dst ? : cfg->dst_port;
 
 	if (ip_tunnel_info_af(info) == AF_INET) {
 		struct vxlan_sock *sock4 = rcu_dereference(vxlan->vn4_sock);
@@ -3373,6 +3447,15 @@ static void vxlan_offload_rx_ports(struct net_device *dev, bool push)
 	}
 }
 
+static void vxlan_free_dev(struct net_device *dev)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_config *cfg = rcu_dereference_protected(vxlan->cfg, 1);
+
+	RCU_INIT_POINTER(vxlan->cfg, NULL);
+	kfree(cfg);
+}
+
 /* Initialize the device structure. */
 static void vxlan_setup(struct net_device *dev)
 {
@@ -3381,6 +3464,8 @@ static void vxlan_setup(struct net_device *dev)
 	eth_hw_addr_random(dev);
 	ether_setup(dev);
 
+	dev->priv_destructor = vxlan_free_dev;
+
 	dev->needs_free_netdev = true;
 	SET_NETDEV_DEVTYPE(dev, &vxlan_type);
 
@@ -3555,9 +3640,11 @@ static int vxlan_get_link_ksettings(struct net_device *dev,
 				    struct ethtool_link_ksettings *cmd)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct net_device *lowerdev = __dev_get_by_index(vxlan->net,
-							 dst->remote_ifindex);
+	const struct vxlan_config *cfg;
+	struct net_device *lowerdev;
+
+	cfg = rtnl_dereference(vxlan->cfg);
+	lowerdev = __dev_get_by_index(vxlan->net, cfg->remote_ifindex);
 
 	if (!lowerdev) {
 		cmd->base.duplex = DUPLEX_UNKNOWN;
@@ -3663,21 +3750,22 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, bool ipv6,
 
 static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 {
-	bool metadata = vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	bool metadata = cfg->flags & VXLAN_F_COLLECT_METADATA;
 	struct vxlan_sock *vs = NULL;
 	struct vxlan_dev_node *node;
 	int l3mdev_index = 0;
 
 	ASSERT_RTNL();
 
-	if (vxlan->cfg.remote_ifindex)
+	if (cfg->remote_ifindex)
 		l3mdev_index = l3mdev_master_upper_ifindex_by_index(
-			vxlan->net, vxlan->cfg.remote_ifindex);
+			vxlan->net, cfg->remote_ifindex);
 
-	if (!vxlan->cfg.no_share) {
+	if (!cfg->no_share) {
 		rcu_read_lock();
 		vs = vxlan_find_sock(vxlan->net, ipv6 ? AF_INET6 : AF_INET,
-				     vxlan->cfg.dst_port, vxlan->cfg.flags,
+				     cfg->dst_port, cfg->flags,
 				     l3mdev_index);
 		if (vs && !refcount_inc_not_zero(&vs->refcnt)) {
 			rcu_read_unlock();
@@ -3687,7 +3775,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 	}
 	if (!vs)
 		vs = vxlan_socket_create(vxlan->net, ipv6,
-					 vxlan->cfg.dst_port, vxlan->cfg.flags,
+					 cfg->dst_port, cfg->flags,
 					 l3mdev_index);
 	if (IS_ERR(vs))
 		return PTR_ERR(vs);
@@ -3702,7 +3790,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 		node = &vxlan->hlist4;
 	}
 
-	if (metadata && (vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (metadata && (cfg->flags & VXLAN_F_VNIFILTER))
 		vxlan_vs_add_vnigrp(vxlan, vs, ipv6);
 	else
 		vxlan_vs_add_dev(vs, vxlan, node);
@@ -3712,11 +3800,14 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 
 static int vxlan_sock_add(struct vxlan_dev *vxlan)
 {
-	bool metadata = vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA;
-	bool ipv6 = vxlan->cfg.flags & VXLAN_F_IPV6 || metadata;
-	bool ipv4 = !ipv6 || metadata;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	bool metadata, ipv6, ipv4;
 	int ret = 0;
 
+	metadata = cfg->flags & VXLAN_F_COLLECT_METADATA;
+	ipv6 = (cfg->flags & VXLAN_F_IPV6) || metadata;
+	ipv4 = !ipv6 || metadata;
+
 	RCU_INIT_POINTER(vxlan->vn4_sock, NULL);
 #if IS_ENABLED(CONFIG_IPV6)
 	RCU_INIT_POINTER(vxlan->vn6_sock, NULL);
@@ -3734,28 +3825,33 @@ static int vxlan_sock_add(struct vxlan_dev *vxlan)
 }
 
 int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,
-		     struct vxlan_config *conf, __be32 vni)
+		     const struct vxlan_config *conf, __be32 vni)
 {
 	struct vxlan_net *vn = net_generic(src_net, vxlan_net_id);
 	struct vxlan_dev *tmp;
 
 	list_for_each_entry(tmp, &vn->vxlan_list, next) {
+		const struct vxlan_config *tmp_cfg;
+
 		if (tmp == vxlan)
 			continue;
-		if (tmp->cfg.flags & VXLAN_F_VNIFILTER) {
+
+		tmp_cfg = rtnl_dereference(tmp->cfg);
+
+		if (tmp_cfg->flags & VXLAN_F_VNIFILTER) {
 			if (!vxlan_vnifilter_lookup(tmp, vni))
 				continue;
-		} else if (tmp->cfg.vni != vni) {
+		} else if (tmp_cfg->vni != vni) {
 			continue;
 		}
-		if (tmp->cfg.dst_port != conf->dst_port)
+		if (tmp_cfg->dst_port != conf->dst_port)
 			continue;
-		if ((tmp->cfg.flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=
+		if ((tmp_cfg->flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=
 		    (conf->flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)))
 			continue;
 
 		if ((conf->flags & VXLAN_F_IPV6_LINKLOCAL) &&
-		    tmp->cfg.remote_ifindex != conf->remote_ifindex)
+		    tmp_cfg->remote_ifindex != conf->remote_ifindex)
 			continue;
 
 		return -EEXIST;
@@ -3886,6 +3982,13 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,
 			return -EINVAL;
 		}
 
+		if ((conf->flags & VXLAN_F_VNIFILTER) && old &&
+		    vxlan_vnifilter_has_multicast(old)) {
+			NL_SET_ERR_MSG(extack,
+				       "Local interface required for multicast remote group");
+			return -EINVAL;
+		}
+
 #if IS_ENABLED(CONFIG_IPV6)
 		if (conf->flags & VXLAN_F_IPV6_LINKLOCAL) {
 			NL_SET_ERR_MSG(extack,
@@ -3917,16 +4020,16 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,
 }
 
 static void vxlan_config_apply(struct net_device *dev,
-			       struct vxlan_config *conf,
+			       struct vxlan_config *new_cfg,
 			       struct net_device *lowerdev,
 			       struct net *src_net,
-			       bool changelink)
+			       bool changelink, bool lowerdev_changed)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
 	unsigned short needed_headroom = ETH_HLEN;
+	struct vxlan_config *old_cfg;
 	int max_mtu = ETH_MAX_MTU;
-	u32 flags = conf->flags;
+	u32 flags = new_cfg->flags;
 
 	if (!changelink) {
 		if (flags & VXLAN_F_GPE)
@@ -3934,19 +4037,14 @@ static void vxlan_config_apply(struct net_device *dev,
 		else
 			vxlan_ether_setup(dev);
 
-		if (conf->mtu)
-			dev->mtu = conf->mtu;
+		if (new_cfg->mtu)
+			WRITE_ONCE(dev->mtu, new_cfg->mtu);
 
 		vxlan->net = src_net;
 	}
 
-	dst->remote_vni = conf->vni;
-
-	memcpy(&dst->remote_ip, &conf->remote_ip, sizeof(conf->remote_ip));
-
+	dev->needed_tailroom = 0;
 	if (lowerdev) {
-		dst->remote_ifindex = conf->remote_ifindex;
-
 		netif_inherit_tso_max(dev, lowerdev);
 
 		needed_headroom = lowerdev->hard_header_len;
@@ -3954,23 +4052,27 @@ static void vxlan_config_apply(struct net_device *dev,
 
 		dev->needed_tailroom = lowerdev->needed_tailroom;
 
-		max_mtu = lowerdev->mtu - vxlan_headroom(flags);
+		max_mtu = READ_ONCE(lowerdev->mtu) - vxlan_headroom(flags);
 		if (max_mtu < ETH_MIN_MTU)
 			max_mtu = ETH_MIN_MTU;
 
-		if (!changelink && !conf->mtu)
-			dev->mtu = max_mtu;
+		if (!changelink && !new_cfg->mtu)
+			WRITE_ONCE(dev->mtu, max_mtu);
 	}
 
-	if (dev->mtu > max_mtu)
-		dev->mtu = max_mtu;
+	/* A changelink leaving the lower device alone must not shrink the mtu */
+	if (lowerdev_changed && READ_ONCE(dev->mtu) > max_mtu)
+		WRITE_ONCE(dev->mtu, max_mtu);
 
 	if (flags & VXLAN_F_COLLECT_METADATA)
 		flags |= VXLAN_F_IPV6;
 	needed_headroom += vxlan_headroom(flags);
 	dev->needed_headroom = needed_headroom;
 
-	memcpy(&vxlan->cfg, conf, sizeof(*conf));
+	old_cfg = rtnl_dereference(vxlan->cfg);
+	rcu_assign_pointer(vxlan->cfg, new_cfg);
+	if (old_cfg)
+		kfree_rcu(old_cfg, rcu);
 }
 
 static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
@@ -3979,13 +4081,18 @@ static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct net_device *lowerdev;
+	struct vxlan_config *new_cfg;
 	int ret;
 
 	ret = vxlan_config_validate(src_net, conf, &lowerdev, vxlan, extack);
 	if (ret)
 		return ret;
 
-	vxlan_config_apply(dev, conf, lowerdev, src_net, false);
+	new_cfg = kmemdup(conf, sizeof(*conf), GFP_KERNEL);
+	if (!new_cfg)
+		return -ENOMEM;
+
+	vxlan_config_apply(dev, new_cfg, lowerdev, src_net, false, true);
 
 	return 0;
 }
@@ -3997,22 +4104,25 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct net_device *remote_dev = NULL;
-	struct vxlan_rdst *dst;
+	const struct vxlan_config *cfg;
 	int err;
 
-	dst = &vxlan->default_dst;
 	err = vxlan_dev_configure(net, dev, conf, extack);
 	if (err)
 		return err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	dev->ethtool_ops = &vxlan_ethtool_ops;
 
 	err = register_netdevice(dev);
-	if (err)
+	if (err) {
+		vxlan_free_dev(dev);
 		return err;
+	}
 
-	if (dst->remote_ifindex) {
-		remote_dev = __dev_get_by_index(net, dst->remote_ifindex);
+	if (cfg->remote_ifindex) {
+		remote_dev = __dev_get_by_index(net, cfg->remote_ifindex);
 		if (!remote_dev) {
 			err = -ENODEV;
 			goto unregister;
@@ -4022,7 +4132,7 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 		if (err)
 			goto unregister;
 
-		dst->remote_dev = remote_dev;
+		vxlan->lowerdev = remote_dev;
 	}
 
 	err = rtnl_configure_link(dev, NULL, 0, NULL);
@@ -4030,16 +4140,18 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 		goto unlink;
 
 	/* create an fdb entry for a valid default destination */
-	if (!vxlan_addr_any(&dst->remote_ip)) {
+	if (!vxlan_addr_any(&cfg->remote_ip)) {
+		union vxlan_addr rip = cfg->remote_ip;
+
 		spin_lock_bh(&vxlan->hash_lock);
 		err = vxlan_fdb_update(vxlan, all_zeros_mac,
-				       &dst->remote_ip,
+				       &rip,
 				       NUD_REACHABLE | NUD_PERMANENT,
 				       NLM_F_EXCL | NLM_F_CREATE,
-				       vxlan->cfg.dst_port,
-				       dst->remote_vni,
-				       dst->remote_vni,
-				       dst->remote_ifindex,
+				       cfg->dst_port,
+				       cfg->vni,
+				       cfg->vni,
+				       cfg->remote_ifindex,
 				       NTF_SELF, 0, true, extack);
 		spin_unlock_bh(&vxlan->hash_lock);
 		if (err)
@@ -4100,8 +4212,12 @@ static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],
 	memset(conf, 0, sizeof(*conf));
 
 	/* if changelink operation, start with old existing cfg */
-	if (changelink)
-		memcpy(conf, &vxlan->cfg, sizeof(*conf));
+	if (changelink) {
+		const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+
+		if (cfg)
+			memcpy(conf, cfg, sizeof(*conf));
+	}
 
 	if (data[IFLA_VXLAN_ID]) {
 		__be32 vni = cpu_to_be32(nla_get_u32(data[IFLA_VXLAN_ID]));
@@ -4447,17 +4563,19 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],
 			    struct nlattr *data[],
 			    struct netlink_ext_ack *extack)
 {
+	bool lowerdev_changed, rem_ip_changed, change_igmp;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	bool rem_ip_changed, change_igmp;
+	const struct vxlan_config *cfg;
+	struct vxlan_config *new_cfg;
 	struct net_device *lowerdev;
 	struct vxlan_config conf;
-	struct vxlan_rdst *dst;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	if (!rtnl_dev_link_net_capable(dev, vxlan->net))
 		return -EPERM;
 
-	dst = &vxlan->default_dst;
 	err = vxlan_nl2conf(tb, data, dev, &conf, true, extack);
 	if (err)
 		return err;
@@ -4467,76 +4585,108 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],
 	if (err)
 		return err;
 
-	if (dst->remote_dev == lowerdev)
-		lowerdev = NULL;
+	new_cfg = kmemdup(&conf, sizeof(conf), GFP_KERNEL);
+	if (!new_cfg)
+		return -ENOMEM;
 
-	err = netdev_adjacent_change_prepare(dst->remote_dev, lowerdev, dev,
-					     extack);
-	if (err)
-		return err;
+	lowerdev_changed = vxlan->lowerdev != lowerdev;
+	if (lowerdev_changed) {
+		err = netdev_adjacent_change_prepare(vxlan->lowerdev, lowerdev,
+						     dev, extack);
+		if (err) {
+			kfree(new_cfg);
+			return err;
+		}
+	}
 
-	rem_ip_changed = !vxlan_addr_equal(&conf.remote_ip, &dst->remote_ip);
+	rem_ip_changed = !vxlan_addr_equal(&conf.remote_ip, &cfg->remote_ip);
 	change_igmp = vxlan->dev->flags & IFF_UP &&
 		      (rem_ip_changed ||
-		       dst->remote_ifindex != conf.remote_ifindex);
+		       cfg->remote_ifindex != conf.remote_ifindex);
 
 	/* handle default dst entry */
-	if (rem_ip_changed) {
+	if (rem_ip_changed || cfg->remote_ifindex != conf.remote_ifindex) {
 		spin_lock_bh(&vxlan->hash_lock);
 		if (!vxlan_addr_any(&conf.remote_ip)) {
 			err = vxlan_fdb_update(vxlan, all_zeros_mac,
 					       &conf.remote_ip,
 					       NUD_REACHABLE | NUD_PERMANENT,
 					       NLM_F_APPEND | NLM_F_CREATE,
-					       vxlan->cfg.dst_port,
+					       cfg->dst_port,
 					       conf.vni, conf.vni,
 					       conf.remote_ifindex,
 					       NTF_SELF, 0, true, extack);
 			if (err) {
 				spin_unlock_bh(&vxlan->hash_lock);
-				netdev_adjacent_change_abort(dst->remote_dev,
-							     lowerdev, dev);
+				if (lowerdev_changed)
+					netdev_adjacent_change_abort(vxlan->lowerdev,
+								     lowerdev, dev);
+				kfree(new_cfg);
 				return err;
 			}
 		}
-		if (!vxlan_addr_any(&dst->remote_ip))
+		if (!vxlan_addr_any(&cfg->remote_ip))
 			__vxlan_fdb_delete(vxlan, all_zeros_mac,
-					   dst->remote_ip,
-					   vxlan->cfg.dst_port,
-					   dst->remote_vni,
-					   dst->remote_vni,
-					   dst->remote_ifindex,
+					   cfg->remote_ip,
+					   cfg->dst_port,
+					   cfg->vni,
+					   cfg->vni,
+					   cfg->remote_ifindex,
 					   true);
 		spin_unlock_bh(&vxlan->hash_lock);
 
-		/* If vni filtering device, also update fdb entries of
-		 * all vnis that were using default remote ip
+		/* If vni filtering device, also update default fdb entries of
+		 * all vnis
 		 */
-		if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
-			err = vxlan_vnilist_update_group(vxlan, &dst->remote_ip,
-							 &conf.remote_ip, extack);
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
+			err = vxlan_vnilist_update_group(vxlan, &cfg->remote_ip,
+							 &conf.remote_ip,
+							 cfg->remote_ifindex,
+							 conf.remote_ifindex,
+							 extack);
 			if (err) {
-				netdev_adjacent_change_abort(dst->remote_dev,
-							     lowerdev, dev);
+				vxlan_update_default_fdb_entry(vxlan, conf.vni,
+							       &conf.remote_ip,
+							       &cfg->remote_ip,
+							       conf.remote_ifindex,
+							       cfg->remote_ifindex,
+							       NULL);
+				if (lowerdev_changed)
+					netdev_adjacent_change_abort(vxlan->lowerdev,
+								     lowerdev, dev);
+				kfree(new_cfg);
 				return err;
 			}
 		}
 	}
 
-	if (change_igmp && vxlan_addr_multicast(&dst->remote_ip))
+	if (change_igmp &&
+	    (vxlan_addr_multicast(&cfg->remote_ip) ||
+	     (cfg->flags & VXLAN_F_VNIFILTER)))
 		err = vxlan_multicast_leave(vxlan);
 
-	if (netif_running(dev) && conf.age_interval != vxlan->cfg.age_interval)
+	if (netif_running(dev) && conf.age_interval != cfg->age_interval)
 		mod_timer(&vxlan->age_timer, jiffies);
 
-	netdev_adjacent_change_commit(dst->remote_dev, lowerdev, dev);
-	if (lowerdev && lowerdev != dst->remote_dev)
-		dst->remote_dev = lowerdev;
-	vxlan_config_apply(dev, &conf, lowerdev, vxlan->net, true);
+	if (lowerdev_changed) {
+		if (lowerdev)
+			netdev_adjacent_change_commit(vxlan->lowerdev, lowerdev,
+						      dev);
+		else
+			netdev_upper_dev_unlink(vxlan->lowerdev, dev);
+		vxlan->lowerdev = lowerdev;
+	}
+	vxlan_config_apply(dev, new_cfg, lowerdev, vxlan->net, true,
+			   lowerdev_changed);
+
+	if (change_igmp &&
+	    (vxlan_addr_multicast(&new_cfg->remote_ip) ||
+	     (new_cfg->flags & VXLAN_F_VNIFILTER))) {
+		int join_err = vxlan_multicast_join(vxlan);
 
-	if (!err && change_igmp &&
-	    vxlan_addr_multicast(&dst->remote_ip))
-		err = vxlan_multicast_join(vxlan);
+		if (join_err)
+			err = join_err;
+	}
 
 	return err;
 }
@@ -4550,8 +4700,8 @@ static void vxlan_dellink(struct net_device *dev, struct list_head *head)
 
 	list_del(&vxlan->next);
 	unregister_netdevice_queue(dev, head);
-	if (vxlan->default_dst.remote_dev)
-		netdev_upper_dev_unlink(vxlan->default_dst.remote_dev, dev);
+	if (vxlan->lowerdev)
+		netdev_upper_dev_unlink(vxlan->lowerdev, dev);
 }
 
 static size_t vxlan_get_size(const struct net_device *dev)
@@ -4595,110 +4745,114 @@ static size_t vxlan_get_size(const struct net_device *dev)
 static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
 {
 	const struct vxlan_dev *vxlan = netdev_priv(dev);
-	const struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct ifla_vxlan_port_range ports = {
-		.low =  htons(vxlan->cfg.port_min),
-		.high = htons(vxlan->cfg.port_max),
-	};
+	struct ifla_vxlan_port_range ports;
+	const struct vxlan_config *cfg;
+
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
 
-	if (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(dst->remote_vni)))
+	if (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(cfg->vni)))
 		goto nla_put_failure;
 
-	if (!vxlan_addr_any(&dst->remote_ip)) {
-		if (dst->remote_ip.sa.sa_family == AF_INET) {
+	if (!vxlan_addr_any(&cfg->remote_ip)) {
+		if (cfg->remote_ip.sa.sa_family == AF_INET) {
 			if (nla_put_in_addr(skb, IFLA_VXLAN_GROUP,
-					    dst->remote_ip.sin.sin_addr.s_addr))
+					    cfg->remote_ip.sin.sin_addr.s_addr))
 				goto nla_put_failure;
 #if IS_ENABLED(CONFIG_IPV6)
 		} else {
 			if (nla_put_in6_addr(skb, IFLA_VXLAN_GROUP6,
-					     &dst->remote_ip.sin6.sin6_addr))
+					     &cfg->remote_ip.sin6.sin6_addr))
 				goto nla_put_failure;
 #endif
 		}
 	}
 
-	if (dst->remote_ifindex && nla_put_u32(skb, IFLA_VXLAN_LINK, dst->remote_ifindex))
+	if (cfg->remote_ifindex && nla_put_u32(skb, IFLA_VXLAN_LINK, cfg->remote_ifindex))
 		goto nla_put_failure;
 
-	if (!vxlan_addr_any(&vxlan->cfg.saddr)) {
-		if (vxlan->cfg.saddr.sa.sa_family == AF_INET) {
+	if (!vxlan_addr_any(&cfg->saddr)) {
+		if (cfg->saddr.sa.sa_family == AF_INET) {
 			if (nla_put_in_addr(skb, IFLA_VXLAN_LOCAL,
-					    vxlan->cfg.saddr.sin.sin_addr.s_addr))
+					    cfg->saddr.sin.sin_addr.s_addr))
 				goto nla_put_failure;
 #if IS_ENABLED(CONFIG_IPV6)
 		} else {
 			if (nla_put_in6_addr(skb, IFLA_VXLAN_LOCAL6,
-					     &vxlan->cfg.saddr.sin6.sin6_addr))
+					     &cfg->saddr.sin6.sin6_addr))
 				goto nla_put_failure;
 #endif
 		}
 	}
 
-	if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->cfg.ttl) ||
+	if (nla_put_u8(skb, IFLA_VXLAN_TTL, cfg->ttl) ||
 	    nla_put_u8(skb, IFLA_VXLAN_TTL_INHERIT,
-		       !!(vxlan->cfg.flags & VXLAN_F_TTL_INHERIT)) ||
-	    nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->cfg.tos) ||
-	    nla_put_u8(skb, IFLA_VXLAN_DF, vxlan->cfg.df) ||
-	    nla_put_be32(skb, IFLA_VXLAN_LABEL, vxlan->cfg.label) ||
-	    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, vxlan->cfg.label_policy) ||
+		       !!(cfg->flags & VXLAN_F_TTL_INHERIT)) ||
+	    nla_put_u8(skb, IFLA_VXLAN_TOS, cfg->tos) ||
+	    nla_put_u8(skb, IFLA_VXLAN_DF, cfg->df) ||
+	    nla_put_be32(skb, IFLA_VXLAN_LABEL, cfg->label) ||
+	    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, cfg->label_policy) ||
 	    nla_put_u8(skb, IFLA_VXLAN_LEARNING,
-		       !!(vxlan->cfg.flags & VXLAN_F_LEARN)) ||
+		       !!(cfg->flags & VXLAN_F_LEARN)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_PROXY,
-		       !!(vxlan->cfg.flags & VXLAN_F_PROXY)) ||
+		       !!(cfg->flags & VXLAN_F_PROXY)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_RSC,
-		       !!(vxlan->cfg.flags & VXLAN_F_RSC)) ||
+		       !!(cfg->flags & VXLAN_F_RSC)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_L2MISS,
-		       !!(vxlan->cfg.flags & VXLAN_F_L2MISS)) ||
+		       !!(cfg->flags & VXLAN_F_L2MISS)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_L3MISS,
-		       !!(vxlan->cfg.flags & VXLAN_F_L3MISS)) ||
+		       !!(cfg->flags & VXLAN_F_L3MISS)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_COLLECT_METADATA,
-		       !!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA)) ||
-	    nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan->cfg.age_interval) ||
-	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan->cfg.addrmax) ||
-	    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan->cfg.dst_port) ||
+		       !!(cfg->flags & VXLAN_F_COLLECT_METADATA)) ||
+	    nla_put_u32(skb, IFLA_VXLAN_AGEING, cfg->age_interval) ||
+	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, cfg->addrmax) ||
+	    nla_put_be16(skb, IFLA_VXLAN_PORT, cfg->dst_port) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_CSUM,
-		       !(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM_TX)) ||
+		       !(cfg->flags & VXLAN_F_UDP_ZERO_CSUM_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_TX,
-		       !!(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
+		       !!(cfg->flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_RX,
-		       !!(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM6_RX)) ||
+		       !!(cfg->flags & VXLAN_F_UDP_ZERO_CSUM6_RX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_TX,
-		       !!(vxlan->cfg.flags & VXLAN_F_REMCSUM_TX)) ||
+		       !!(cfg->flags & VXLAN_F_REMCSUM_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_RX,
-		       !!(vxlan->cfg.flags & VXLAN_F_REMCSUM_RX)) ||
+		       !!(cfg->flags & VXLAN_F_REMCSUM_RX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_LOCALBYPASS,
-		       !!(vxlan->cfg.flags & VXLAN_F_LOCALBYPASS)))
+		       !!(cfg->flags & VXLAN_F_LOCALBYPASS)))
 		goto nla_put_failure;
 
+	ports.low = htons(cfg->port_min);
+	ports.high = htons(cfg->port_max);
 	if (nla_put(skb, IFLA_VXLAN_PORT_RANGE, sizeof(ports), &ports))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_GBP &&
+	if (cfg->flags & VXLAN_F_GBP &&
 	    nla_put_flag(skb, IFLA_VXLAN_GBP))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_GPE &&
+	if (cfg->flags & VXLAN_F_GPE &&
 	    nla_put_flag(skb, IFLA_VXLAN_GPE))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_REMCSUM_NOPARTIAL &&
+	if (cfg->flags & VXLAN_F_REMCSUM_NOPARTIAL &&
 	    nla_put_flag(skb, IFLA_VXLAN_REMCSUM_NOPARTIAL))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER &&
+	if (cfg->flags & VXLAN_F_VNIFILTER &&
 	    nla_put_u8(skb, IFLA_VXLAN_VNIFILTER,
-		       !!(vxlan->cfg.flags & VXLAN_F_VNIFILTER)))
+		       !!(cfg->flags & VXLAN_F_VNIFILTER)))
 		goto nla_put_failure;
 
 	if (nla_put(skb, IFLA_VXLAN_RESERVED_BITS,
-		    sizeof(vxlan->cfg.reserved_bits),
-		    &vxlan->cfg.reserved_bits))
+		    sizeof(cfg->reserved_bits),
+		    &cfg->reserved_bits))
 		goto nla_put_failure;
 
+	rcu_read_unlock();
 	return 0;
 
 nla_put_failure:
+	rcu_read_unlock();
 	return -EMSGSIZE;
 }
 
@@ -4731,7 +4885,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,
 	LIST_HEAD(list_kill);
 
 	list_for_each_entry_safe(vxlan, next, &vn->vxlan_list, next) {
-		struct vxlan_rdst *dst = &vxlan->default_dst;
+		const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 
 		/* In case we created vxlan device with carrier
 		 * and we loose the carrier due to module unload
@@ -4739,7 +4893,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,
 		 * cases, it's not necessary and remote_ifindex
 		 * is 0 here, so no matches.
 		 */
-		if (dst->remote_ifindex == dev->ifindex)
+		if (cfg->remote_ifindex == dev->ifindex)
 			vxlan_dellink(vxlan->dev, &list_kill);
 	}
 
@@ -4752,6 +4906,10 @@ static int vxlan_netdevice_event(struct notifier_block *unused,
 	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
 
+	if ((event == NETDEV_REGISTER || event == NETDEV_UNREGISTER) &&
+	    netif_is_vxlan(dev))
+		vxlan_vnifilter_seq_inc(dev_net(dev));
+
 	if (event == NETDEV_UNREGISTER)
 		vxlan_handle_lowerdev_unregister(vn, dev);
 	else if (event == NETDEV_UDP_TUNNEL_PUSH_INFO)
diff --git a/drivers/net/vxlan/vxlan_mdb.c b/drivers/net/vxlan/vxlan_mdb.c
index 841f42ffecb9a..c1a0551990555 100644
--- a/drivers/net/vxlan/vxlan_mdb.c
+++ b/drivers/net/vxlan/vxlan_mdb.c
@@ -165,6 +165,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 				     const struct vxlan_mdb_entry *mdb_entry,
 				     const struct vxlan_mdb_remote *remote)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	struct vxlan_rdst *rd = rtnl_dereference(remote->rd);
 	struct br_mdb_entry e;
 	struct nlattr *nest;
@@ -189,12 +190,12 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 	    vxlan_nla_put_addr(skb, MDBA_MDB_EATTR_DST, &rd->remote_ip))
 		goto nest_err;
 
-	if (rd->remote_port && rd->remote_port != vxlan->cfg.dst_port &&
+	if (rd->remote_port && rd->remote_port != cfg->dst_port &&
 	    nla_put_u16(skb, MDBA_MDB_EATTR_DST_PORT,
 			be16_to_cpu(rd->remote_port)))
 		goto nest_err;
 
-	if (rd->remote_vni != vxlan->default_dst.remote_vni &&
+	if (rd->remote_vni != cfg->vni &&
 	    nla_put_u32(skb, MDBA_MDB_EATTR_VNI, be32_to_cpu(rd->remote_vni)))
 		goto nest_err;
 
@@ -202,7 +203,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 	    nla_put_u32(skb, MDBA_MDB_EATTR_IFINDEX, rd->remote_ifindex))
 		goto nest_err;
 
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) &&
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) &&
 	    mdb_entry->key.vni && nla_put_u32(skb, MDBA_MDB_EATTR_SRC_VNI,
 					      be32_to_cpu(mdb_entry->key.vni)))
 		goto nest_err;
@@ -613,16 +614,19 @@ static int vxlan_mdb_config_init(struct vxlan_mdb_config *cfg,
 {
 	struct br_mdb_entry *entry = nla_data(tb[MDBA_SET_ENTRY]);
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *vcfg;
+
+	vcfg = rtnl_dereference(vxlan->cfg);
 
 	memset(cfg, 0, sizeof(*cfg));
 	cfg->vxlan = vxlan;
-	cfg->group.vni = vxlan->default_dst.remote_vni;
+	cfg->group.vni = vcfg->vni;
 	INIT_LIST_HEAD(&cfg->src_list);
 	cfg->nlflags = nlmsg_flags;
 	cfg->filter_mode = MCAST_EXCLUDE;
 	cfg->rt_protocol = RTPROT_STATIC;
-	cfg->remote_vni = vxlan->default_dst.remote_vni;
-	cfg->remote_port = vxlan->cfg.dst_port;
+	cfg->remote_vni = vcfg->vni;
+	cfg->remote_port = vcfg->dst_port;
 
 	if (entry->ifindex != dev->ifindex) {
 		NL_SET_ERR_MSG_MOD(extack, "Port net device must be the VXLAN net device");
@@ -957,11 +961,12 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,
 			    const struct vxlan_mdb_entry *mdb_entry,
 			    const struct vxlan_mdb_remote *remote)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	const struct vxlan_mdb_entry_key *group = &mdb_entry->key;
 	struct vxlan_rdst *rd = rtnl_dereference(remote->rd);
 	size_t nlmsg_size;
 
-		     /* MDBA_MDB_ENTRY_INFO */
+	/* MDBA_MDB_ENTRY_INFO */
 	nlmsg_size = nla_total_size(sizeof(struct br_mdb_entry)) +
 		     /* MDBA_MDB_EATTR_TIMER */
 		     nla_total_size(sizeof(u32));
@@ -978,16 +983,16 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,
 	/* MDBA_MDB_EATTR_DST */
 	nlmsg_size += nla_total_size(vxlan_addr_size(&rd->remote_ip));
 	/* MDBA_MDB_EATTR_DST_PORT */
-	if (rd->remote_port && rd->remote_port != vxlan->cfg.dst_port)
+	if (rd->remote_port && rd->remote_port != cfg->dst_port)
 		nlmsg_size += nla_total_size(sizeof(u16));
 	/* MDBA_MDB_EATTR_VNI */
-	if (rd->remote_vni != vxlan->default_dst.remote_vni)
+	if (rd->remote_vni != cfg->vni)
 		nlmsg_size += nla_total_size(sizeof(u32));
 	/* MDBA_MDB_EATTR_IFINDEX */
 	if (rd->remote_ifindex)
 		nlmsg_size += nla_total_size(sizeof(u32));
 	/* MDBA_MDB_EATTR_SRC_VNI */
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) && group->vni)
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) && group->vni)
 		nlmsg_size += nla_total_size(sizeof(u32));
 
 	return nlmsg_size;
@@ -1216,7 +1221,7 @@ vxlan_mdb_entry_get(struct vxlan_dev *vxlan,
 		goto err_free_entry;
 
 	if (hlist_is_singular_node(&mdb_entry->mdb_node, &vxlan->mdb_list))
-		vxlan->cfg.flags |= VXLAN_F_MDB;
+		set_bit(VXLAN_DEV_F_MDB, &vxlan->flags);
 
 	return mdb_entry;
 
@@ -1233,7 +1238,7 @@ static void vxlan_mdb_entry_put(struct vxlan_dev *vxlan,
 		return;
 
 	if (hlist_is_singular_node(&mdb_entry->mdb_node, &vxlan->mdb_list))
-		vxlan->cfg.flags &= ~VXLAN_F_MDB;
+		clear_bit(VXLAN_DEV_F_MDB, &vxlan->flags);
 
 	rhashtable_remove_fast(&vxlan->mdb_tbl, &mdb_entry->rhnode,
 			       vxlan_mdb_rht_params);
@@ -1483,11 +1488,13 @@ static int vxlan_mdb_get_parse(struct net_device *dev, struct nlattr *tb[],
 {
 	struct br_mdb_entry *entry = nla_data(tb[MDBA_GET_ENTRY]);
 	struct nlattr *mdbe_attrs[MDBE_ATTR_MAX + 1];
+	const struct vxlan_config *cfg;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
 	memset(group, 0, sizeof(*group));
-	group->vni = vxlan->default_dst.remote_vni;
+	group->vni = cfg->vni;
 
 	if (!tb[MDBA_GET_ENTRY_ATTRS]) {
 		vxlan_mdb_group_set(group, entry, NULL);
@@ -1621,6 +1628,7 @@ int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,
 }
 
 struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
+						const struct vxlan_config *cfg,
 						struct sk_buff *skb,
 						__be32 src_vni)
 {
@@ -1634,8 +1642,8 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
 	/* When not in collect metadata mode, 'src_vni' is zero, but MDB
 	 * entries are stored with the VNI of the VXLAN device.
 	 */
-	if (!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA))
-		src_vni = vxlan->default_dst.remote_vni;
+	if (!(cfg->flags & VXLAN_F_COLLECT_METADATA))
+		src_vni = cfg->vni;
 
 	memset(&group, 0, sizeof(group));
 	group.vni = src_vni;
@@ -1700,6 +1708,7 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
 }
 
 netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg,
 			   const struct vxlan_mdb_entry *mdb_entry,
 			   struct sk_buff *skb)
 {
@@ -1721,12 +1730,12 @@ netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
 
 		skb1 = skb_clone(skb, GFP_ATOMIC);
 		if (skb1)
-			vxlan_xmit_one(skb1, vxlan->dev, src_vni,
+			vxlan_xmit_one(skb1, vxlan->dev, cfg, src_vni,
 				       rcu_dereference(remote->rd), false);
 	}
 
 	if (fremote)
-		vxlan_xmit_one(skb, vxlan->dev, src_vni,
+		vxlan_xmit_one(skb, vxlan->dev, cfg, src_vni,
 			       rcu_dereference(fremote->rd), false);
 	else
 		kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
@@ -1757,7 +1766,7 @@ void vxlan_mdb_fini(struct vxlan_dev *vxlan)
 	struct vxlan_mdb_flush_desc desc = {};
 
 	vxlan_mdb_flush(vxlan, &desc);
-	WARN_ON_ONCE(vxlan->cfg.flags & VXLAN_F_MDB);
+	WARN_ON_ONCE(test_bit(VXLAN_DEV_F_MDB, &vxlan->flags));
 	rhashtable_free_and_destroy(&vxlan->mdb_tbl, vxlan_mdb_check_empty,
 				    NULL);
 }
diff --git a/drivers/net/vxlan/vxlan_multicast.c b/drivers/net/vxlan/vxlan_multicast.c
index 3b75b48dc726d..a5ef6f66f00d4 100644
--- a/drivers/net/vxlan/vxlan_multicast.c
+++ b/drivers/net/vxlan/vxlan_multicast.c
@@ -14,11 +14,12 @@
 /* Update multicast group membership when first VNI on
  * multicast address is brought up
  */
-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		    int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &vxlan->default_dst.remote_ip);
-	int ifindex = (rifindex ? : vxlan->default_dst.remote_ifindex);
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *ip = (rip ? : &cfg->remote_ip);
+	int ifindex = (rifindex ? : cfg->remote_ifindex);
 	int ret = -EINVAL;
 	struct sock *sk;
 
@@ -47,11 +48,12 @@ int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
 	return ret;
 }
 
-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		     int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &vxlan->default_dst.remote_ip);
-	int ifindex = (rifindex ? : vxlan->default_dst.remote_ifindex);
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *ip = (rip ? : &cfg->remote_ip);
+	int ifindex = (rifindex ? : cfg->remote_ifindex);
 	int ret = -EINVAL;
 	struct sock *sk;
 
@@ -80,8 +82,8 @@ int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
 	return ret;
 }
 
-static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,
-				   union vxlan_addr *rip, int rifindex)
+static bool vxlan_group_used_match(const union vxlan_addr *ip, int ifindex,
+				   const union vxlan_addr *rip, int rifindex)
 {
 	if (!vxlan_addr_multicast(rip))
 		return false;
@@ -96,14 +98,16 @@ static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,
 }
 
 static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
-					  union vxlan_addr *ip, int ifindex)
+					  const struct vxlan_config *cfg,
+					  const union vxlan_addr *ip,
+					  int ifindex)
 {
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
 	struct vxlan_vni_node *v, *tmp;
 
 	if (vxlan_group_used_match(ip, ifindex,
-				   &vxlan->default_dst.remote_ip,
-				   vxlan->default_dst.remote_ifindex))
+				   &cfg->remote_ip,
+				   cfg->remote_ifindex))
 		return true;
 
 	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
@@ -112,7 +116,7 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
 
 		if (vxlan_group_used_match(ip, ifindex,
 					   &v->remote_ip,
-					   vxlan->default_dst.remote_ifindex))
+					   cfg->remote_ifindex))
 			return true;
 	}
 
@@ -121,16 +125,17 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
 
 /* See if multicast group is already in use by other ID */
 bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
-		      __be32 vni, union vxlan_addr *rip, int rifindex)
+		      __be32 vni, const union vxlan_addr *rip, int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &dev->default_dst.remote_ip);
-	int ifindex = (rifindex ? : dev->default_dst.remote_ifindex);
+	const struct vxlan_config *dev_cfg = rtnl_dereference(dev->cfg);
+	const union vxlan_addr *ip = (rip ? : &dev_cfg->remote_ip);
+	int ifindex = (rifindex ? : dev_cfg->remote_ifindex);
 	struct vxlan_dev *vxlan;
 	struct vxlan_sock *sock4;
 #if IS_ENABLED(CONFIG_IPV6)
 	struct vxlan_sock *sock6;
 #endif
-	unsigned short family = dev->default_dst.remote_ip.sa.sa_family;
+	unsigned short family = dev_cfg->remote_ip.sa.sa_family;
 
 	sock4 = rtnl_dereference(dev->vn4_sock);
 
@@ -147,6 +152,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 #endif
 
 	list_for_each_entry(vxlan, &vn->vxlan_list, next) {
+		const struct vxlan_config *cfg;
+
 		if (!netif_running(vxlan->dev) || vxlan == dev)
 			continue;
 
@@ -158,13 +165,15 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 		    rtnl_dereference(vxlan->vn6_sock) != sock6)
 			continue;
 #endif
-		if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
-			if (!vxlan_group_used_by_vnifilter(vxlan, ip, ifindex))
+		cfg = rtnl_dereference(vxlan->cfg);
+
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
+			if (!vxlan_group_used_by_vnifilter(vxlan, cfg, ip, ifindex))
 				continue;
 		} else {
 			if (!vxlan_group_used_match(ip, ifindex,
-						    &vxlan->default_dst.remote_ip,
-						    vxlan->default_dst.remote_ifindex))
+						    &cfg->remote_ip,
+						    cfg->remote_ifindex))
 				continue;
 		}
 
@@ -174,7 +183,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 	return false;
 }
 
-static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
+static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan,
+				       const struct vxlan_config *cfg)
 {
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
 	struct vxlan_vni_node *v, *tmp, *vgood = NULL;
@@ -185,7 +195,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 			continue;
 		/* skip if address is same as default address */
 		if (vxlan_addr_equal(&v->remote_ip,
-				     &vxlan->default_dst.remote_ip))
+				     &cfg->remote_ip))
 			continue;
 		ret = vxlan_igmp_join(vxlan, &v->remote_ip, 0);
 		if (ret == -EADDRINUSE)
@@ -200,7 +210,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 			if (!vxlan_addr_multicast(&v->remote_ip))
 				continue;
 			if (vxlan_addr_equal(&v->remote_ip,
-					     &vxlan->default_dst.remote_ip))
+					     &cfg->remote_ip))
 				continue;
 			vxlan_igmp_leave(vxlan, &v->remote_ip, 0);
 			if (v == vgood)
@@ -211,7 +221,8 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 	return ret;
 }
 
-static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
+static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan,
+					const struct vxlan_config *cfg)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
@@ -219,10 +230,17 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
 	int last_err = 0, ret;
 
 	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
-		if (vxlan_addr_multicast(&v->remote_ip) &&
-		    !vxlan_group_used(vn, vxlan, v->vni, &v->remote_ip,
+		if (!vxlan_addr_multicast(&v->remote_ip))
+			continue;
+		/* skip if address is same as default address */
+		if (vxlan_addr_equal(&v->remote_ip,
+				     &cfg->remote_ip))
+			continue;
+		if (!vxlan_group_used(vn, vxlan, v->vni, &v->remote_ip,
 				      0)) {
 			ret = vxlan_igmp_leave(vxlan, &v->remote_ip, 0);
+			if (ret == -EADDRNOTAVAIL)
+				ret = 0;
 			if (ret)
 				last_err = ret;
 		}
@@ -233,38 +251,42 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
 
 int vxlan_multicast_join(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	int ret = 0;
 
-	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip)) {
-		ret = vxlan_igmp_join(vxlan, &vxlan->default_dst.remote_ip,
-				      vxlan->default_dst.remote_ifindex);
+	if (vxlan_addr_multicast(&cfg->remote_ip)) {
+		ret = vxlan_igmp_join(vxlan, &cfg->remote_ip,
+				      cfg->remote_ifindex);
 		if (ret == -EADDRINUSE)
 			ret = 0;
 		if (ret)
 			return ret;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
-		return vxlan_multicast_join_vnigrp(vxlan);
+	if (cfg->flags & VXLAN_F_VNIFILTER)
+		return vxlan_multicast_join_vnigrp(vxlan, cfg);
 
 	return 0;
 }
 
 int vxlan_multicast_leave(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
 	int ret = 0;
 
-	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip) &&
+	if (vxlan_addr_multicast(&cfg->remote_ip) &&
 	    !vxlan_group_used(vn, vxlan, 0, NULL, 0)) {
-		ret = vxlan_igmp_leave(vxlan, &vxlan->default_dst.remote_ip,
-				       vxlan->default_dst.remote_ifindex);
+		ret = vxlan_igmp_leave(vxlan, &cfg->remote_ip,
+				       cfg->remote_ifindex);
+		if (ret == -EADDRNOTAVAIL)
+			ret = 0;
 		if (ret)
 			return ret;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
-		return vxlan_multicast_leave_vnigrp(vxlan);
+	if (cfg->flags & VXLAN_F_VNIFILTER)
+		return vxlan_multicast_leave_vnigrp(vxlan, cfg);
 
 	return 0;
 }
diff --git a/drivers/net/vxlan/vxlan_private.h b/drivers/net/vxlan/vxlan_private.h
index b1eec22163608..f0ee85f83732a 100644
--- a/drivers/net/vxlan/vxlan_private.h
+++ b/drivers/net/vxlan/vxlan_private.h
@@ -22,6 +22,8 @@ struct vxlan_net {
 	/* sock_list is protected by rtnl lock */
 	struct hlist_head sock_list[PORT_HASH_SIZE];
 	struct notifier_block nexthop_notifier_block;
+	/* Generation counter for RTM_GETTUNNEL dumps */
+	atomic_t vnifilter_seq;
 };
 
 struct vxlan_fdb_key {
@@ -177,6 +179,13 @@ vxlan_vnifilter_lookup(struct vxlan_dev *vxlan, __be32 vni)
 				      vxlan_vni_rht_params);
 }
 
+static inline void vxlan_vnifilter_seq_inc(const struct net *net)
+{
+	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
+
+	atomic_inc(&vn->vnifilter_seq);
+}
+
 /* vxlan_core.c */
 int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		     const u8 *mac, union vxlan_addr *ip,
@@ -195,9 +204,10 @@ int vxlan_fdb_update(struct vxlan_dev *vxlan,
 		     __u32 ifindex, __u16 ndm_flags, u32 nhid,
 		     bool swdev_notify, struct netlink_ext_ack *extack);
 void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
+		    const struct vxlan_config *cfg,
 		    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc);
 int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,
-		     struct vxlan_config *conf, __be32 vni);
+		     const struct vxlan_config *conf, __be32 vni);
 
 /* vxlan_vnifilter.c */
 int vxlan_vnigroup_init(struct vxlan_dev *vxlan);
@@ -205,7 +215,8 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan);
 
 int vxlan_vnifilter_init(void);
 void vxlan_vnifilter_uninit(void);
-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,
+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg, __be32 vni,
 			   struct vxlan_vni_node *vninode,
 			   int type, unsigned int len);
 
@@ -213,9 +224,16 @@ void vxlan_vs_add_vnigrp(struct vxlan_dev *vxlan,
 			 struct vxlan_sock *vs,
 			 bool ipv6);
 void vxlan_vs_del_vnigrp(struct vxlan_dev *vxlan);
+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan);
+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
+				   const union vxlan_addr *old_remote_ip,
+				   const union vxlan_addr *remote_ip,
+				   u32 old_ifindex, u32 new_ifindex,
+				   struct netlink_ext_ack *extack);
 int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
-			       union vxlan_addr *old_remote_ip,
-			       union vxlan_addr *new_remote_ip,
+			       const union vxlan_addr *old_remote_ip,
+			       const union vxlan_addr *new_remote_ip,
+			       u32 old_ifindex, u32 new_ifindex,
 			       struct netlink_ext_ack *extack);
 
 
@@ -223,10 +241,10 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
 int vxlan_multicast_join(struct vxlan_dev *vxlan);
 int vxlan_multicast_leave(struct vxlan_dev *vxlan);
 bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
-		      __be32 vni, union vxlan_addr *rip, int rifindex);
-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+		      __be32 vni, const union vxlan_addr *rip, int rifindex);
+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		    int rifindex);
-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		     int rifindex);
 
 /* vxlan_mdb.c */
@@ -241,9 +259,11 @@ int vxlan_mdb_del_bulk(struct net_device *dev, struct nlattr *tb[],
 int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,
 		  u32 seq, struct netlink_ext_ack *extack);
 struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
+						const struct vxlan_config *cfg,
 						struct sk_buff *skb,
 						__be32 src_vni);
 netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg,
 			   const struct vxlan_mdb_entry *mdb_entry,
 			   struct sk_buff *skb);
 int vxlan_mdb_init(struct vxlan_dev *vxlan);
diff --git a/drivers/net/vxlan/vxlan_vnifilter.c b/drivers/net/vxlan/vxlan_vnifilter.c
index dd94085e08865..41095096b698d 100644
--- a/drivers/net/vxlan/vxlan_vnifilter.c
+++ b/drivers/net/vxlan/vxlan_vnifilter.c
@@ -171,13 +171,14 @@ static void vxlan_vnifilter_stats_add(struct vxlan_vni_node *vninode,
 	u64_stats_update_end(&pstats->syncp);
 }
 
-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,
+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg, __be32 vni,
 			   struct vxlan_vni_node *vninode,
 			   int type, unsigned int len)
 {
 	struct vxlan_vni_node *vnode;
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (!(cfg->flags & VXLAN_F_VNIFILTER))
 		return;
 
 	if (vninode) {
@@ -333,22 +334,36 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 				    struct sk_buff *skb,
 				    struct netlink_callback *cb)
 {
-	struct vxlan_vni_node *tmp, *v, *vbegin = NULL, *vend = NULL;
+	struct vxlan_vni_node *v, *vbegin = NULL, *vend = NULL;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct tunnel_msg *new_tmsg, *tmsg;
-	int idx = 0, s_idx = cb->args[1];
+	const struct vxlan_config *cfg;
 	struct vxlan_vni_group *vg;
 	struct nlmsghdr *nlh;
+	int idx = 0, s_idx;
 	bool dump_stats;
 	int err = 0;
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (cb->args[2] != dev->ifindex) {
+		cb->args[1] = 0;
+		cb->args[2] = dev->ifindex;
+	}
+	s_idx = cb->args[1];
+
+	cfg = rcu_dereference(vxlan->cfg);
+	if (!(cfg->flags & VXLAN_F_VNIFILTER)) {
+		cb->args[1] = 0;
+		cb->args[2] = 0;
 		return -EINVAL;
+	}
 
 	/* RCU needed because of the vni locking rules (rcu || rtnl) */
 	vg = rcu_dereference(vxlan->vnigrp);
-	if (!vg || !vg->num_vnis)
+	if (!vg || !vg->num_vnis) {
+		cb->args[1] = 0;
+		cb->args[2] = 0;
 		return 0;
+	}
 
 	tmsg = nlmsg_data(cb->nlh);
 	dump_stats = !!(tmsg->flags & TUNNEL_MSG_FLAG_STATS);
@@ -362,7 +377,7 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 	new_tmsg->family = PF_BRIDGE;
 	new_tmsg->ifindex = dev->ifindex;
 
-	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
+	list_for_each_entry_rcu(v, &vg->vni_list, vlist) {
 		if (idx < s_idx) {
 			idx++;
 			continue;
@@ -394,12 +409,26 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 	}
 
 	cb->args[1] = err ? idx : 0;
+	cb->args[2] = err ? dev->ifindex : 0;
 
 	nlmsg_end(skb, nlh);
 
+	nl_dump_check_consistent(cb, nlh);
+
 	return err;
 }
 
+static u32 vxlan_vnifilter_base_seq(const struct net *net)
+{
+	const struct vxlan_net *vn = net_generic(net, vxlan_net_id);
+	u32 res = atomic_read(&vn->vnifilter_seq);
+
+	/* Must not return 0 (see nl_dump_check_consistent()) */
+	if (!res)
+		res = 0x80000000;
+	return res;
+}
+
 static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	int idx = 0, err = 0, s_idx = cb->args[0];
@@ -419,6 +448,9 @@ static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb
 	}
 
 	rcu_read_lock();
+
+	cb->seq = vxlan_vnifilter_base_seq(net);
+
 	if (tmsg->ifindex) {
 		dev = dev_get_by_index_rcu(net, tmsg->ifindex);
 		if (!dev) {
@@ -470,26 +502,50 @@ static const struct nla_policy vni_filter_policy[VXLAN_VNIFILTER_MAX + 1] = {
 	[VXLAN_VNIFILTER_ENTRY] = { .type = NLA_NESTED },
 };
 
-static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
-					  union vxlan_addr *old_remote_ip,
-					  union vxlan_addr *remote_ip,
-					  struct netlink_ext_ack *extack)
+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan)
 {
-	struct vxlan_rdst *dst = &vxlan->default_dst;
+	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
+	struct vxlan_vni_node *v;
+
+	if (!vg)
+		return false;
+
+	list_for_each_entry(v, &vg->vni_list, vlist) {
+		if (vxlan_addr_multicast(&v->remote_ip))
+			return true;
+	}
+
+	return false;
+}
+
+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
+				   const union vxlan_addr *old_remote_ip,
+				   const union vxlan_addr *remote_ip,
+				   u32 old_ifindex, u32 new_ifindex,
+				   struct netlink_ext_ack *extack)
+{
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	int err = 0;
 
+	if (old_remote_ip && remote_ip &&
+	    vxlan_addr_equal(old_remote_ip, remote_ip) &&
+	    old_ifindex == new_ifindex)
+		return 0;
+
 	spin_lock_bh(&vxlan->hash_lock);
 	if (remote_ip && !vxlan_addr_any(remote_ip)) {
+		union vxlan_addr rip = *remote_ip;
+
 		err = vxlan_fdb_update(vxlan, all_zeros_mac,
-				       remote_ip,
+				       &rip,
 				       NUD_REACHABLE | NUD_PERMANENT,
 				       NLM_F_APPEND | NLM_F_CREATE,
-				       vxlan->cfg.dst_port,
+				       cfg->dst_port,
 				       vni,
 				       vni,
-				       dst->remote_ifindex,
+				       new_ifindex,
 				       NTF_SELF, 0, true, extack);
-		if (err) {
+		if (err && extack) {
 			spin_unlock_bh(&vxlan->hash_lock);
 			return err;
 		}
@@ -498,9 +554,9 @@ static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
 	if (old_remote_ip && !vxlan_addr_any(old_remote_ip)) {
 		__vxlan_fdb_delete(vxlan, all_zeros_mac,
 				   *old_remote_ip,
-				   vxlan->cfg.dst_port,
+				   cfg->dst_port,
 				   vni, vni,
-				   dst->remote_ifindex,
+				   old_ifindex,
 				   true);
 	}
 	spin_unlock_bh(&vxlan->hash_lock);
@@ -515,8 +571,8 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 				  struct netlink_ext_ack *extack)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	union vxlan_addr *newrip = NULL, *oldrip = NULL;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *newrip = NULL, *oldrip = NULL;
 	union vxlan_addr old_remote_ip;
 	int ret = 0;
 
@@ -528,15 +584,16 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 	if (group && !vxlan_addr_any(group)) {
 		newrip = group;
 	} else {
-		if (!vxlan_addr_any(&dst->remote_ip))
-			newrip = &dst->remote_ip;
+		if (!vxlan_addr_any(&cfg->remote_ip))
+			newrip = &cfg->remote_ip;
 	}
 
-	/* if old rip exists, and no newrip,
-	 * explicitly delete old rip
-	 */
-	if (!newrip && !vxlan_addr_any(&old_remote_ip))
-		oldrip = &old_remote_ip;
+	if (!create) {
+		if (!vxlan_addr_any(&old_remote_ip))
+			oldrip = &old_remote_ip;
+		else if (!vxlan_addr_any(&cfg->remote_ip))
+			oldrip = &cfg->remote_ip;
+	}
 
 	if (!newrip && !oldrip)
 		return 0;
@@ -546,20 +603,27 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 
 	ret = vxlan_update_default_fdb_entry(vxlan, vninode->vni,
 					     oldrip, newrip,
+					     cfg->remote_ifindex,
+					     cfg->remote_ifindex,
 					     extack);
 	if (ret)
 		goto out;
 
-	if (group)
+	if (group) {
 		memcpy(&vninode->remote_ip, group, sizeof(vninode->remote_ip));
+		if (!create)
+			vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
+	}
 
 	if (vxlan->dev->flags & IFF_UP) {
 		if (vxlan_addr_multicast(&old_remote_ip) &&
 		    !vxlan_group_used(vn, vxlan, vninode->vni,
 				      &old_remote_ip,
-				      vxlan->default_dst.remote_ifindex)) {
+				      cfg->remote_ifindex)) {
 			ret = vxlan_igmp_leave(vxlan, &old_remote_ip,
 					       0);
+			if (ret == -EADDRNOTAVAIL)
+				ret = 0;
 			if (ret)
 				goto out;
 		}
@@ -581,10 +645,12 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 }
 
 int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
-			       union vxlan_addr *old_remote_ip,
-			       union vxlan_addr *new_remote_ip,
+			       const union vxlan_addr *old_remote_ip,
+			       const union vxlan_addr *new_remote_ip,
+			       u32 old_ifindex, u32 new_ifindex,
 			       struct netlink_ext_ack *extack)
 {
+	const union vxlan_addr *oldrip, *newrip;
 	struct list_head *headp, *hpos;
 	struct vxlan_vni_group *vg;
 	struct vxlan_vni_node *vent;
@@ -595,37 +661,66 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
 	headp = &vg->vni_list;
 	list_for_each_prev(hpos, headp) {
 		vent = list_entry(hpos, struct vxlan_vni_node, vlist);
+
 		if (vxlan_addr_any(&vent->remote_ip)) {
-			ret = vxlan_update_default_fdb_entry(vxlan, vent->vni,
-							     old_remote_ip,
-							     new_remote_ip,
-							     extack);
-			if (ret)
-				return ret;
+			oldrip = old_remote_ip;
+			newrip = new_remote_ip;
+		} else {
+			/* A vni with its own group keeps it, but its fdb entry
+			 * is still keyed on the device remote_ifindex.
+			 */
+			oldrip = &vent->remote_ip;
+			newrip = &vent->remote_ip;
 		}
+
+		ret = vxlan_update_default_fdb_entry(vxlan, vent->vni,
+						     oldrip, newrip,
+						     old_ifindex, new_ifindex,
+						     extack);
+		if (ret)
+			goto err_unwind;
 	}
 
 	return 0;
+
+err_unwind:
+	list_for_each_continue(hpos, headp) {
+		vent = list_entry(hpos, struct vxlan_vni_node, vlist);
+
+		if (vxlan_addr_any(&vent->remote_ip)) {
+			oldrip = old_remote_ip;
+			newrip = new_remote_ip;
+		} else {
+			oldrip = &vent->remote_ip;
+			newrip = &vent->remote_ip;
+		}
+
+		vxlan_update_default_fdb_entry(vxlan, vent->vni,
+					       newrip, oldrip,
+					       new_ifindex, old_ifindex,
+					       NULL);
+	}
+	return ret;
 }
 
 static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,
 				   struct vxlan_vni_node *vninode)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 
 	/* if per vni remote_ip not present, delete the
 	 * default dst remote_ip previously added for this vni
 	 */
 	if (!vxlan_addr_any(&vninode->remote_ip) ||
-	    !vxlan_addr_any(&dst->remote_ip)) {
+	    !vxlan_addr_any(&cfg->remote_ip)) {
 		spin_lock_bh(&vxlan->hash_lock);
 		__vxlan_fdb_delete(vxlan, all_zeros_mac,
 				   (vxlan_addr_any(&vninode->remote_ip) ?
-				   dst->remote_ip : vninode->remote_ip),
-				   vxlan->cfg.dst_port,
+				   cfg->remote_ip : vninode->remote_ip),
+				   cfg->dst_port,
 				   vninode->vni, vninode->vni,
-				   dst->remote_ifindex,
+				   cfg->remote_ifindex,
 				   true);
 		spin_unlock_bh(&vxlan->hash_lock);
 	}
@@ -634,7 +729,7 @@ static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,
 		if (vxlan_addr_multicast(&vninode->remote_ip) &&
 		    !vxlan_group_used(vn, vxlan, vninode->vni,
 				      &vninode->remote_ip,
-				      dst->remote_ifindex)) {
+				      cfg->remote_ifindex)) {
 			vxlan_igmp_leave(vxlan, &vninode->remote_ip, 0);
 		}
 	}
@@ -665,7 +760,8 @@ static int vxlan_vni_update(struct vxlan_dev *vxlan,
 	return 0;
 }
 
-static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,
+static void __vxlan_vni_add_list(struct vxlan_dev *vxlan,
+				 struct vxlan_vni_group *vg,
 				 struct vxlan_vni_node *v)
 {
 	struct list_head *headp, *hpos;
@@ -681,13 +777,16 @@ static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,
 	}
 	list_add_rcu(&v->vlist, hpos);
 	vg->num_vnis++;
+	vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
 }
 
-static void __vxlan_vni_del_list(struct vxlan_vni_group *vg,
+static void __vxlan_vni_del_list(struct vxlan_dev *vxlan,
+				 struct vxlan_vni_group *vg,
 				 struct vxlan_vni_node *v)
 {
 	list_del_rcu(&v->vlist);
 	vg->num_vnis--;
+	vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
 }
 
 static struct vxlan_vni_node *vxlan_vni_alloc(struct vxlan_dev *vxlan,
@@ -723,6 +822,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 			 u32 vni, union vxlan_addr *group,
 			 struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_vni_node *vninode;
 	__be32 v = cpu_to_be32(vni);
 	bool changed = false;
@@ -731,7 +831,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 	if (vxlan_vnifilter_lookup(vxlan, v))
 		return vxlan_vni_update(vxlan, vg, v, group, &changed, extack);
 
-	err = vxlan_vni_in_use(vxlan->net, vxlan, &vxlan->cfg, v);
+	err = vxlan_vni_in_use(vxlan->net, vxlan, cfg, v);
 	if (err) {
 		NL_SET_ERR_MSG(extack, "VNI in use");
 		return err;
@@ -749,7 +849,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 		return err;
 	}
 
-	__vxlan_vni_add_list(vg, vninode);
+	__vxlan_vni_add_list(vxlan, vg, vninode);
 
 	if (vxlan->dev->flags & IFF_UP)
 		vxlan_vs_add_del_vninode(vxlan, vninode, false);
@@ -795,7 +895,7 @@ static int vxlan_vni_del(struct vxlan_dev *vxlan,
 	if (err)
 		goto out;
 
-	__vxlan_vni_del_list(vg, vninode);
+	__vxlan_vni_del_list(vxlan, vg, vninode);
 
 	vxlan_vnifilter_notify(vxlan, vninode, RTM_DELTUNNEL);
 
@@ -844,6 +944,7 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,
 				    int cmd, struct netlink_ext_ack *extack)
 {
 	struct nlattr *vattrs[VXLAN_VNIFILTER_ENTRY_MAX + 1];
+	const struct vxlan_config *cfg;
 	u32 vni_start = 0, vni_end = 0;
 	union vxlan_addr group;
 	int err;
@@ -881,7 +982,8 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,
 		memset(&group, 0, sizeof(group));
 	}
 
-	if (vxlan_addr_multicast(&group) && !vxlan->default_dst.remote_ifindex) {
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (vxlan_addr_multicast(&group) && !cfg->remote_ifindex) {
 		NL_SET_ERR_MSG(extack,
 			       "Local interface required for multicast remote group");
 
@@ -909,7 +1011,7 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan)
 #if IS_ENABLED(CONFIG_IPV6)
 		hlist_del_init_rcu(&v->hlist6.hlist);
 #endif
-		__vxlan_vni_del_list(vg, v);
+		__vxlan_vni_del_list(vxlan, vg, v);
 		vxlan_vnifilter_notify(vxlan, v, RTM_DELTUNNEL);
 		call_rcu(&v->rcu, vxlan_vni_node_rcu_free);
 	}
@@ -940,6 +1042,7 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,
 				   struct netlink_ext_ack *extack)
 {
 	struct net *net = sock_net(skb->sk);
+	const struct vxlan_config *cfg;
 	struct tunnel_msg *tmsg;
 	struct vxlan_dev *vxlan;
 	struct net_device *dev;
@@ -964,8 +1067,9 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,
 	}
 
 	vxlan = netdev_priv(dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (!(cfg->flags & VXLAN_F_VNIFILTER))
 		return -EOPNOTSUPP;
 
 	nlmsg_for_each_attr_type(attr, VXLAN_VNIFILTER_ENTRY, nlh,
diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index 7b82075055237..c3c9f2ccc3d1b 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -204,7 +204,6 @@ struct vxlan_rdst {
 	u8			 offloaded:1;
 	__be32			 remote_vni;
 	u32			 remote_ifindex;
-	struct net_device	 *remote_dev;
 	struct list_head	 list;
 	struct rcu_head		 rcu;
 	struct dst_cache	 dst_cache;
@@ -229,6 +228,7 @@ struct vxlan_config {
 	bool				no_share;
 	enum ifla_vxlan_df		df;
 	struct vxlanhdr			reserved_bits;
+	struct rcu_head			rcu;
 };
 
 enum {
@@ -294,14 +294,15 @@ struct vxlan_dev {
 #endif
 	struct net_device *dev;
 	struct net	  *net;		/* netns for packet i/o */
-	struct vxlan_rdst default_dst;	/* default destination */
+	struct net_device *lowerdev;
 
 	struct timer_list age_timer;
 	spinlock_t	  hash_lock;
 	unsigned int	  addrcnt;
 	struct gro_cells  gro_cells;
+	unsigned long	  flags;
 
-	struct vxlan_config	cfg;
+	struct vxlan_config __rcu	*cfg;
 
 	struct vxlan_vni_group  __rcu *vnigrp;
 
@@ -313,6 +314,10 @@ struct vxlan_dev {
 	unsigned int mdb_seq;
 };
 
+enum vxlan_dev_flags {
+	VXLAN_DEV_F_MDB,
+};
+
 #define VXLAN_F_LEARN			0x01
 #define VXLAN_F_PROXY			0x02
 #define VXLAN_F_RSC			0x04
@@ -331,7 +336,6 @@ struct vxlan_dev {
 #define VXLAN_F_IPV6_LINKLOCAL		0x8000
 #define VXLAN_F_TTL_INHERIT		0x10000
 #define VXLAN_F_VNIFILTER               0x20000
-#define VXLAN_F_MDB			0x40000
 #define VXLAN_F_LOCALBYPASS		0x80000
 #define VXLAN_F_MC_ROUTE		0x100000
 

3/2 2026/09/25 02:07 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 50242
							output: 27
							thoughts: 456
						

Thoughts:
**Analyzing the VXLAN Driver Changes**

Okay, so this patch is clearly making some significant changes to the VXLAN driver. I see it's touching several core files: `vxlan_core.c`, `vxlan_mdb.c`, `vxlan_multicast.c`, `vxlan_vnifilter.c`, and `include/net/vxlan.h`, plus some Mellanox driver files. My initial focus is on the VXLAN modifications themselves.

First, they've shifted `vxlan->cfg` from an embedded `struct vxlan_config` within `struct vxlan_dev` to an RCU-protected pointer, `struct vxlan_config __rcu *cfg`. That's a big deal. This is almost certainly aimed at improving concurrency and allowing for more efficient read/write access to the configuration data. I'll need to pay close attention to where they've added RCU dereferences. Speaking of which, the removal of `vxlan->default_dst` and its replacement with `vxlan->lowerdev` and accessing remote fields via `cfg` is interesting. It indicates a move towards potentially a more flexible or generalized configuration approach.

The addition of RCU dereferences across rx/tx/fdb/mdb paths is exactly what I expected, given the pointer change. I need to track down all the places they've added those calls. Memory allocation is also changed, they seem to be using `kmemdup` for `new_cfg` in `vxlan_dev_configure`, `vxlan_changelink`, and `kfree_rcu(old_cfg, rcu)` in `vxlan_config_apply`, with associated freeing in `vxlan_free_dev`. I need to ensure that the memory management is correct and doesn't introduce any new leaks or vulnerabilities. Also, they've updated the bit flag with `VXLAN_DEV_F_MDB` in `cfg->flags`, which replaces `VXLAN_F_MDB`. The change in the `vxlan_vnifilter_seq_inc` and the seq counter added for dumps also catches my eye.

Okay, let's step back for a moment. This is a pretty substantial set of changes. Before I dig deeper, I need to understand the impact in a fuzzing context.

Ah, good! VXLAN *is* reachable through standard QEMU fuzzing, which means the changes will be actively tested. Syzkaller regularly fuzzez this driver. The use of rtnetlink (`RTM_NEWLINK`, `RTM_SETLINK`, `RTM_DELLINK`, `RTM_GETLINK`), sockets, and VXLAN tx/rx packets means it's already well-covered, and I should expect good fuzzing coverage of these modifications.

But hold on, let me quickly verify the kernel configuration. I need to make sure that the required Kconfig options are enabled for this particular VXLAN setup. I'll check with `grepper` or by reading the necessary files. That's a must before moving on.



4/2 2026/09/25 02:07 tool grepper
0m Args:
{
  "Expression": "config VXLAN",
  "PathPrefix": "drivers/net/"
}

Results:
{
  "Output": "drivers/net/Kconfig=216=config IPVTAP\n--\ndrivers/net/Kconfig-229-\ndrivers/net/Kconfig:230:config VXLAN\ndrivers/net/Kconfig-231-\ttristate \"Virtual eXtensible Local Area Network (VXLAN)\"\n"
}

5/2 2026/09/25 02:07 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 51468
							output: 34
							thoughts: 29
						

Thoughts:
Here's a summary of my thought process, as requested:

**Investigating the VXLAN Configuration in the Kernel**

Okay, so I need to understand the dependencies for `CONFIG_VXLAN`. I'll start by checking out the `drivers/net/Kconfig` file. Specifically, I'm going to look around line 230 – that's usually where the relevant section for VXLAN is located. I'm assuming that the developers have kept the Kconfig file relatively well-organized and that the VXLAN configuration options are grouped near that line. My goal is to quickly pinpoint any other kernel features or modules that VXLAN relies upon, whether directly or indirectly, in order to get a full picture of the configuration implications.



6/2 2026/09/25 02:07 tool read-file
0m Args:
{
  "File": "drivers/net/Kconfig",
  "FirstLine": 228,
  "LineCount": 25
}

Results:
{
  "Contents": " 228:\t\t  will be called ipvtap.\n 229:\t\n 230:\tconfig VXLAN\n 231:\t\ttristate \"Virtual eXtensible Local Area Network (VXLAN)\"\n 232:\t\tdepends on INET\n 233:\t\tselect NET_UDP_TUNNEL\n 234:\t\tselect GRO_CELLS\n 235:\t\thelp\n 236:\t\t  This allows one to create vxlan virtual interfaces that provide\n 237:\t\t  Layer 2 Networks over Layer 3 Networks. VXLAN is often used\n 238:\t\t  to tunnel virtual network infrastructure in virtualized environments.\n 239:\t\t  For more information see:\n 240:\t\t    http://tools.ietf.org/html/draft-mahalingam-dutt-dcops-vxlan-02\n 241:\t\n 242:\t\t  To compile this driver as a module, choose M here: the module\n 243:\t\t  will be called vxlan.\n 244:\t\n 245:\tconfig GENEVE\n 246:\t\ttristate \"Generic Network Virtualization Encapsulation\"\n 247:\t\tdepends on INET\n 248:\t\tselect NET_UDP_TUNNEL\n 249:\t\tselect GRO_CELLS\n 250:\t\thelp\n 251:\t\t  This allows one to create geneve virtual interfaces that provide\n 252:\t\t  Layer 2 Networks over Layer 3 Networks. GENEVE is often used\n"
}

7/2 2026/09/25 02:07 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 52075
							output: 34
							thoughts: 85
						

Thoughts:
Okay, let's break this down.

**Investigating VXLAN Configuration Dependencies**

Alright, so `CONFIG_VXLAN` is clearly dependent on `INET`, that's the baseline. I need to make sure I understand the full configuration picture. I'm especially interested in IPv6 support here. Let me see if there are any conditional compilation paths in the VXLAN code that might use IPv6. Ah, good, I see `#if IS_ENABLED(CONFIG_IPV6)` used in the VXLAN files. That confirms it, enabling `CONFIG_IPV6` would be smart to ensure the VXLAN configuration covers all potential IPv6 scenarios. To make certain I have a complete understanding, I'll need to check the `drivers/net/vxlan/Makefile` as well, to see if there are any other configuration variables that are important for this functionality.



8/2 2026/09/25 02:07 tool read-file
0m Args:
{
  "File": "drivers/net/vxlan/Makefile",
  "FirstLine": 1,
  "LineCount": 10
}

Results:
{
  "Contents": "   1:\t#\n   2:\t# Makefile for the vxlan driver\n   3:\t#\n   4:\t\n   5:\tobj-$(CONFIG_VXLAN) += vxlan.o\n   6:\t\n   7:\tvxlan-objs := vxlan_core.o vxlan_multicast.o vxlan_vnifilter.o vxlan_mdb.o\n"
}

9/2 2026/09/25 02:07 llm patch-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 52479
							output: 283
							thoughts: 2533
						

Thoughts:
**VXLAN Driver Concurrency and Configuration Changes - Worthy of Fuzzing**

Okay, so looking at this patch, it's clear the VXLAN driver is undergoing some significant changes, specifically regarding how its configuration data is handled. My initial observation is that all the modified files belong to `CONFIG_VXLAN`, and there are also related changes in `mlx5` and `mlxsw` files that are adapting to these VXLAN changes. The fact that the core VXLAN driver is purely virtual and frequently fuzzed makes these changes even more interesting.

The major change here is the switch from an embedded `vxlan->cfg` to a dynamically allocated structure, protected by RCU. Specifically, `kmemdup` is now used for allocation, and `kfree_rcu` or `kfree` is used for freeing, which immediately sets off alarms regarding RCU lifetime, potential use-after-free conditions, race conditions, memory leaks, null pointer dereferences, and even rollback issues. There's an especially concerning point in `vxlan_free_dev`, where `kfree(cfg)` is called *immediately* after `RCU_INIT_POINTER(vxlan->cfg, NULL)`, which seems wrong. What if an RCU read-side is still dereferencing that config pointer?

The `vxlan_config_apply` and `vxlan_changelink` functions further reinforce these concerns, where the allocation, assignment using `rcu_assign_pointer`, and deferred freeing using `kfree_rcu` seem like obvious targets. And let's not forget the unwinding logic in `vxlan_vnilist_update_group` which needs to be verified. The change to `vxlan_xmit` now using `rcu_read_lock()` at the top, and numerous unlock exits is also heavily reworked concurrency/RCU protection, and worth extra attention.

Now, let's look at the modified functions, and see which ones are the best candidates to act as "FocusSymbols." We're looking for functions that specifically implement the logic being added or altered, or direct API entrypoints for the subsystem feature under review, and not for ubiquitous lifecycle hot-paths.

Here's the list of relevant functions:
*   `vxlan_changelink` - netlink config handling.
*   `vxlan_config_apply` - applies the configuration.
*   `vxlan_config_validate` - validates the configuration.
*   `vxlan_xmit` - the core transmit path.
*   `vxlan_rcv` - the core receive path.
*   `vxlan_vnifilter_dump_dev` - netlink vni filter dump.
*   `vxlan_vnilist_update_group` - handles group updates for vni filter.
*   `vxlan_update_default_fdb_entry` - default fdb updates.
*   `vxlan_mdb_entry_skb_get` and `vxlan_mdb_xmit` - MDB handling.
*   `vxlan_free_dev` - netdevice priv\_destructor.

I've also verified that these functions are actually visible in the symbol tables. They can be found in `vxlan_core.c`, `vxlan_vnifilter.c`, `vxlan_mdb.c` and `vxlan_multicast.c`

So, after careful consideration, the key FocusSymbols are: `vxlan_changelink`, `vxlan_config_apply`, `vxlan_dev_configure`, `vxlan_xmit`, `vxlan_rcv`, `vxlan_vnilist_update_group`, `vxlan_update_default_fdb_entry`, `vxlan_vnifilter_dump`, `vxlan_multicast_join`, `vxlan_multicast_leave` and, critically, `vxlan_free_dev`.

To summarize the requirements:
*   `EnableConfigs`: `["VXLAN", "INET", "IPV6"]` - These are needed for VXLAN functionality.
*   `FocusSymbols`: `["vxlan_changelink", "vxlan_config_apply", "vxlan_dev_configure", "vxlan_xmit", "vxlan_rcv", "vxlan_vnilist_update_group", "vxlan_update_default_fdb_entry", "vxlan_vnifilter_dump", "vxlan_multicast_join", "vxlan_multicast_leave", "vxlan_free_dev"]`
*   `Reasoning`: The patch represents a major shift in how VXLAN config is handled - from embedded to RCU-protected, dynamic allocation.


10/2 2026/09/25 02:07 tool set-results
0m Args:
{
  "EnableConfigs": [
    "VXLAN",
    "INET",
    "IPV6"
  ],
  "FocusSymbols": [
    "vxlan_changelink",
    "vxlan_config_apply",
    "vxlan_dev_configure",
    "vxlan_xmit",
    "vxlan_rcv",
    "vxlan_vnilist_update_group",
    "vxlan_update_default_fdb_entry",
    "vxlan_vnifilter_dump",
    "vxlan_multicast_join",
    "vxlan_multicast_leave",
    "vxlan_free_dev"
  ],
  "Reasoning": "The patch refactors the VXLAN driver by converting the embedded vxlan_config in struct vxlan_dev to a dynamically allocated RCU-protected pointer (vxlan-\u003ecfg). It updates device configuration, changelink, transmission (vxlan_xmit), reception (vxlan_rcv), multicast, and VNI-filter operations to access configuration under RCU, and adds memory management (kmemdup, kfree_rcu, and priv_destructor vxlan_free_dev) and rollback handling. These reachable networking changes introduce subtle lifetime and concurrency considerations that warrant fuzzing via netlink and packet I/O.",
  "WorthFuzzing": true
}

Results:
{
  "EnableConfigs": [
    "VXLAN",
    "INET",
    "IPV6"
  ],
  "FocusSymbols": [
    "vxlan_changelink",
    "vxlan_config_apply",
    "vxlan_dev_configure",
    "vxlan_xmit",
    "vxlan_rcv",
    "vxlan_vnilist_update_group",
    "vxlan_update_default_fdb_entry",
    "vxlan_vnifilter_dump",
    "vxlan_multicast_join",
    "vxlan_multicast_leave",
    "vxlan_free_dev"
  ],
  "Reasoning": "The patch refactors the VXLAN driver by converting the embedded vxlan_config in struct vxlan_dev to a dynamically allocated RCU-protected pointer (vxlan-\u003ecfg). It updates device configuration, changelink, transmission (vxlan_xmit), reception (vxlan_rcv), multicast, and VNI-filter operations to access configuration under RCU, and adds memory management (kmemdup, kfree_rcu, and priv_destructor vxlan_free_dev) and rollback handling. These reachable networking changes introduce subtle lifetime and concurrency considerations that warrant fuzzing via netlink and packet I/O.",
  "WorthFuzzing": true
}

11/1 2026/09/25 02:07 action If
2m Args:
{
  "WorthFuzzing": true
}

Results:
null

12/2 2026/09/25 02:07 agent kmsan-evaluator
2m Model:
core

Args:
null

Results:
{
  "KMSANReasoning": "The patch refactors vxlan device configuration (`vxlan-\u003ecfg`) from an embedded struct to a dynamically-allocated, RCU-protected pointer (`struct vxlan_config __rcu *cfg`) to allow atomic configuration updates under RCU. It also replaces `vxlan-\u003edefault_dst` with fields directly in `cfg` and adds RCU dereference helpers across vxlan core, mlx5, and mlxsw drivers.\n\nEvaluation of uninitialized memory risks:\n1. All dynamically allocated `struct vxlan_config` instances are created with `kmemdup` from a `struct vxlan_config conf` that is explicitly zero-initialized via `memset(conf, 0, sizeof(*conf))` in `vxlan_nl2conf()`.\n2. In `vxlan_fill_info()`, `struct ifla_vxlan_port_range ports` contains only two `__be16` fields (`low` and `high`) without padding, both of which are fully assigned before `nla_put()`.\n3. In `mlxsw_sp_bridge_vlan_aware_vxlan_join()`, `struct mlxsw_sp_nve_params params` has all its active fields explicitly assigned, and is solely consumed internally within the mlxsw driver without exposure to userspace.\n4. No uninitialized memory is read, copied to userspace, or branched upon.\n\nThe primary bug risks associated with this refactor involve RCU pointer lifetimes, use-after-free of old configuration objects during grace periods, and NULL pointer dereferences. These classes of bugs are thoroughly covered by KASAN, PROVE_RCU, and standard kernel debugging facilities, not KMSAN. Therefore, a dedicated KMSAN fuzzing session is not warranted.",
  "NeedsKMSAN": false
}

Instruction:
You are an expert Linux kernel security engineer specializing in kernel memory
error detectors (KASAN and KMSAN). Your job is to review the provided patch series and
determine if the code changes justify spawning a dedicated KMSAN (KernelMemorySanitizer)
fuzzing session in addition to standard KASAN fuzzing.

CRITICAL DISTINCTION BETWEEN KASAN AND KMSAN:
- Standard KASAN kernel builds (upstream-apparmor-kasan.config) already enable
  a comprehensive suite of debugging tools and sanitizers, including KASAN
  (out-of-bounds accesses, use-after-free, double free, invalid free), LOCKDEP
  (locking bugs and deadlocks), UB-sanitizers, and memory corruption checks.
- KMSAN (KernelMemorySanitizer) detects reads of UNINITIALIZED memory (stack, heap,
  or page allocations) and kernel-to-user memory info-leaks.

Rule: THERE IS NO SENSE IN RUNNING A KMSAN SESSION IF A BUG CAN BE CAUGHT BY KASAN,
LOCKDEP, OR OTHER STANDARD BUG DETECTORS.
A dedicated KMSAN fuzzing session incurs significant resource costs. You must ONLY
set NeedsKMSAN=true if the code changes introduce or expose UNINITIALIZED MEMORY risks
that are detected ONLY by KMSAN.

Look holistically at the patch series and surrounding code. Even if no direct
uninitialized field accesses or new buffer allocations are added in the diff itself,
a patch may alter control flow, bounds checking, or data length calculations in ways
that change how the rest of the code operates on existing buffers (e.g. allowing
uninitialized stack/heap memory to be read, copied to user space, or used in control
flow). Do not hesitate to use your code access tools to inspect the surrounding code,
called functions, and callers.

Set NeedsKMSAN=true ONLY IF the patch introduces or modifies:
1. Kernel structures sent to user space (via copy_to_user, put_user, netlink skb
   attributes, ioctl output arguments, socket options, or BPF buffers) where fields
   or structure padding might not be fully initialized/zeroed.
2. Conditional logic or branching that depends on potentially uninitialized variables
   or struct fields.
3. Allocation or initialization of complex data structures where uninitialized fields
   could be read later in reachable code paths.
4. Bounds checks, lengths, or logic in a way that allows surrounding code to access
   uninitialized bytes of existing buffers.

Set NeedsKMSAN=false IF:
- The code changes primarily risk out-of-bounds access, array overflows, NULL pointer
  dereferences, locking deadlocks, or use-after-free bugs (these are already caught
  by KASAN, LOCKDEP, or standard bug detectors).
- All stack/heap structures touched or introduced by the patch are fully zeroed
  or initialized (e.g. using = {0}, memset, kzalloc) before being read or copied.
- The patch does not introduce any risk of uninitialized memory usage or info-leaks.

Use your code access tools to inspect the surrounding code if necessary, then provide
detailed KMSANReasoning contrasting KASAN vs KMSAN applicability for this patch.
Prefer calling several tools at the same time to save round-trips.


Use set-results tool to provide results of the analysis.
It must be called exactly once before the final reply.
Ignore results of this tool.

Prompt:
Target architecture: amd64

For your convenience, here is the diff of the changes:
commit c6e7846ceee7f17a881c3bd9be86fc5863645aba
Author: syz-cluster <triage@syzkaller.com>
Date:   Fri Sep 25 02:07:19 2026 +0000

    syz-cluster: applied patch under review

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
index 7a18a469961db..467fbe43b89e9 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en/tc_tun_vxlan.c
@@ -241,9 +241,16 @@ static bool mlx5e_tc_tun_encap_info_equal_vxlan(struct mlx5e_encap_key *a,
 static int mlx5e_tc_tun_get_remote_ifindex(struct net_device *mirred_dev)
 {
 	const struct vxlan_dev *vxlan = netdev_priv(mirred_dev);
-	const struct vxlan_rdst *dst = &vxlan->default_dst;
+	const struct vxlan_config *cfg;
+	int ifindex = 0;
 
-	return dst->remote_ifindex;
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+	if (cfg)
+		ifindex = cfg->remote_ifindex;
+	rcu_read_unlock();
+
+	return ifindex;
 }
 
 struct mlx5e_tc_tunnel vxlan_tunnel = {
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
index 52c2fe3644d4b..21e7b499f3879 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c
@@ -59,8 +59,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,
 					   const struct mlxsw_sp_nve_params *params,
 					   struct netlink_ext_ack *extack)
 {
-	struct vxlan_dev *vxlan = netdev_priv(params->dev);
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
+	struct vxlan_dev *vxlan;
+
+	vxlan = netdev_priv(params->dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	if (vxlan_addr_multicast(&cfg->remote_ip)) {
 		NL_SET_ERR_MSG_MOD(extack, "VxLAN: Multicast destination IP is not supported");
@@ -92,6 +95,11 @@ static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,
 		return false;
 	}
 
+	if (test_bit(VXLAN_DEV_F_MDB, &vxlan->flags)) {
+		NL_SET_ERR_MSG_MOD(extack, "VxLAN: MDB entries are not supported");
+		return false;
+	}
+
 	switch (cfg->saddr.sa.sa_family) {
 	case AF_INET:
 		if (!mlxsw_sp_nve_vxlan_ipv4_flags_check(cfg, extack))
@@ -148,8 +156,11 @@ static void mlxsw_sp_nve_vxlan_config(const struct mlxsw_sp_nve *nve,
 				      const struct mlxsw_sp_nve_params *params,
 				      struct mlxsw_sp_nve_config *config)
 {
-	struct vxlan_dev *vxlan = netdev_priv(params->dev);
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
+	struct vxlan_dev *vxlan;
+
+	vxlan = netdev_priv(params->dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	config->type = MLXSW_SP_NVE_TYPE_VXLAN;
 	config->ttl = cfg->ttl;
diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index fe45e533a4b2e..de60b698bea98 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -2513,15 +2513,17 @@ mlxsw_sp_bridge_vlan_aware_vxlan_join(struct mlxsw_sp_bridge_device *bridge_devi
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device->dev);
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	struct mlxsw_sp_nve_params params = {
-		.type = MLXSW_SP_NVE_TYPE_VXLAN,
-		.vni = vxlan->cfg.vni,
-		.dev = vxlan_dev,
-		.ethertype = ethertype,
-	};
+	struct mlxsw_sp_nve_params params;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	params.type = MLXSW_SP_NVE_TYPE_VXLAN;
+	params.vni = cfg->vni;
+	params.dev = vxlan_dev;
+	params.ethertype = ethertype;
+
 	/* If the VLAN is 0, we need to find the VLAN that is configured as
 	 * PVID and egress untagged on the bridge port of the VxLAN device.
 	 * It is possible no such VLAN exists
@@ -2704,15 +2706,17 @@ mlxsw_sp_bridge_8021d_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,
 {
 	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device->dev);
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	struct mlxsw_sp_nve_params params = {
-		.type = MLXSW_SP_NVE_TYPE_VXLAN,
-		.vni = vxlan->cfg.vni,
-		.dev = vxlan_dev,
-		.ethertype = ETH_P_8021Q,
-	};
+	struct mlxsw_sp_nve_params params;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	params.type = MLXSW_SP_NVE_TYPE_VXLAN;
+	params.vni = cfg->vni;
+	params.dev = vxlan_dev;
+	params.ethertype = ETH_P_8021Q;
+
 	fid = mlxsw_sp_fid_8021d_get(mlxsw_sp, bridge_device->dev->ifindex);
 	if (IS_ERR(fid)) {
 		NL_SET_ERR_MSG_MOD(extack, "Failed to create 802.1D FID");
@@ -2933,10 +2937,13 @@ static void __mlxsw_sp_bridge_vxlan_leave(struct mlxsw_sp *mlxsw_sp,
 					  const struct net_device *vxlan_dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	/* If the VxLAN device is down, then the FID does not have a VNI */
-	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vxlan->cfg.vni);
+	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, cfg->vni);
 	if (!fid)
 		return;
 
@@ -3029,11 +3036,13 @@ static void mlxsw_sp_fdb_vxlan_call_notifiers(struct net_device *dev,
 	struct switchdev_notifier_vxlan_fdb_info info;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	enum switchdev_notifier_type type;
+	const struct vxlan_config *cfg;
 
+	cfg = rtnl_dereference(vxlan->cfg);
 	type = adding ? SWITCHDEV_VXLAN_FDB_ADD_TO_BRIDGE :
 			SWITCHDEV_VXLAN_FDB_DEL_TO_BRIDGE;
 	mlxsw_sp_switchdev_addr_vxlan_convert(proto, addr, &info.remote_ip);
-	info.remote_port = vxlan->cfg.dst_port;
+	info.remote_port = cfg->dst_port;
 	info.remote_vni = vni;
 	info.remote_ifindex = 0;
 	ether_addr_copy(info.eth_addr, mac);
@@ -3236,8 +3245,10 @@ __mlxsw_sp_fdb_notify_mac_uc_tunnel_process(struct mlxsw_sp *mlxsw_sp,
 
 	if (adding && netif_is_vxlan(dev)) {
 		struct vxlan_dev *vxlan = netdev_priv(dev);
+		const struct vxlan_config *cfg;
 
-		if (!(vxlan->cfg.flags & VXLAN_F_LEARN))
+		cfg = rtnl_dereference(vxlan->cfg);
+		if (!(cfg->flags & VXLAN_F_LEARN))
 			return -EINVAL;
 	}
 
@@ -3722,9 +3733,11 @@ mlxsw_sp_switchdev_vxlan_work_prepare(struct mlxsw_sp_switchdev_event_work *
 {
 	struct vxlan_dev *vxlan = netdev_priv(switchdev_work->dev);
 	struct switchdev_notifier_vxlan_fdb_info *vxlan_fdb_info;
-	struct vxlan_config *cfg = &vxlan->cfg;
+	const struct vxlan_config *cfg;
 	struct netlink_ext_ack *extack;
 
+	cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	extack = switchdev_notifier_info_to_extack(info);
 	vxlan_fdb_info = container_of(info,
 				      struct switchdev_notifier_vxlan_fdb_info,
@@ -3851,11 +3864,15 @@ mlxsw_sp_switchdev_vxlan_vlan_add(struct mlxsw_sp *mlxsw_sp,
 				  struct netlink_ext_ack *extack)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	__be32 vni = vxlan->cfg.vni;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
 	u16 old_vid;
+	__be32 vni;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	/* We cannot have the same VLAN as PVID and egress untagged on multiple
 	 * VxLAN devices. Note that we get this notification before the VLAN is
 	 * actually added to the bridge's database, so it is not possible for
@@ -3935,12 +3952,16 @@ mlxsw_sp_switchdev_vxlan_vlan_del(struct mlxsw_sp *mlxsw_sp,
 				  const struct net_device *vxlan_dev, u16 vid)
 {
 	struct vxlan_dev *vxlan = netdev_priv(vxlan_dev);
-	__be32 vni = vxlan->cfg.vni;
+	const struct vxlan_config *cfg;
 	struct mlxsw_sp_fid *fid;
+	__be32 vni;
 
 	if (!netif_running(vxlan_dev))
 		return;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	fid = mlxsw_sp_fid_lookup_by_vni(mlxsw_sp, vni);
 	if (!fid)
 		return;
diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c
index 347245cc1de4e..0e5ab0055ff8c 100644
--- a/drivers/net/vxlan/vxlan_core.c
+++ b/drivers/net/vxlan/vxlan_core.c
@@ -110,20 +110,23 @@ static struct vxlan_dev *vxlan_vs_find_vni(struct vxlan_sock *vs,
 		vni = 0;
 
 	hlist_for_each_entry_rcu(node, vni_head(vs, vni), hlist) {
+		const struct vxlan_config *cfg;
+
 		if (!node->vxlan)
 			continue;
+
+		cfg = rcu_dereference(node->vxlan->cfg);
+
 		vnode = NULL;
-		if (node->vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
 			vnode = vxlan_vnifilter_lookup(node->vxlan, vni);
 			if (!vnode)
 				continue;
-		} else if (node->vxlan->default_dst.remote_vni != vni) {
+		} else if (cfg->vni != vni) {
 			continue;
 		}
 
 		if (IS_ENABLED(CONFIG_IPV6)) {
-			const struct vxlan_config *cfg = &node->vxlan->cfg;
-
 			if ((cfg->flags & VXLAN_F_IPV6_LINKLOCAL) &&
 			    cfg->remote_ifindex != ifindex)
 				continue;
@@ -157,6 +160,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			  u32 portid, u32 seq, int type, unsigned int flags,
 			  const struct vxlan_rdst *rdst)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	unsigned long now = jiffies;
 	struct nda_cacheinfo ci;
 	bool send_ip, send_eth;
@@ -216,10 +220,10 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			goto nla_put_failure;
 
 		if (rdst->remote_port &&
-		    rdst->remote_port != vxlan->cfg.dst_port &&
+		    rdst->remote_port != cfg->dst_port &&
 		    nla_put_be16(skb, NDA_PORT, rdst->remote_port))
 			goto nla_put_failure;
-		if (rdst->remote_vni != vxlan->default_dst.remote_vni &&
+		if (rdst->remote_vni != cfg->vni &&
 		    nla_put_u32(skb, NDA_VNI, be32_to_cpu(rdst->remote_vni)))
 			goto nla_put_failure;
 		if (rdst->remote_ifindex &&
@@ -227,7 +231,7 @@ static int vxlan_fdb_info(struct sk_buff *skb, struct vxlan_dev *vxlan,
 			goto nla_put_failure;
 	}
 
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) && fdb->key.vni &&
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) && fdb->key.vni &&
 	    nla_put_u32(skb, NDA_SRC_VNI,
 			be32_to_cpu(fdb->key.vni)))
 		goto nla_put_failure;
@@ -377,14 +381,15 @@ static void vxlan_fdb_miss(struct vxlan_dev *vxlan, const u8 eth_addr[ETH_ALEN])
 
 /* Look up Ethernet address in forwarding table */
 static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,
+					    const struct vxlan_config *cfg,
 					    const u8 *mac, __be32 vni)
 {
 	struct vxlan_fdb_key key;
 
 	memset(&key, 0, sizeof(key));
 	memcpy(key.eth_addr, mac, sizeof(key.eth_addr));
-	if (!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA))
-		key.vni = vxlan->default_dst.remote_vni;
+	if (!(cfg->flags & VXLAN_F_COLLECT_METADATA))
+		key.vni = cfg->vni;
 	else
 		key.vni = vni;
 
@@ -393,11 +398,12 @@ static struct vxlan_fdb *vxlan_find_mac_rcu(struct vxlan_dev *vxlan,
 }
 
 static struct vxlan_fdb *vxlan_find_mac_tx(struct vxlan_dev *vxlan,
+					   const struct vxlan_config *cfg,
 					   const u8 *mac, __be32 vni)
 {
 	struct vxlan_fdb *f;
 
-	f = vxlan_find_mac_rcu(vxlan, mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, mac, vni);
 	if (f) {
 		unsigned long now = jiffies;
 
@@ -416,7 +422,7 @@ static struct vxlan_fdb *vxlan_find_mac(struct vxlan_dev *vxlan,
 	lockdep_assert_held_once(&vxlan->hash_lock);
 
 	rcu_read_lock();
-	f = vxlan_find_mac_rcu(vxlan, mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan->cfg), mac, vni);
 	rcu_read_unlock();
 
 	return f;
@@ -457,7 +463,7 @@ int vxlan_fdb_find_uc(struct net_device *dev, const u8 *mac, __be32 vni,
 
 	rcu_read_lock();
 
-	f = vxlan_find_mac_rcu(vxlan, eth_addr, vni);
+	f = vxlan_find_mac_rcu(vxlan, rcu_dereference(vxlan->cfg), eth_addr, vni);
 	if (f)
 		rdst = first_remote_rcu(f);
 	if (!rdst) {
@@ -798,6 +804,7 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
 			       u32 nhid, struct netlink_ext_ack *extack)
 {
 	struct nexthop *old_nh = rtnl_dereference(fdb->nh);
+	const struct vxlan_config *cfg;
 	struct nexthop *nh;
 	int err = -EINVAL;
 
@@ -826,7 +833,8 @@ static int vxlan_fdb_nh_update(struct vxlan_dev *vxlan, struct vxlan_fdb *fdb,
 	}
 
 	/* check nexthop group family */
-	switch (vxlan->default_dst.remote_ip.sa.sa_family) {
+	cfg = rtnl_dereference(vxlan->cfg);
+	switch (cfg->remote_ip.sa.sa_family) {
 	case AF_INET:
 		if (!nexthop_has_v4(nh)) {
 			err = -EAFNOSUPPORT;
@@ -863,12 +871,13 @@ int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		     u32 nhid, struct vxlan_fdb **fdb,
 		     struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	struct vxlan_rdst *rd = NULL;
 	struct vxlan_fdb *f;
 	int rc;
 
-	if (vxlan->cfg.addrmax &&
-	    vxlan->addrcnt >= vxlan->cfg.addrmax)
+	if (cfg->addrmax &&
+	    vxlan->addrcnt >= cfg->addrmax)
 		return -ENOSPC;
 
 	netdev_dbg(vxlan->dev, "add %pM -> %pIS\n", mac, ip);
@@ -1154,6 +1163,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 			   __be32 *vni, u32 *ifindex, u32 *nhid,
 			   struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct net *net = dev_net(vxlan->dev);
 	int err;
 
@@ -1170,7 +1180,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 			return err;
 		}
 	} else {
-		union vxlan_addr *remote = &vxlan->default_dst.remote_ip;
+		const union vxlan_addr *remote = &cfg->remote_ip;
 
 		if (remote->sa.sa_family == AF_INET) {
 			ip->sin.sin_addr.s_addr = htonl(INADDR_ANY);
@@ -1190,7 +1200,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*port = nla_get_be16(tb[NDA_PORT]);
 	} else {
-		*port = vxlan->cfg.dst_port;
+		*port = cfg->dst_port;
 	}
 
 	if (tb[NDA_VNI]) {
@@ -1200,7 +1210,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*vni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));
 	} else {
-		*vni = vxlan->default_dst.remote_vni;
+		*vni = cfg->vni;
 	}
 
 	if (tb[NDA_SRC_VNI]) {
@@ -1210,7 +1220,7 @@ static int vxlan_fdb_parse(struct nlattr *tb[], struct vxlan_dev *vxlan,
 		}
 		*src_vni = cpu_to_be32(nla_get_u32(tb[NDA_SRC_VNI]));
 	} else {
-		*src_vni = vxlan->default_dst.remote_vni;
+		*src_vni = cfg->vni;
 	}
 
 	if (tb[NDA_IFINDEX]) {
@@ -1241,6 +1251,7 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 			 const unsigned char *addr, u16 vid, u16 flags,
 			 bool *notified, struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	/* struct net *net = dev_net(vxlan->dev); */
 	union vxlan_addr ip;
@@ -1268,7 +1279,8 @@ static int vxlan_fdb_add(struct ndmsg *ndm, struct nlattr *tb[],
 		return -EINVAL;
 	}
 
-	if (vxlan->default_dst.remote_ip.sa.sa_family != ip.sa.sa_family)
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (cfg->remote_ip.sa.sa_family != ip.sa.sa_family)
 		return -EAFNOSUPPORT;
 
 	spin_lock_bh(&vxlan->hash_lock);
@@ -1405,18 +1417,21 @@ static int vxlan_fdb_get(struct sk_buff *skb,
 			 struct netlink_ext_ack *extack)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	struct vxlan_fdb *f;
 	__be32 vni;
 	int err;
 
+	cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	if (tb[NDA_VNI])
 		vni = cpu_to_be32(nla_get_u32(tb[NDA_VNI]));
 	else
-		vni = vxlan->default_dst.remote_vni;
+		vni = cfg->vni;
 
 	rcu_read_lock();
 
-	f = vxlan_find_mac_rcu(vxlan, addr, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, addr, vni);
 	if (!f) {
 		NL_SET_ERR_MSG(extack, "Fdb entry not found");
 		err = -ENOENT;
@@ -1434,6 +1449,7 @@ static int vxlan_fdb_get(struct sk_buff *skb,
  * and Tunnel endpoint.
  */
 static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
+					const struct vxlan_config *cfg,
 					union vxlan_addr *src_ip,
 					const u8 *src_mac, u32 src_ifindex,
 					__be32 vni)
@@ -1452,7 +1468,7 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
 		ifindex = src_ifindex;
 #endif
 
-	f = vxlan_find_mac_rcu(vxlan, src_mac, vni);
+	f = vxlan_find_mac_rcu(vxlan, cfg, src_mac, vni);
 	if (likely(f)) {
 		struct vxlan_rdst *rdst = first_remote_rcu(f);
 		unsigned long now = jiffies;
@@ -1488,9 +1504,9 @@ static enum skb_drop_reason vxlan_snoop(struct net_device *dev,
 			vxlan_fdb_update(vxlan, src_mac, src_ip,
 					 NUD_REACHABLE,
 					 NLM_F_EXCL|NLM_F_CREATE,
-					 vxlan->cfg.dst_port,
+					 cfg->dst_port,
 					 vni,
-					 vxlan->default_dst.remote_vni,
+					 cfg->vni,
 					 ifindex, NTF_SELF, 0, true, NULL);
 		spin_unlock(&vxlan->hash_lock);
 	}
@@ -1518,6 +1534,7 @@ static bool __vxlan_sock_release_prep(struct vxlan_sock *vs)
 
 static void vxlan_sock_release(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_sock *sock4 = rtnl_dereference(vxlan->vn4_sock);
 #if IS_ENABLED(CONFIG_IPV6)
 	struct vxlan_sock *sock6 = rtnl_dereference(vxlan->vn6_sock);
@@ -1527,7 +1544,7 @@ static void vxlan_sock_release(struct vxlan_dev *vxlan)
 
 	RCU_INIT_POINTER(vxlan->vn4_sock, NULL);
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vs_del_vnigrp(vxlan);
 	else
 		vxlan_vs_del_dev(vxlan);
@@ -1598,6 +1615,7 @@ static void vxlan_parse_gbp_hdr(struct sk_buff *skb, u32 vxflags,
 }
 
 static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,
+					  const struct vxlan_config *cfg,
 					  struct vxlan_sock *vs,
 					  struct sk_buff *skb, __be32 vni)
 {
@@ -1623,10 +1641,10 @@ static enum skb_drop_reason vxlan_set_mac(struct vxlan_dev *vxlan,
 #endif
 	}
 
-	if (!(vxlan->cfg.flags & VXLAN_F_LEARN))
+	if (!(cfg->flags & VXLAN_F_LEARN))
 		return SKB_NOT_DROPPED_YET;
 
-	return vxlan_snoop(skb->dev, &saddr, eth_hdr(skb)->h_source,
+	return vxlan_snoop(skb->dev, cfg, &saddr, eth_hdr(skb)->h_source,
 			   ifindex, vni);
 }
 
@@ -1657,18 +1675,21 @@ static bool vxlan_ecn_decapsulate(struct vxlan_sock *vs, void *oiph,
 static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 {
 	struct vxlan_vni_node *vninode = NULL;
-	const struct vxlanhdr *vh;
-	struct vxlan_dev *vxlan;
-	struct vxlan_sock *vs;
-	struct vxlan_metadata _md;
-	struct vxlan_metadata *md = &_md;
 	__be16 protocol = htons(ETH_P_TEB);
+	const struct vxlan_config *cfg;
 	enum skb_drop_reason reason;
+	const struct vxlanhdr *vh;
+	struct vxlan_metadata *md;
+	struct vxlan_metadata _md;
+	struct vxlan_dev *vxlan;
 	bool raw_proto = false;
-	void *oiph;
+	struct vxlan_sock *vs;
 	__be32 vni = 0;
+	void *oiph;
 	int nh;
 
+	md = &_md;
+
 	/* Need UDP and VXLAN header to be present */
 	reason = pskb_may_pull_reason(skb, VXLAN_HLEN);
 	if (reason)
@@ -1696,8 +1717,10 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		goto drop;
 	}
 
-	if (vh->vx_flags & vxlan->cfg.reserved_bits.vx_flags ||
-	    vh->vx_vni & vxlan->cfg.reserved_bits.vx_vni) {
+	cfg = rcu_dereference(vxlan->cfg);
+
+	if (vh->vx_flags & cfg->reserved_bits.vx_flags ||
+	    vh->vx_vni & cfg->reserved_bits.vx_vni) {
 		/* If the header uses bits besides those enabled by the
 		 * netdevice configuration, treat this as a malformed packet.
 		 * This behavior diverges from VXLAN RFC (RFC7348) which
@@ -1709,12 +1732,12 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		reason = SKB_DROP_REASON_VXLAN_INVALID_HDR;
 		DEV_STATS_INC(vxlan->dev, rx_frame_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_GPE) {
+	if (cfg->flags & VXLAN_F_GPE) {
 		if (!vxlan_parse_gpe_proto(vh, &protocol))
 			goto drop;
 		raw_proto = true;
@@ -1726,8 +1749,8 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		goto drop;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_REMCSUM_RX) {
-		reason = vxlan_remcsum(skb, vxlan->cfg.flags);
+	if (cfg->flags & VXLAN_F_REMCSUM_RX) {
+		reason = vxlan_remcsum(skb, cfg->flags);
 		if (unlikely(reason))
 			goto drop;
 	}
@@ -1752,14 +1775,14 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		memset(md, 0, sizeof(*md));
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_GBP)
-		vxlan_parse_gbp_hdr(skb, vxlan->cfg.flags, md);
+	if (cfg->flags & VXLAN_F_GBP)
+		vxlan_parse_gbp_hdr(skb, cfg->flags, md);
 	/* Note that GBP and GPE can never be active together. This is
 	 * ensured in vxlan_dev_configure.
 	 */
 
 	if (!raw_proto) {
-		reason = vxlan_set_mac(vxlan, vs, skb, vni);
+		reason = vxlan_set_mac(vxlan, cfg, vs, skb, vni);
 		if (reason)
 			goto drop;
 	} else {
@@ -1780,7 +1803,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	if (reason) {
 		DEV_STATS_INC(vxlan->dev, rx_length_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
@@ -1792,7 +1815,7 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 		reason = SKB_DROP_REASON_IP_TUNNEL_ECN;
 		DEV_STATS_INC(vxlan->dev, rx_frame_errors);
 		DEV_STATS_INC(vxlan->dev, rx_errors);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_ERRORS, 0);
 		goto drop;
 	}
@@ -1802,14 +1825,15 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
 	if (unlikely(!(vxlan->dev->flags & IFF_UP))) {
 		rcu_read_unlock();
 		dev_dstats_rx_dropped(vxlan->dev);
-		vxlan_vnifilter_count(vxlan, vni, vninode,
+		vxlan_vnifilter_count(vxlan, cfg, vni, vninode,
 				      VXLAN_VNI_STATS_RX_DROPS, 0);
 		reason = SKB_DROP_REASON_DEV_READY;
 		goto drop;
 	}
 
 	dev_dstats_rx_add(vxlan->dev, skb->len);
-	vxlan_vnifilter_count(vxlan, vni, vninode, VXLAN_VNI_STATS_RX, skb->len);
+	vxlan_vnifilter_count(vxlan, cfg, vni, vninode, VXLAN_VNI_STATS_RX,
+			      skb->len);
 	gro_cells_receive(&vxlan->gro_cells, skb);
 
 	rcu_read_unlock();
@@ -1850,7 +1874,8 @@ static int vxlan_err_lookup(struct sock *sk, struct sk_buff *skb)
 	return 0;
 }
 
-static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
+static int arp_reduce(struct net_device *dev, struct sk_buff *skb,
+		      const struct vxlan_config *cfg, __be32 vni)
 {
 	struct neigh_table *tbl = arp_table(dev_net(dev));
 	struct vxlan_dev *vxlan = netdev_priv(dev);
@@ -1864,7 +1889,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 	if (!pskb_network_may_pull(skb, arp_hdr_len(dev))) {
 		dev_dstats_tx_dropped(dev);
-		vxlan_vnifilter_count(vxlan, vni, NULL,
+		vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 				      VXLAN_VNI_STATS_TX_DROPS, 0);
 		goto out;
 	}
@@ -1905,7 +1930,7 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 		neigh_ha_snapshot(ha, n, n->dev);
 
 		rcu_read_lock();
-		f = vxlan_find_mac_tx(vxlan, ha, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, ha, vni);
 		if (f)
 			rdst = first_remote_rcu(f);
 		if (rdst && vxlan_addr_any(&rdst->remote_ip)) {
@@ -1931,11 +1956,11 @@ static int arp_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 		if (netif_rx(reply) == NET_RX_DROP) {
 			dev_dstats_rx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_RX_DROPS, 0);
 		}
 
-	} else if (vxlan->cfg.flags & VXLAN_F_L3MISS) {
+	} else if (cfg->flags & VXLAN_F_L3MISS) {
 		union vxlan_addr ipa = {
 			.sin.sin_addr.s_addr = tip,
 			.sin.sin_family = AF_INET,
@@ -2043,7 +2068,8 @@ static struct sk_buff *vxlan_na_create(struct sk_buff *request,
 	return reply;
 }
 
-static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
+static int neigh_reduce(struct net_device *dev, struct sk_buff *skb,
+			const struct vxlan_config *cfg, __be32 vni)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	const struct in6_addr *daddr;
@@ -2077,7 +2103,7 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 		}
 
 		neigh_ha_snapshot(ha, n, n->dev);
-		f = vxlan_find_mac_tx(vxlan, ha, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, ha, vni);
 		if (f)
 			rdst = first_remote_rcu(f);
 		if (rdst && vxlan_addr_any(&rdst->remote_ip)) {
@@ -2096,10 +2122,10 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 
 		if (netif_rx(reply) == NET_RX_DROP) {
 			dev_dstats_rx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_RX_DROPS, 0);
 		}
-	} else if (vxlan->cfg.flags & VXLAN_F_L3MISS) {
+	} else if (cfg->flags & VXLAN_F_L3MISS) {
 		union vxlan_addr ipa = {
 			.sin6.sin6_addr = msg->target,
 			.sin6.sin6_family = AF_INET6,
@@ -2115,9 +2141,9 @@ static int neigh_reduce(struct net_device *dev, struct sk_buff *skb, __be32 vni)
 }
 #endif
 
-static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
+static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb,
+			       const struct vxlan_config *cfg)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct neigh_table *tbl;
 	struct neighbour *n;
 
@@ -2136,7 +2162,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
 		tbl = arp_table(dev_net(dev));
 		pip = ip_hdr(skb);
 		n = neigh_lookup(tbl, &pip->daddr, dev);
-		if (!n && (vxlan->cfg.flags & VXLAN_F_L3MISS)) {
+		if (!n && (cfg->flags & VXLAN_F_L3MISS)) {
 			union vxlan_addr ipa = {
 				.sin.sin_addr.s_addr = pip->daddr,
 				.sin.sin_family = AF_INET,
@@ -2164,7 +2190,7 @@ static bool route_shortcircuit(struct net_device *dev, struct sk_buff *skb)
 		tbl = nd_table(dev_net(dev));
 		pip6 = ipv6_hdr(skb);
 		n = neigh_lookup(tbl, &pip6->daddr, dev);
-		if (!n && (vxlan->cfg.flags & VXLAN_F_L3MISS)) {
+		if (!n && (cfg->flags & VXLAN_F_L3MISS)) {
 			union vxlan_addr ipa = {
 				.sin6.sin6_addr = pip6->daddr,
 				.sin6.sin6_family = AF_INET6,
@@ -2282,20 +2308,28 @@ static int vxlan_build_skb(struct sk_buff *skb, struct dst_entry *dst,
 
 /* Bypass encapsulation if the destination is local */
 static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
-			       struct vxlan_dev *dst_vxlan, __be32 vni,
-			       bool snoop)
+			       struct vxlan_dev *dst_vxlan,
+			       const struct vxlan_config *src_cfg,
+			       __be32 vni, bool snoop)
 {
+	const struct vxlan_config *dst_cfg;
 	union vxlan_addr loopback;
-	union vxlan_addr *remote_ip = &dst_vxlan->default_dst.remote_ip;
 	unsigned int len = skb->len;
-	struct net_device *dev;
+	struct net_device *dev = dst_vxlan->dev;
 
 	skb->pkt_type = PACKET_HOST;
 	skb->encapsulation = 0;
-	skb->dev = dst_vxlan->dev;
+	skb->dev = dev;
 	__skb_pull(skb, skb_network_offset(skb));
 
-	if (remote_ip->sa.sa_family == AF_INET) {
+	rcu_read_lock();
+	dst_cfg = rcu_dereference(dst_vxlan->cfg);
+	if (unlikely(!(dev->flags & IFF_UP))) {
+		kfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);
+		goto drop;
+	}
+
+	if (dst_cfg->remote_ip.sa.sa_family == AF_INET) {
 		loopback.sin.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
 		loopback.sa.sa_family =  AF_INET;
 #if IS_ENABLED(CONFIG_IPV6)
@@ -2305,27 +2339,20 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
 #endif
 	}
 
-	rcu_read_lock();
-	dev = skb->dev;
-	if (unlikely(!(dev->flags & IFF_UP))) {
-		kfree_skb_reason(skb, SKB_DROP_REASON_DEV_READY);
-		goto drop;
-	}
-
-	if ((dst_vxlan->cfg.flags & VXLAN_F_LEARN) && snoop)
-		vxlan_snoop(dev, &loopback, eth_hdr(skb)->h_source, 0, vni);
+	if ((dst_cfg->flags & VXLAN_F_LEARN) && snoop)
+		vxlan_snoop(dev, dst_cfg, &loopback, eth_hdr(skb)->h_source, 0, vni);
 
 	dev_dstats_tx_add(src_vxlan->dev, len);
-	vxlan_vnifilter_count(src_vxlan, vni, NULL, VXLAN_VNI_STATS_TX, len);
+	vxlan_vnifilter_count(src_vxlan, src_cfg, vni, NULL, VXLAN_VNI_STATS_TX, len);
 
 	if (__netif_rx(skb) == NET_RX_SUCCESS) {
 		dev_dstats_rx_add(dst_vxlan->dev, len);
-		vxlan_vnifilter_count(dst_vxlan, vni, NULL, VXLAN_VNI_STATS_RX,
+		vxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL, VXLAN_VNI_STATS_RX,
 				      len);
 	} else {
 drop:
 		dev_dstats_rx_dropped(dev);
-		vxlan_vnifilter_count(dst_vxlan, vni, NULL,
+		vxlan_vnifilter_count(dst_vxlan, dst_cfg, vni, NULL,
 				      VXLAN_VNI_STATS_RX_DROPS, 0);
 	}
 	rcu_read_unlock();
@@ -2333,6 +2360,7 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan,
 
 static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 				 struct vxlan_dev *vxlan,
+				 const struct vxlan_config *cfg,
 				 int addr_family,
 				 __be16 dst_port, int dst_ifindex, __be32 vni,
 				 struct dst_entry *dst,
@@ -2348,22 +2376,22 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 	/* Bypass encapsulation if the destination is local */
 	if (rt_flags & RTCF_LOCAL &&
 	    !(rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) &&
-	    vxlan->cfg.flags & VXLAN_F_LOCALBYPASS) {
+	    cfg->flags & VXLAN_F_LOCALBYPASS) {
 		struct vxlan_dev *dst_vxlan;
 
 		dst_release(dst);
 		dst_vxlan = vxlan_find_vni(vxlan->net, dst_ifindex, vni,
 					   addr_family, dst_port,
-					   vxlan->cfg.flags);
+					   cfg->flags);
 		if (!dst_vxlan) {
 			DEV_STATS_INC(dev, tx_errors);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_TX_ERRORS, 0);
 			kfree_skb_reason(skb, SKB_DROP_REASON_VXLAN_VNI_NOT_FOUND);
 
 			return -ENOENT;
 		}
-		vxlan_encap_bypass(skb, vxlan, dst_vxlan, vni, true);
+		vxlan_encap_bypass(skb, vxlan, dst_vxlan, cfg, vni, true);
 		return 1;
 	}
 
@@ -2371,30 +2399,35 @@ static int encap_bypass_if_local(struct sk_buff *skb, struct net_device *dev,
 }
 
 void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
+		    const struct vxlan_config *cfg,
 		    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc)
 {
+	unsigned int pkt_len = skb->len;
+	struct vxlan_metadata _md = {};
+	__be16 src_port = 0, dst_port;
+	struct dst_entry *ndst = NULL;
+	enum skb_drop_reason reason;
 	struct dst_cache *dst_cache;
+	const struct iphdr *old_iph;
 	struct ip_tunnel_info *info;
 	struct ip_tunnel_key *pkey;
+	struct vxlan_metadata *md;
 	struct ip_tunnel_key key;
-	struct vxlan_dev *vxlan = netdev_priv(dev);
-	const struct iphdr *old_iph;
-	struct vxlan_metadata _md = {};
-	struct vxlan_metadata *md = &_md;
-	unsigned int pkt_len = skb->len;
-	__be16 src_port = 0, dst_port;
-	struct dst_entry *ndst = NULL;
+	struct vxlan_dev *vxlan;
+	u32 flags = cfg->flags;
+	bool udp_sum = false;
+	bool no_eth_encap;
 	int addr_family;
+	bool use_cache;
+	__be32 vni = 0;
 	__u8 tos, ttl;
 	int ifindex;
 	int err = 0;
-	u32 flags = vxlan->cfg.flags;
-	bool use_cache;
-	bool udp_sum = false;
-	bool xnet = !net_eq(vxlan->net, dev_net(vxlan->dev));
-	enum skb_drop_reason reason;
-	bool no_eth_encap;
-	__be32 vni = 0;
+	bool xnet;
+
+	vxlan = netdev_priv(dev);
+	xnet = !net_eq(vxlan->net, dev_net(vxlan->dev));
+	md = &_md;
 
 	no_eth_encap = flags & VXLAN_F_GPE && skb->protocol != htons(ETH_P_TEB);
 	reason = skb_vlan_inet_prepare(skb, no_eth_encap);
@@ -2414,23 +2447,23 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (vxlan_addr_any(&rdst->remote_ip)) {
 			if (did_rsc) {
 				/* short-circuited back to local bridge */
-				vxlan_encap_bypass(skb, vxlan, vxlan,
+				vxlan_encap_bypass(skb, vxlan, vxlan, cfg,
 						   default_vni, true);
 				return;
 			}
 			goto drop;
 		}
 
-		addr_family = vxlan->cfg.saddr.sa.sa_family;
-		dst_port = rdst->remote_port ? rdst->remote_port : vxlan->cfg.dst_port;
+		addr_family = cfg->saddr.sa.sa_family;
+		dst_port = rdst->remote_port ? rdst->remote_port : cfg->dst_port;
 		vni = (rdst->remote_vni) ? : default_vni;
 		ifindex = rdst->remote_ifindex;
 
 		if (addr_family == AF_INET) {
-			key.u.ipv4.src = vxlan->cfg.saddr.sin.sin_addr.s_addr;
+			key.u.ipv4.src = cfg->saddr.sin.sin_addr.s_addr;
 			key.u.ipv4.dst = rdst->remote_ip.sin.sin_addr.s_addr;
 		} else {
-			key.u.ipv6.src = vxlan->cfg.saddr.sin6.sin6_addr;
+			key.u.ipv6.src = cfg->saddr.sin6.sin6_addr;
 			key.u.ipv6.dst = rdst->remote_ip.sin6.sin6_addr;
 		}
 
@@ -2439,11 +2472,11 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (flags & VXLAN_F_TTL_INHERIT) {
 			ttl = ip_tunnel_get_ttl(old_iph, skb);
 		} else {
-			ttl = vxlan->cfg.ttl;
+			ttl = cfg->ttl;
 			if (!ttl && vxlan_addr_multicast(&rdst->remote_ip))
 				ttl = 1;
 		}
-		tos = vxlan->cfg.tos;
+		tos = cfg->tos;
 		if (tos == 1)
 			tos = ip_tunnel_get_dsfield(old_iph, skb);
 		if (tos && !info)
@@ -2454,9 +2487,9 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		else
 			udp_sum = !(flags & VXLAN_F_UDP_ZERO_CSUM6_TX);
 #if IS_ENABLED(CONFIG_IPV6)
-		switch (vxlan->cfg.label_policy) {
+		switch (cfg->label_policy) {
 		case VXLAN_LABEL_FIXED:
-			key.label = vxlan->cfg.label;
+			key.label = cfg->label;
 			break;
 		case VXLAN_LABEL_INHERIT:
 			key.label = ip_tunnel_get_flowlabel(old_iph, skb);
@@ -2474,7 +2507,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		}
 		pkey = &info->key;
 		addr_family = ip_tunnel_info_af(info);
-		dst_port = info->key.tp_dst ? : vxlan->cfg.dst_port;
+		dst_port = info->key.tp_dst ? : cfg->dst_port;
 		vni = tunnel_id_to_key32(info->key.tun_id);
 		ifindex = 0;
 		dst_cache = &info->dst_cache;
@@ -2487,8 +2520,8 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		tos = info->key.tos;
 		udp_sum = test_bit(IP_TUNNEL_CSUM_BIT, info->key.tun_flags);
 	}
-	src_port = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
-				     vxlan->cfg.port_max, true);
+	src_port = udp_flow_src_port(dev_net(dev), skb, cfg->port_min,
+				     cfg->port_max, true);
 
 	rcu_read_lock();
 	if (addr_family == AF_INET) {
@@ -2521,15 +2554,15 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 
 		if (!info) {
 			/* Bypass encapsulation if the destination is local */
-			err = encap_bypass_if_local(skb, dev, vxlan, AF_INET,
+			err = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET,
 						    dst_port, ifindex, vni,
 						    &rt->dst, rt->rt_flags);
 			if (err)
 				goto out_unlock;
 
-			if (vxlan->cfg.df == VXLAN_DF_SET) {
+			if (cfg->df == VXLAN_DF_SET) {
 				df = htons(IP_DF);
-			} else if (vxlan->cfg.df == VXLAN_DF_INHERIT) {
+			} else if (cfg->df == VXLAN_DF_INHERIT) {
 				struct ethhdr *eth = eth_hdr(skb);
 
 				if (ntohs(eth->h_proto) == ETH_P_IPV6 ||
@@ -2558,7 +2591,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				unclone->key.u.ipv4.src = pkey->u.ipv4.dst;
 				unclone->key.u.ipv4.dst = saddr;
 			}
-			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
+			vxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);
 			dst_release(ndst);
 			goto out_unlock;
 		}
@@ -2608,7 +2641,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		if (!info) {
 			u32 rt6i_flags = dst_rt6_info(ndst)->rt6i_flags;
 
-			err = encap_bypass_if_local(skb, dev, vxlan, AF_INET6,
+			err = encap_bypass_if_local(skb, dev, vxlan, cfg, AF_INET6,
 						    dst_port, ifindex, vni,
 						    ndst, rt6i_flags);
 			if (err)
@@ -2632,7 +2665,7 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				unclone->key.u.ipv6.dst = saddr;
 			}
 
-			vxlan_encap_bypass(skb, vxlan, vxlan, vni, false);
+			vxlan_encap_bypass(skb, vxlan, vxlan, cfg, vni, false);
 			dst_release(ndst);
 			goto out_unlock;
 		}
@@ -2653,14 +2686,14 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 				     ip6cb_flags);
 #endif
 	}
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX, pkt_len);
 out_unlock:
 	rcu_read_unlock();
 	return;
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_DROPS, 0);
 	kfree_skb_reason(skb, reason);
 	return;
 
@@ -2672,11 +2705,12 @@ void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
 		DEV_STATS_INC(dev, tx_carrier_errors);
 	dst_release(ndst);
 	DEV_STATS_INC(dev, tx_errors);
-	vxlan_vnifilter_count(vxlan, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);
+	vxlan_vnifilter_count(vxlan, cfg, vni, NULL, VXLAN_VNI_STATS_TX_ERRORS, 0);
 	kfree_skb_reason(skb, reason);
 }
 
 static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
+			  const struct vxlan_config *cfg,
 			  struct vxlan_fdb *f, __be32 vni, bool did_rsc)
 {
 	struct vxlan_rdst nh_rdst;
@@ -2693,7 +2727,7 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
 	do_xmit = vxlan_fdb_nh_path_select(nh, hash, &nh_rdst);
 
 	if (likely(do_xmit))
-		vxlan_xmit_one(skb, dev, vni, &nh_rdst, did_rsc);
+		vxlan_xmit_one(skb, dev, cfg, vni, &nh_rdst, did_rsc);
 	else
 		goto drop;
 
@@ -2701,15 +2735,15 @@ static void vxlan_xmit_nh(struct sk_buff *skb, struct net_device *dev,
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(netdev_priv(dev), vni, NULL,
+	vxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,
 			      VXLAN_VNI_STATS_TX_DROPS, 0);
 	dev_kfree_skb(skb);
 }
 
 static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
-				   u32 nhid, __be32 vni)
+				   u32 nhid, __be32 vni,
+				   const struct vxlan_config *cfg)
 {
-	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst nh_rdst;
 	struct nexthop *nh;
 	bool do_xmit;
@@ -2727,11 +2761,11 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
 	do_xmit = vxlan_fdb_nh_path_select(nh, hash, &nh_rdst);
 	rcu_read_unlock();
 
-	if (vxlan->cfg.saddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)
+	if (cfg->saddr.sa.sa_family != nh_rdst.remote_ip.sa.sa_family)
 		goto drop;
 
 	if (likely(do_xmit))
-		vxlan_xmit_one(skb, dev, vni, &nh_rdst, false);
+		vxlan_xmit_one(skb, dev, cfg, vni, &nh_rdst, false);
 	else
 		goto drop;
 
@@ -2739,7 +2773,7 @@ static netdev_tx_t vxlan_xmit_nhid(struct sk_buff *skb, struct net_device *dev,
 
 drop:
 	dev_dstats_tx_dropped(dev);
-	vxlan_vnifilter_count(netdev_priv(dev), vni, NULL,
+	vxlan_vnifilter_count(netdev_priv(dev), cfg, vni, NULL,
 			      VXLAN_VNI_STATS_TX_DROPS, 0);
 	dev_kfree_skb(skb);
 	return NETDEV_TX_OK;
@@ -2756,34 +2790,47 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct vxlan_rdst *rdst, *fdst = NULL;
 	const struct ip_tunnel_info *info;
+	const struct vxlan_config *cfg;
+	__be32 default_vni;
 	struct vxlan_fdb *f;
 	struct ethhdr *eth;
 	__be32 vni = 0;
-	u32 nhid = 0;
 	bool did_rsc;
+	u32 nhid = 0;
+	u32 flags;
+
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+	flags = cfg->flags;
+	default_vni = cfg->vni;
 
 	info = skb_tunnel_info(skb);
 
 	skb_reset_mac_header(skb);
 
-	if (vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) {
+	if (flags & VXLAN_F_COLLECT_METADATA) {
 		if (info && info->mode & IP_TUNNEL_INFO_BRIDGE &&
 		    info->mode & IP_TUNNEL_INFO_TX) {
 			vni = tunnel_id_to_key32(info->key.tun_id);
 			nhid = info->key.nhid;
 		} else {
 			if (info && info->mode & IP_TUNNEL_INFO_TX)
-				vxlan_xmit_one(skb, dev, vni, NULL, false);
+				vxlan_xmit_one(skb, dev, cfg, vni, NULL, false);
 			else
 				kfree_skb_reason(skb, SKB_DROP_REASON_TUNNEL_TXINFO);
+			rcu_read_unlock();
 			return NETDEV_TX_OK;
 		}
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_PROXY) {
+	if (flags & VXLAN_F_PROXY) {
 		eth = eth_hdr(skb);
-		if (ntohs(eth->h_proto) == ETH_P_ARP)
-			return arp_reduce(dev, skb, vni);
+		if (ntohs(eth->h_proto) == ETH_P_ARP) {
+			netdev_tx_t res = arp_reduce(dev, skb, cfg, vni);
+
+			rcu_read_unlock();
+			return res;
+		}
 #if IS_ENABLED(CONFIG_IPV6)
 		else if (ntohs(eth->h_proto) == ETH_P_IPV6 &&
 			 pskb_network_may_pull(skb, sizeof(struct ipv6hdr) +
@@ -2792,53 +2839,57 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 			struct nd_msg *m = (struct nd_msg *)(ipv6_hdr(skb) + 1);
 
 			if (m->icmph.icmp6_code == 0 &&
-			    m->icmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION)
-				return neigh_reduce(dev, skb, vni);
+			    m->icmph.icmp6_type == NDISC_NEIGHBOUR_SOLICITATION) {
+				netdev_tx_t res = neigh_reduce(dev, skb, cfg, vni);
+
+				rcu_read_unlock();
+				return res;
+			}
 		}
 #endif
 	}
 
-	if (nhid)
-		return vxlan_xmit_nhid(skb, dev, nhid, vni);
+	if (nhid) {
+		netdev_tx_t res = vxlan_xmit_nhid(skb, dev, nhid, vni, cfg);
+
+		rcu_read_unlock();
+		return res;
+	}
 
-	if (vxlan->cfg.flags & VXLAN_F_MDB) {
+	if (test_bit(VXLAN_DEV_F_MDB, &vxlan->flags)) {
 		struct vxlan_mdb_entry *mdb_entry;
 
-		rcu_read_lock();
-		mdb_entry = vxlan_mdb_entry_skb_get(vxlan, skb, vni);
+		mdb_entry = vxlan_mdb_entry_skb_get(vxlan, cfg, skb, vni);
 		if (mdb_entry) {
-			netdev_tx_t ret;
+			netdev_tx_t ret = vxlan_mdb_xmit(vxlan, cfg, mdb_entry, skb);
 
-			ret = vxlan_mdb_xmit(vxlan, mdb_entry, skb);
 			rcu_read_unlock();
 			return ret;
 		}
-		rcu_read_unlock();
 	}
 
 	eth = eth_hdr(skb);
-	rcu_read_lock();
-	f = vxlan_find_mac_tx(vxlan, eth->h_dest, vni);
+	f = vxlan_find_mac_tx(vxlan, cfg, eth->h_dest, vni);
 	did_rsc = false;
 
-	if (f && (f->flags & NTF_ROUTER) && (vxlan->cfg.flags & VXLAN_F_RSC) &&
+	if (f && (f->flags & NTF_ROUTER) && (flags & VXLAN_F_RSC) &&
 	    (ntohs(eth->h_proto) == ETH_P_IP ||
 	     ntohs(eth->h_proto) == ETH_P_IPV6)) {
-		did_rsc = route_shortcircuit(dev, skb);
+		did_rsc = route_shortcircuit(dev, skb, cfg);
 		eth = eth_hdr(skb);
 		if (did_rsc)
-			f = vxlan_find_mac_tx(vxlan, eth->h_dest, vni);
+			f = vxlan_find_mac_tx(vxlan, cfg, eth->h_dest, vni);
 	}
 
 	if (f == NULL) {
-		f = vxlan_find_mac_tx(vxlan, all_zeros_mac, vni);
+		f = vxlan_find_mac_tx(vxlan, cfg, all_zeros_mac, vni);
 		if (f == NULL) {
-			if ((vxlan->cfg.flags & VXLAN_F_L2MISS) &&
+			if ((flags & VXLAN_F_L2MISS) &&
 			    !is_multicast_ether_addr(eth->h_dest))
 				vxlan_fdb_miss(vxlan, eth->h_dest);
 
 			dev_dstats_tx_dropped(dev);
-			vxlan_vnifilter_count(vxlan, vni, NULL,
+			vxlan_vnifilter_count(vxlan, cfg, vni, NULL,
 					      VXLAN_VNI_STATS_TX_DROPS, 0);
 			kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
 			goto out;
@@ -2846,8 +2897,8 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	if (rcu_access_pointer(f->nh)) {
-		vxlan_xmit_nh(skb, dev, f,
-			      (vni ? : vxlan->default_dst.remote_vni), did_rsc);
+		vxlan_xmit_nh(skb, dev, cfg, f,
+			      (vni ? : default_vni), did_rsc);
 	} else {
 		list_for_each_entry_rcu(rdst, &f->remotes, list) {
 			struct sk_buff *skb1;
@@ -2858,10 +2909,10 @@ static netdev_tx_t vxlan_xmit(struct sk_buff *skb, struct net_device *dev)
 			}
 			skb1 = skb_clone(skb, GFP_ATOMIC);
 			if (skb1)
-				vxlan_xmit_one(skb1, dev, vni, rdst, did_rsc);
+				vxlan_xmit_one(skb1, dev, cfg, vni, rdst, did_rsc);
 		}
 		if (fdst)
-			vxlan_xmit_one(skb, dev, vni, fdst, did_rsc);
+			vxlan_xmit_one(skb, dev, cfg, vni, fdst, did_rsc);
 		else
 			kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
 	}
@@ -2876,12 +2927,15 @@ static void vxlan_cleanup(struct timer_list *t)
 {
 	struct vxlan_dev *vxlan = timer_container_of(vxlan, t, age_timer);
 	unsigned long next_timer = jiffies + FDB_AGE_INTERVAL;
+	const struct vxlan_config *cfg;
 	struct vxlan_fdb *f;
 
 	if (!netif_running(vxlan->dev))
 		return;
 
 	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
+
 	hlist_for_each_entry_rcu(f, &vxlan->fdb_list, fdb_node) {
 		unsigned long timeout;
 
@@ -2891,7 +2945,7 @@ static void vxlan_cleanup(struct timer_list *t)
 		if (f->flags & NTF_EXT_LEARNED)
 			continue;
 
-		timeout = READ_ONCE(f->updated) + vxlan->cfg.age_interval * HZ;
+		timeout = READ_ONCE(f->updated) + cfg->age_interval * HZ;
 		if (time_before_eq(timeout, jiffies)) {
 			spin_lock(&vxlan->hash_lock);
 			if (!hlist_unhashed(&f->fdb_node)) {
@@ -2923,10 +2977,14 @@ static void vxlan_vs_del_dev(struct vxlan_dev *vxlan)
 static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,
 			     struct vxlan_dev_node *node)
 {
-	__be32 vni = vxlan->default_dst.remote_vni;
+	const struct vxlan_config *cfg;
+	__be32 vni;
 
 	ASSERT_RTNL();
 
+	cfg = rtnl_dereference(vxlan->cfg);
+	vni = cfg->vni;
+
 	node->vxlan = vxlan;
 	hlist_add_head_rcu(&node->hlist, vni_head(vs, vni));
 }
@@ -2935,13 +2993,16 @@ static void vxlan_vs_add_dev(struct vxlan_sock *vs, struct vxlan_dev *vxlan,
 static int vxlan_init(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	err = rhashtable_init(&vxlan->fdb_hash_tbl, &vxlan_fdb_rht_params);
 	if (err)
 		return err;
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
+	if (cfg->flags & VXLAN_F_VNIFILTER) {
 		err = vxlan_vnigroup_init(vxlan);
 		if (err)
 			goto err_rhashtable_destroy;
@@ -2961,7 +3022,7 @@ static int vxlan_init(struct net_device *dev)
 err_gro_cells_destroy:
 	gro_cells_destroy(&vxlan->gro_cells);
 err_vnigroup_uninit:
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vnigroup_uninit(vxlan);
 err_rhashtable_destroy:
 	rhashtable_destroy(&vxlan->fdb_hash_tbl);
@@ -2971,10 +3032,13 @@ static int vxlan_init(struct net_device *dev)
 static void vxlan_uninit(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
+
+	cfg = rtnl_dereference(vxlan->cfg);
 
 	vxlan_mdb_fini(vxlan);
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
+	if (cfg->flags & VXLAN_F_VNIFILTER)
 		vxlan_vnigroup_uninit(vxlan);
 
 	gro_cells_destroy(&vxlan->gro_cells);
@@ -2986,6 +3050,7 @@ static void vxlan_uninit(struct net_device *dev)
 static int vxlan_open(struct net_device *dev)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *cfg;
 	int ret;
 
 	ret = vxlan_sock_add(vxlan);
@@ -2998,7 +3063,8 @@ static int vxlan_open(struct net_device *dev)
 		return ret;
 	}
 
-	if (vxlan->cfg.age_interval)
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (cfg->age_interval)
 		mod_timer(&vxlan->age_timer, jiffies + FDB_AGE_INTERVAL);
 
 	return ret;
@@ -3020,8 +3086,10 @@ struct vxlan_fdb_flush_desc {
 static bool vxlan_fdb_is_default_entry(const struct vxlan_fdb *f,
 				       const struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
+
 	return is_zero_ether_addr(f->key.eth_addr) &&
-	       f->key.vni == vxlan->cfg.vni;
+	       f->key.vni == cfg->vni;
 }
 
 static bool vxlan_fdb_nhid_matches(const struct vxlan_fdb *f, u32 nhid)
@@ -3238,15 +3306,18 @@ static void vxlan_set_multicast_list(struct net_device *dev)
 static int vxlan_change_mtu(struct net_device *dev, int new_mtu)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct net_device *lowerdev = __dev_get_by_index(vxlan->net,
-							 dst->remote_ifindex);
+	const struct vxlan_config *cfg;
+	struct net_device *lowerdev;
+
+	cfg = rtnl_dereference(vxlan->cfg);
+
+	lowerdev = __dev_get_by_index(vxlan->net, cfg->remote_ifindex);
 
 	/* This check is different than dev->max_mtu, because it looks at
 	 * the lowerdev->mtu, rather than the static dev->max_mtu
 	 */
 	if (lowerdev) {
-		int max_mtu = lowerdev->mtu - vxlan_headroom(vxlan->cfg.flags);
+		int max_mtu = lowerdev->mtu - vxlan_headroom(cfg->flags);
 		if (new_mtu > max_mtu)
 			return -EINVAL;
 	}
@@ -3259,11 +3330,14 @@ static int vxlan_fill_metadata_dst(struct net_device *dev, struct sk_buff *skb)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct ip_tunnel_info *info = skb_tunnel_info(skb);
+	const struct vxlan_config *cfg;
 	__be16 sport, dport;
 
-	sport = udp_flow_src_port(dev_net(dev), skb, vxlan->cfg.port_min,
-				  vxlan->cfg.port_max, true);
-	dport = info->key.tp_dst ? : vxlan->cfg.dst_port;
+	cfg = rcu_dereference(vxlan->cfg);
+
+	sport = udp_flow_src_port(dev_net(dev), skb, cfg->port_min,
+				  cfg->port_max, true);
+	dport = info->key.tp_dst ? : cfg->dst_port;
 
 	if (ip_tunnel_info_af(info) == AF_INET) {
 		struct vxlan_sock *sock4 = rcu_dereference(vxlan->vn4_sock);
@@ -3373,6 +3447,15 @@ static void vxlan_offload_rx_ports(struct net_device *dev, bool push)
 	}
 }
 
+static void vxlan_free_dev(struct net_device *dev)
+{
+	struct vxlan_dev *vxlan = netdev_priv(dev);
+	struct vxlan_config *cfg = rcu_dereference_protected(vxlan->cfg, 1);
+
+	RCU_INIT_POINTER(vxlan->cfg, NULL);
+	kfree(cfg);
+}
+
 /* Initialize the device structure. */
 static void vxlan_setup(struct net_device *dev)
 {
@@ -3381,6 +3464,8 @@ static void vxlan_setup(struct net_device *dev)
 	eth_hw_addr_random(dev);
 	ether_setup(dev);
 
+	dev->priv_destructor = vxlan_free_dev;
+
 	dev->needs_free_netdev = true;
 	SET_NETDEV_DEVTYPE(dev, &vxlan_type);
 
@@ -3555,9 +3640,11 @@ static int vxlan_get_link_ksettings(struct net_device *dev,
 				    struct ethtool_link_ksettings *cmd)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct net_device *lowerdev = __dev_get_by_index(vxlan->net,
-							 dst->remote_ifindex);
+	const struct vxlan_config *cfg;
+	struct net_device *lowerdev;
+
+	cfg = rtnl_dereference(vxlan->cfg);
+	lowerdev = __dev_get_by_index(vxlan->net, cfg->remote_ifindex);
 
 	if (!lowerdev) {
 		cmd->base.duplex = DUPLEX_UNKNOWN;
@@ -3663,21 +3750,22 @@ static struct vxlan_sock *vxlan_socket_create(struct net *net, bool ipv6,
 
 static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 {
-	bool metadata = vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	bool metadata = cfg->flags & VXLAN_F_COLLECT_METADATA;
 	struct vxlan_sock *vs = NULL;
 	struct vxlan_dev_node *node;
 	int l3mdev_index = 0;
 
 	ASSERT_RTNL();
 
-	if (vxlan->cfg.remote_ifindex)
+	if (cfg->remote_ifindex)
 		l3mdev_index = l3mdev_master_upper_ifindex_by_index(
-			vxlan->net, vxlan->cfg.remote_ifindex);
+			vxlan->net, cfg->remote_ifindex);
 
-	if (!vxlan->cfg.no_share) {
+	if (!cfg->no_share) {
 		rcu_read_lock();
 		vs = vxlan_find_sock(vxlan->net, ipv6 ? AF_INET6 : AF_INET,
-				     vxlan->cfg.dst_port, vxlan->cfg.flags,
+				     cfg->dst_port, cfg->flags,
 				     l3mdev_index);
 		if (vs && !refcount_inc_not_zero(&vs->refcnt)) {
 			rcu_read_unlock();
@@ -3687,7 +3775,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 	}
 	if (!vs)
 		vs = vxlan_socket_create(vxlan->net, ipv6,
-					 vxlan->cfg.dst_port, vxlan->cfg.flags,
+					 cfg->dst_port, cfg->flags,
 					 l3mdev_index);
 	if (IS_ERR(vs))
 		return PTR_ERR(vs);
@@ -3702,7 +3790,7 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 		node = &vxlan->hlist4;
 	}
 
-	if (metadata && (vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (metadata && (cfg->flags & VXLAN_F_VNIFILTER))
 		vxlan_vs_add_vnigrp(vxlan, vs, ipv6);
 	else
 		vxlan_vs_add_dev(vs, vxlan, node);
@@ -3712,11 +3800,14 @@ static int __vxlan_sock_add(struct vxlan_dev *vxlan, bool ipv6)
 
 static int vxlan_sock_add(struct vxlan_dev *vxlan)
 {
-	bool metadata = vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA;
-	bool ipv6 = vxlan->cfg.flags & VXLAN_F_IPV6 || metadata;
-	bool ipv4 = !ipv6 || metadata;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	bool metadata, ipv6, ipv4;
 	int ret = 0;
 
+	metadata = cfg->flags & VXLAN_F_COLLECT_METADATA;
+	ipv6 = (cfg->flags & VXLAN_F_IPV6) || metadata;
+	ipv4 = !ipv6 || metadata;
+
 	RCU_INIT_POINTER(vxlan->vn4_sock, NULL);
 #if IS_ENABLED(CONFIG_IPV6)
 	RCU_INIT_POINTER(vxlan->vn6_sock, NULL);
@@ -3734,28 +3825,33 @@ static int vxlan_sock_add(struct vxlan_dev *vxlan)
 }
 
 int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,
-		     struct vxlan_config *conf, __be32 vni)
+		     const struct vxlan_config *conf, __be32 vni)
 {
 	struct vxlan_net *vn = net_generic(src_net, vxlan_net_id);
 	struct vxlan_dev *tmp;
 
 	list_for_each_entry(tmp, &vn->vxlan_list, next) {
+		const struct vxlan_config *tmp_cfg;
+
 		if (tmp == vxlan)
 			continue;
-		if (tmp->cfg.flags & VXLAN_F_VNIFILTER) {
+
+		tmp_cfg = rtnl_dereference(tmp->cfg);
+
+		if (tmp_cfg->flags & VXLAN_F_VNIFILTER) {
 			if (!vxlan_vnifilter_lookup(tmp, vni))
 				continue;
-		} else if (tmp->cfg.vni != vni) {
+		} else if (tmp_cfg->vni != vni) {
 			continue;
 		}
-		if (tmp->cfg.dst_port != conf->dst_port)
+		if (tmp_cfg->dst_port != conf->dst_port)
 			continue;
-		if ((tmp->cfg.flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=
+		if ((tmp_cfg->flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)) !=
 		    (conf->flags & (VXLAN_F_RCV_FLAGS | VXLAN_F_IPV6)))
 			continue;
 
 		if ((conf->flags & VXLAN_F_IPV6_LINKLOCAL) &&
-		    tmp->cfg.remote_ifindex != conf->remote_ifindex)
+		    tmp_cfg->remote_ifindex != conf->remote_ifindex)
 			continue;
 
 		return -EEXIST;
@@ -3886,6 +3982,13 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,
 			return -EINVAL;
 		}
 
+		if ((conf->flags & VXLAN_F_VNIFILTER) && old &&
+		    vxlan_vnifilter_has_multicast(old)) {
+			NL_SET_ERR_MSG(extack,
+				       "Local interface required for multicast remote group");
+			return -EINVAL;
+		}
+
 #if IS_ENABLED(CONFIG_IPV6)
 		if (conf->flags & VXLAN_F_IPV6_LINKLOCAL) {
 			NL_SET_ERR_MSG(extack,
@@ -3917,16 +4020,16 @@ static int vxlan_config_validate(struct net *src_net, struct vxlan_config *conf,
 }
 
 static void vxlan_config_apply(struct net_device *dev,
-			       struct vxlan_config *conf,
+			       struct vxlan_config *new_cfg,
 			       struct net_device *lowerdev,
 			       struct net *src_net,
-			       bool changelink)
+			       bool changelink, bool lowerdev_changed)
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
 	unsigned short needed_headroom = ETH_HLEN;
+	struct vxlan_config *old_cfg;
 	int max_mtu = ETH_MAX_MTU;
-	u32 flags = conf->flags;
+	u32 flags = new_cfg->flags;
 
 	if (!changelink) {
 		if (flags & VXLAN_F_GPE)
@@ -3934,19 +4037,14 @@ static void vxlan_config_apply(struct net_device *dev,
 		else
 			vxlan_ether_setup(dev);
 
-		if (conf->mtu)
-			dev->mtu = conf->mtu;
+		if (new_cfg->mtu)
+			WRITE_ONCE(dev->mtu, new_cfg->mtu);
 
 		vxlan->net = src_net;
 	}
 
-	dst->remote_vni = conf->vni;
-
-	memcpy(&dst->remote_ip, &conf->remote_ip, sizeof(conf->remote_ip));
-
+	dev->needed_tailroom = 0;
 	if (lowerdev) {
-		dst->remote_ifindex = conf->remote_ifindex;
-
 		netif_inherit_tso_max(dev, lowerdev);
 
 		needed_headroom = lowerdev->hard_header_len;
@@ -3954,23 +4052,27 @@ static void vxlan_config_apply(struct net_device *dev,
 
 		dev->needed_tailroom = lowerdev->needed_tailroom;
 
-		max_mtu = lowerdev->mtu - vxlan_headroom(flags);
+		max_mtu = READ_ONCE(lowerdev->mtu) - vxlan_headroom(flags);
 		if (max_mtu < ETH_MIN_MTU)
 			max_mtu = ETH_MIN_MTU;
 
-		if (!changelink && !conf->mtu)
-			dev->mtu = max_mtu;
+		if (!changelink && !new_cfg->mtu)
+			WRITE_ONCE(dev->mtu, max_mtu);
 	}
 
-	if (dev->mtu > max_mtu)
-		dev->mtu = max_mtu;
+	/* A changelink leaving the lower device alone must not shrink the mtu */
+	if (lowerdev_changed && READ_ONCE(dev->mtu) > max_mtu)
+		WRITE_ONCE(dev->mtu, max_mtu);
 
 	if (flags & VXLAN_F_COLLECT_METADATA)
 		flags |= VXLAN_F_IPV6;
 	needed_headroom += vxlan_headroom(flags);
 	dev->needed_headroom = needed_headroom;
 
-	memcpy(&vxlan->cfg, conf, sizeof(*conf));
+	old_cfg = rtnl_dereference(vxlan->cfg);
+	rcu_assign_pointer(vxlan->cfg, new_cfg);
+	if (old_cfg)
+		kfree_rcu(old_cfg, rcu);
 }
 
 static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
@@ -3979,13 +4081,18 @@ static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,
 {
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct net_device *lowerdev;
+	struct vxlan_config *new_cfg;
 	int ret;
 
 	ret = vxlan_config_validate(src_net, conf, &lowerdev, vxlan, extack);
 	if (ret)
 		return ret;
 
-	vxlan_config_apply(dev, conf, lowerdev, src_net, false);
+	new_cfg = kmemdup(conf, sizeof(*conf), GFP_KERNEL);
+	if (!new_cfg)
+		return -ENOMEM;
+
+	vxlan_config_apply(dev, new_cfg, lowerdev, src_net, false, true);
 
 	return 0;
 }
@@ -3997,22 +4104,25 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct net_device *remote_dev = NULL;
-	struct vxlan_rdst *dst;
+	const struct vxlan_config *cfg;
 	int err;
 
-	dst = &vxlan->default_dst;
 	err = vxlan_dev_configure(net, dev, conf, extack);
 	if (err)
 		return err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	dev->ethtool_ops = &vxlan_ethtool_ops;
 
 	err = register_netdevice(dev);
-	if (err)
+	if (err) {
+		vxlan_free_dev(dev);
 		return err;
+	}
 
-	if (dst->remote_ifindex) {
-		remote_dev = __dev_get_by_index(net, dst->remote_ifindex);
+	if (cfg->remote_ifindex) {
+		remote_dev = __dev_get_by_index(net, cfg->remote_ifindex);
 		if (!remote_dev) {
 			err = -ENODEV;
 			goto unregister;
@@ -4022,7 +4132,7 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 		if (err)
 			goto unregister;
 
-		dst->remote_dev = remote_dev;
+		vxlan->lowerdev = remote_dev;
 	}
 
 	err = rtnl_configure_link(dev, NULL, 0, NULL);
@@ -4030,16 +4140,18 @@ static int vxlan_dev_create(struct net *net, struct net_device *dev,
 		goto unlink;
 
 	/* create an fdb entry for a valid default destination */
-	if (!vxlan_addr_any(&dst->remote_ip)) {
+	if (!vxlan_addr_any(&cfg->remote_ip)) {
+		union vxlan_addr rip = cfg->remote_ip;
+
 		spin_lock_bh(&vxlan->hash_lock);
 		err = vxlan_fdb_update(vxlan, all_zeros_mac,
-				       &dst->remote_ip,
+				       &rip,
 				       NUD_REACHABLE | NUD_PERMANENT,
 				       NLM_F_EXCL | NLM_F_CREATE,
-				       vxlan->cfg.dst_port,
-				       dst->remote_vni,
-				       dst->remote_vni,
-				       dst->remote_ifindex,
+				       cfg->dst_port,
+				       cfg->vni,
+				       cfg->vni,
+				       cfg->remote_ifindex,
 				       NTF_SELF, 0, true, extack);
 		spin_unlock_bh(&vxlan->hash_lock);
 		if (err)
@@ -4100,8 +4212,12 @@ static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],
 	memset(conf, 0, sizeof(*conf));
 
 	/* if changelink operation, start with old existing cfg */
-	if (changelink)
-		memcpy(conf, &vxlan->cfg, sizeof(*conf));
+	if (changelink) {
+		const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+
+		if (cfg)
+			memcpy(conf, cfg, sizeof(*conf));
+	}
 
 	if (data[IFLA_VXLAN_ID]) {
 		__be32 vni = cpu_to_be32(nla_get_u32(data[IFLA_VXLAN_ID]));
@@ -4447,17 +4563,19 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],
 			    struct nlattr *data[],
 			    struct netlink_ext_ack *extack)
 {
+	bool lowerdev_changed, rem_ip_changed, change_igmp;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
-	bool rem_ip_changed, change_igmp;
+	const struct vxlan_config *cfg;
+	struct vxlan_config *new_cfg;
 	struct net_device *lowerdev;
 	struct vxlan_config conf;
-	struct vxlan_rdst *dst;
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
+
 	if (!rtnl_dev_link_net_capable(dev, vxlan->net))
 		return -EPERM;
 
-	dst = &vxlan->default_dst;
 	err = vxlan_nl2conf(tb, data, dev, &conf, true, extack);
 	if (err)
 		return err;
@@ -4467,76 +4585,108 @@ static int vxlan_changelink(struct net_device *dev, struct nlattr *tb[],
 	if (err)
 		return err;
 
-	if (dst->remote_dev == lowerdev)
-		lowerdev = NULL;
+	new_cfg = kmemdup(&conf, sizeof(conf), GFP_KERNEL);
+	if (!new_cfg)
+		return -ENOMEM;
 
-	err = netdev_adjacent_change_prepare(dst->remote_dev, lowerdev, dev,
-					     extack);
-	if (err)
-		return err;
+	lowerdev_changed = vxlan->lowerdev != lowerdev;
+	if (lowerdev_changed) {
+		err = netdev_adjacent_change_prepare(vxlan->lowerdev, lowerdev,
+						     dev, extack);
+		if (err) {
+			kfree(new_cfg);
+			return err;
+		}
+	}
 
-	rem_ip_changed = !vxlan_addr_equal(&conf.remote_ip, &dst->remote_ip);
+	rem_ip_changed = !vxlan_addr_equal(&conf.remote_ip, &cfg->remote_ip);
 	change_igmp = vxlan->dev->flags & IFF_UP &&
 		      (rem_ip_changed ||
-		       dst->remote_ifindex != conf.remote_ifindex);
+		       cfg->remote_ifindex != conf.remote_ifindex);
 
 	/* handle default dst entry */
-	if (rem_ip_changed) {
+	if (rem_ip_changed || cfg->remote_ifindex != conf.remote_ifindex) {
 		spin_lock_bh(&vxlan->hash_lock);
 		if (!vxlan_addr_any(&conf.remote_ip)) {
 			err = vxlan_fdb_update(vxlan, all_zeros_mac,
 					       &conf.remote_ip,
 					       NUD_REACHABLE | NUD_PERMANENT,
 					       NLM_F_APPEND | NLM_F_CREATE,
-					       vxlan->cfg.dst_port,
+					       cfg->dst_port,
 					       conf.vni, conf.vni,
 					       conf.remote_ifindex,
 					       NTF_SELF, 0, true, extack);
 			if (err) {
 				spin_unlock_bh(&vxlan->hash_lock);
-				netdev_adjacent_change_abort(dst->remote_dev,
-							     lowerdev, dev);
+				if (lowerdev_changed)
+					netdev_adjacent_change_abort(vxlan->lowerdev,
+								     lowerdev, dev);
+				kfree(new_cfg);
 				return err;
 			}
 		}
-		if (!vxlan_addr_any(&dst->remote_ip))
+		if (!vxlan_addr_any(&cfg->remote_ip))
 			__vxlan_fdb_delete(vxlan, all_zeros_mac,
-					   dst->remote_ip,
-					   vxlan->cfg.dst_port,
-					   dst->remote_vni,
-					   dst->remote_vni,
-					   dst->remote_ifindex,
+					   cfg->remote_ip,
+					   cfg->dst_port,
+					   cfg->vni,
+					   cfg->vni,
+					   cfg->remote_ifindex,
 					   true);
 		spin_unlock_bh(&vxlan->hash_lock);
 
-		/* If vni filtering device, also update fdb entries of
-		 * all vnis that were using default remote ip
+		/* If vni filtering device, also update default fdb entries of
+		 * all vnis
 		 */
-		if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
-			err = vxlan_vnilist_update_group(vxlan, &dst->remote_ip,
-							 &conf.remote_ip, extack);
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
+			err = vxlan_vnilist_update_group(vxlan, &cfg->remote_ip,
+							 &conf.remote_ip,
+							 cfg->remote_ifindex,
+							 conf.remote_ifindex,
+							 extack);
 			if (err) {
-				netdev_adjacent_change_abort(dst->remote_dev,
-							     lowerdev, dev);
+				vxlan_update_default_fdb_entry(vxlan, conf.vni,
+							       &conf.remote_ip,
+							       &cfg->remote_ip,
+							       conf.remote_ifindex,
+							       cfg->remote_ifindex,
+							       NULL);
+				if (lowerdev_changed)
+					netdev_adjacent_change_abort(vxlan->lowerdev,
+								     lowerdev, dev);
+				kfree(new_cfg);
 				return err;
 			}
 		}
 	}
 
-	if (change_igmp && vxlan_addr_multicast(&dst->remote_ip))
+	if (change_igmp &&
+	    (vxlan_addr_multicast(&cfg->remote_ip) ||
+	     (cfg->flags & VXLAN_F_VNIFILTER)))
 		err = vxlan_multicast_leave(vxlan);
 
-	if (netif_running(dev) && conf.age_interval != vxlan->cfg.age_interval)
+	if (netif_running(dev) && conf.age_interval != cfg->age_interval)
 		mod_timer(&vxlan->age_timer, jiffies);
 
-	netdev_adjacent_change_commit(dst->remote_dev, lowerdev, dev);
-	if (lowerdev && lowerdev != dst->remote_dev)
-		dst->remote_dev = lowerdev;
-	vxlan_config_apply(dev, &conf, lowerdev, vxlan->net, true);
+	if (lowerdev_changed) {
+		if (lowerdev)
+			netdev_adjacent_change_commit(vxlan->lowerdev, lowerdev,
+						      dev);
+		else
+			netdev_upper_dev_unlink(vxlan->lowerdev, dev);
+		vxlan->lowerdev = lowerdev;
+	}
+	vxlan_config_apply(dev, new_cfg, lowerdev, vxlan->net, true,
+			   lowerdev_changed);
+
+	if (change_igmp &&
+	    (vxlan_addr_multicast(&new_cfg->remote_ip) ||
+	     (new_cfg->flags & VXLAN_F_VNIFILTER))) {
+		int join_err = vxlan_multicast_join(vxlan);
 
-	if (!err && change_igmp &&
-	    vxlan_addr_multicast(&dst->remote_ip))
-		err = vxlan_multicast_join(vxlan);
+		if (join_err)
+			err = join_err;
+	}
 
 	return err;
 }
@@ -4550,8 +4700,8 @@ static void vxlan_dellink(struct net_device *dev, struct list_head *head)
 
 	list_del(&vxlan->next);
 	unregister_netdevice_queue(dev, head);
-	if (vxlan->default_dst.remote_dev)
-		netdev_upper_dev_unlink(vxlan->default_dst.remote_dev, dev);
+	if (vxlan->lowerdev)
+		netdev_upper_dev_unlink(vxlan->lowerdev, dev);
 }
 
 static size_t vxlan_get_size(const struct net_device *dev)
@@ -4595,110 +4745,114 @@ static size_t vxlan_get_size(const struct net_device *dev)
 static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)
 {
 	const struct vxlan_dev *vxlan = netdev_priv(dev);
-	const struct vxlan_rdst *dst = &vxlan->default_dst;
-	struct ifla_vxlan_port_range ports = {
-		.low =  htons(vxlan->cfg.port_min),
-		.high = htons(vxlan->cfg.port_max),
-	};
+	struct ifla_vxlan_port_range ports;
+	const struct vxlan_config *cfg;
+
+	rcu_read_lock();
+	cfg = rcu_dereference(vxlan->cfg);
 
-	if (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(dst->remote_vni)))
+	if (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(cfg->vni)))
 		goto nla_put_failure;
 
-	if (!vxlan_addr_any(&dst->remote_ip)) {
-		if (dst->remote_ip.sa.sa_family == AF_INET) {
+	if (!vxlan_addr_any(&cfg->remote_ip)) {
+		if (cfg->remote_ip.sa.sa_family == AF_INET) {
 			if (nla_put_in_addr(skb, IFLA_VXLAN_GROUP,
-					    dst->remote_ip.sin.sin_addr.s_addr))
+					    cfg->remote_ip.sin.sin_addr.s_addr))
 				goto nla_put_failure;
 #if IS_ENABLED(CONFIG_IPV6)
 		} else {
 			if (nla_put_in6_addr(skb, IFLA_VXLAN_GROUP6,
-					     &dst->remote_ip.sin6.sin6_addr))
+					     &cfg->remote_ip.sin6.sin6_addr))
 				goto nla_put_failure;
 #endif
 		}
 	}
 
-	if (dst->remote_ifindex && nla_put_u32(skb, IFLA_VXLAN_LINK, dst->remote_ifindex))
+	if (cfg->remote_ifindex && nla_put_u32(skb, IFLA_VXLAN_LINK, cfg->remote_ifindex))
 		goto nla_put_failure;
 
-	if (!vxlan_addr_any(&vxlan->cfg.saddr)) {
-		if (vxlan->cfg.saddr.sa.sa_family == AF_INET) {
+	if (!vxlan_addr_any(&cfg->saddr)) {
+		if (cfg->saddr.sa.sa_family == AF_INET) {
 			if (nla_put_in_addr(skb, IFLA_VXLAN_LOCAL,
-					    vxlan->cfg.saddr.sin.sin_addr.s_addr))
+					    cfg->saddr.sin.sin_addr.s_addr))
 				goto nla_put_failure;
 #if IS_ENABLED(CONFIG_IPV6)
 		} else {
 			if (nla_put_in6_addr(skb, IFLA_VXLAN_LOCAL6,
-					     &vxlan->cfg.saddr.sin6.sin6_addr))
+					     &cfg->saddr.sin6.sin6_addr))
 				goto nla_put_failure;
 #endif
 		}
 	}
 
-	if (nla_put_u8(skb, IFLA_VXLAN_TTL, vxlan->cfg.ttl) ||
+	if (nla_put_u8(skb, IFLA_VXLAN_TTL, cfg->ttl) ||
 	    nla_put_u8(skb, IFLA_VXLAN_TTL_INHERIT,
-		       !!(vxlan->cfg.flags & VXLAN_F_TTL_INHERIT)) ||
-	    nla_put_u8(skb, IFLA_VXLAN_TOS, vxlan->cfg.tos) ||
-	    nla_put_u8(skb, IFLA_VXLAN_DF, vxlan->cfg.df) ||
-	    nla_put_be32(skb, IFLA_VXLAN_LABEL, vxlan->cfg.label) ||
-	    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, vxlan->cfg.label_policy) ||
+		       !!(cfg->flags & VXLAN_F_TTL_INHERIT)) ||
+	    nla_put_u8(skb, IFLA_VXLAN_TOS, cfg->tos) ||
+	    nla_put_u8(skb, IFLA_VXLAN_DF, cfg->df) ||
+	    nla_put_be32(skb, IFLA_VXLAN_LABEL, cfg->label) ||
+	    nla_put_u32(skb, IFLA_VXLAN_LABEL_POLICY, cfg->label_policy) ||
 	    nla_put_u8(skb, IFLA_VXLAN_LEARNING,
-		       !!(vxlan->cfg.flags & VXLAN_F_LEARN)) ||
+		       !!(cfg->flags & VXLAN_F_LEARN)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_PROXY,
-		       !!(vxlan->cfg.flags & VXLAN_F_PROXY)) ||
+		       !!(cfg->flags & VXLAN_F_PROXY)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_RSC,
-		       !!(vxlan->cfg.flags & VXLAN_F_RSC)) ||
+		       !!(cfg->flags & VXLAN_F_RSC)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_L2MISS,
-		       !!(vxlan->cfg.flags & VXLAN_F_L2MISS)) ||
+		       !!(cfg->flags & VXLAN_F_L2MISS)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_L3MISS,
-		       !!(vxlan->cfg.flags & VXLAN_F_L3MISS)) ||
+		       !!(cfg->flags & VXLAN_F_L3MISS)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_COLLECT_METADATA,
-		       !!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA)) ||
-	    nla_put_u32(skb, IFLA_VXLAN_AGEING, vxlan->cfg.age_interval) ||
-	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, vxlan->cfg.addrmax) ||
-	    nla_put_be16(skb, IFLA_VXLAN_PORT, vxlan->cfg.dst_port) ||
+		       !!(cfg->flags & VXLAN_F_COLLECT_METADATA)) ||
+	    nla_put_u32(skb, IFLA_VXLAN_AGEING, cfg->age_interval) ||
+	    nla_put_u32(skb, IFLA_VXLAN_LIMIT, cfg->addrmax) ||
+	    nla_put_be16(skb, IFLA_VXLAN_PORT, cfg->dst_port) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_CSUM,
-		       !(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM_TX)) ||
+		       !(cfg->flags & VXLAN_F_UDP_ZERO_CSUM_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_TX,
-		       !!(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
+		       !!(cfg->flags & VXLAN_F_UDP_ZERO_CSUM6_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_UDP_ZERO_CSUM6_RX,
-		       !!(vxlan->cfg.flags & VXLAN_F_UDP_ZERO_CSUM6_RX)) ||
+		       !!(cfg->flags & VXLAN_F_UDP_ZERO_CSUM6_RX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_TX,
-		       !!(vxlan->cfg.flags & VXLAN_F_REMCSUM_TX)) ||
+		       !!(cfg->flags & VXLAN_F_REMCSUM_TX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_REMCSUM_RX,
-		       !!(vxlan->cfg.flags & VXLAN_F_REMCSUM_RX)) ||
+		       !!(cfg->flags & VXLAN_F_REMCSUM_RX)) ||
 	    nla_put_u8(skb, IFLA_VXLAN_LOCALBYPASS,
-		       !!(vxlan->cfg.flags & VXLAN_F_LOCALBYPASS)))
+		       !!(cfg->flags & VXLAN_F_LOCALBYPASS)))
 		goto nla_put_failure;
 
+	ports.low = htons(cfg->port_min);
+	ports.high = htons(cfg->port_max);
 	if (nla_put(skb, IFLA_VXLAN_PORT_RANGE, sizeof(ports), &ports))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_GBP &&
+	if (cfg->flags & VXLAN_F_GBP &&
 	    nla_put_flag(skb, IFLA_VXLAN_GBP))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_GPE &&
+	if (cfg->flags & VXLAN_F_GPE &&
 	    nla_put_flag(skb, IFLA_VXLAN_GPE))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_REMCSUM_NOPARTIAL &&
+	if (cfg->flags & VXLAN_F_REMCSUM_NOPARTIAL &&
 	    nla_put_flag(skb, IFLA_VXLAN_REMCSUM_NOPARTIAL))
 		goto nla_put_failure;
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER &&
+	if (cfg->flags & VXLAN_F_VNIFILTER &&
 	    nla_put_u8(skb, IFLA_VXLAN_VNIFILTER,
-		       !!(vxlan->cfg.flags & VXLAN_F_VNIFILTER)))
+		       !!(cfg->flags & VXLAN_F_VNIFILTER)))
 		goto nla_put_failure;
 
 	if (nla_put(skb, IFLA_VXLAN_RESERVED_BITS,
-		    sizeof(vxlan->cfg.reserved_bits),
-		    &vxlan->cfg.reserved_bits))
+		    sizeof(cfg->reserved_bits),
+		    &cfg->reserved_bits))
 		goto nla_put_failure;
 
+	rcu_read_unlock();
 	return 0;
 
 nla_put_failure:
+	rcu_read_unlock();
 	return -EMSGSIZE;
 }
 
@@ -4731,7 +4885,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,
 	LIST_HEAD(list_kill);
 
 	list_for_each_entry_safe(vxlan, next, &vn->vxlan_list, next) {
-		struct vxlan_rdst *dst = &vxlan->default_dst;
+		const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 
 		/* In case we created vxlan device with carrier
 		 * and we loose the carrier due to module unload
@@ -4739,7 +4893,7 @@ static void vxlan_handle_lowerdev_unregister(struct vxlan_net *vn,
 		 * cases, it's not necessary and remote_ifindex
 		 * is 0 here, so no matches.
 		 */
-		if (dst->remote_ifindex == dev->ifindex)
+		if (cfg->remote_ifindex == dev->ifindex)
 			vxlan_dellink(vxlan->dev, &list_kill);
 	}
 
@@ -4752,6 +4906,10 @@ static int vxlan_netdevice_event(struct notifier_block *unused,
 	struct net_device *dev = netdev_notifier_info_to_dev(ptr);
 	struct vxlan_net *vn = net_generic(dev_net(dev), vxlan_net_id);
 
+	if ((event == NETDEV_REGISTER || event == NETDEV_UNREGISTER) &&
+	    netif_is_vxlan(dev))
+		vxlan_vnifilter_seq_inc(dev_net(dev));
+
 	if (event == NETDEV_UNREGISTER)
 		vxlan_handle_lowerdev_unregister(vn, dev);
 	else if (event == NETDEV_UDP_TUNNEL_PUSH_INFO)
diff --git a/drivers/net/vxlan/vxlan_mdb.c b/drivers/net/vxlan/vxlan_mdb.c
index 841f42ffecb9a..c1a0551990555 100644
--- a/drivers/net/vxlan/vxlan_mdb.c
+++ b/drivers/net/vxlan/vxlan_mdb.c
@@ -165,6 +165,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 				     const struct vxlan_mdb_entry *mdb_entry,
 				     const struct vxlan_mdb_remote *remote)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	struct vxlan_rdst *rd = rtnl_dereference(remote->rd);
 	struct br_mdb_entry e;
 	struct nlattr *nest;
@@ -189,12 +190,12 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 	    vxlan_nla_put_addr(skb, MDBA_MDB_EATTR_DST, &rd->remote_ip))
 		goto nest_err;
 
-	if (rd->remote_port && rd->remote_port != vxlan->cfg.dst_port &&
+	if (rd->remote_port && rd->remote_port != cfg->dst_port &&
 	    nla_put_u16(skb, MDBA_MDB_EATTR_DST_PORT,
 			be16_to_cpu(rd->remote_port)))
 		goto nest_err;
 
-	if (rd->remote_vni != vxlan->default_dst.remote_vni &&
+	if (rd->remote_vni != cfg->vni &&
 	    nla_put_u32(skb, MDBA_MDB_EATTR_VNI, be32_to_cpu(rd->remote_vni)))
 		goto nest_err;
 
@@ -202,7 +203,7 @@ static int vxlan_mdb_entry_info_fill(const struct vxlan_dev *vxlan,
 	    nla_put_u32(skb, MDBA_MDB_EATTR_IFINDEX, rd->remote_ifindex))
 		goto nest_err;
 
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) &&
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) &&
 	    mdb_entry->key.vni && nla_put_u32(skb, MDBA_MDB_EATTR_SRC_VNI,
 					      be32_to_cpu(mdb_entry->key.vni)))
 		goto nest_err;
@@ -613,16 +614,19 @@ static int vxlan_mdb_config_init(struct vxlan_mdb_config *cfg,
 {
 	struct br_mdb_entry *entry = nla_data(tb[MDBA_SET_ENTRY]);
 	struct vxlan_dev *vxlan = netdev_priv(dev);
+	const struct vxlan_config *vcfg;
+
+	vcfg = rtnl_dereference(vxlan->cfg);
 
 	memset(cfg, 0, sizeof(*cfg));
 	cfg->vxlan = vxlan;
-	cfg->group.vni = vxlan->default_dst.remote_vni;
+	cfg->group.vni = vcfg->vni;
 	INIT_LIST_HEAD(&cfg->src_list);
 	cfg->nlflags = nlmsg_flags;
 	cfg->filter_mode = MCAST_EXCLUDE;
 	cfg->rt_protocol = RTPROT_STATIC;
-	cfg->remote_vni = vxlan->default_dst.remote_vni;
-	cfg->remote_port = vxlan->cfg.dst_port;
+	cfg->remote_vni = vcfg->vni;
+	cfg->remote_port = vcfg->dst_port;
 
 	if (entry->ifindex != dev->ifindex) {
 		NL_SET_ERR_MSG_MOD(extack, "Port net device must be the VXLAN net device");
@@ -957,11 +961,12 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,
 			    const struct vxlan_mdb_entry *mdb_entry,
 			    const struct vxlan_mdb_remote *remote)
 {
+	const struct vxlan_config *cfg = rcu_dereference_rtnl(vxlan->cfg);
 	const struct vxlan_mdb_entry_key *group = &mdb_entry->key;
 	struct vxlan_rdst *rd = rtnl_dereference(remote->rd);
 	size_t nlmsg_size;
 
-		     /* MDBA_MDB_ENTRY_INFO */
+	/* MDBA_MDB_ENTRY_INFO */
 	nlmsg_size = nla_total_size(sizeof(struct br_mdb_entry)) +
 		     /* MDBA_MDB_EATTR_TIMER */
 		     nla_total_size(sizeof(u32));
@@ -978,16 +983,16 @@ vxlan_mdb_nlmsg_remote_size(const struct vxlan_dev *vxlan,
 	/* MDBA_MDB_EATTR_DST */
 	nlmsg_size += nla_total_size(vxlan_addr_size(&rd->remote_ip));
 	/* MDBA_MDB_EATTR_DST_PORT */
-	if (rd->remote_port && rd->remote_port != vxlan->cfg.dst_port)
+	if (rd->remote_port && rd->remote_port != cfg->dst_port)
 		nlmsg_size += nla_total_size(sizeof(u16));
 	/* MDBA_MDB_EATTR_VNI */
-	if (rd->remote_vni != vxlan->default_dst.remote_vni)
+	if (rd->remote_vni != cfg->vni)
 		nlmsg_size += nla_total_size(sizeof(u32));
 	/* MDBA_MDB_EATTR_IFINDEX */
 	if (rd->remote_ifindex)
 		nlmsg_size += nla_total_size(sizeof(u32));
 	/* MDBA_MDB_EATTR_SRC_VNI */
-	if ((vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA) && group->vni)
+	if ((cfg->flags & VXLAN_F_COLLECT_METADATA) && group->vni)
 		nlmsg_size += nla_total_size(sizeof(u32));
 
 	return nlmsg_size;
@@ -1216,7 +1221,7 @@ vxlan_mdb_entry_get(struct vxlan_dev *vxlan,
 		goto err_free_entry;
 
 	if (hlist_is_singular_node(&mdb_entry->mdb_node, &vxlan->mdb_list))
-		vxlan->cfg.flags |= VXLAN_F_MDB;
+		set_bit(VXLAN_DEV_F_MDB, &vxlan->flags);
 
 	return mdb_entry;
 
@@ -1233,7 +1238,7 @@ static void vxlan_mdb_entry_put(struct vxlan_dev *vxlan,
 		return;
 
 	if (hlist_is_singular_node(&mdb_entry->mdb_node, &vxlan->mdb_list))
-		vxlan->cfg.flags &= ~VXLAN_F_MDB;
+		clear_bit(VXLAN_DEV_F_MDB, &vxlan->flags);
 
 	rhashtable_remove_fast(&vxlan->mdb_tbl, &mdb_entry->rhnode,
 			       vxlan_mdb_rht_params);
@@ -1483,11 +1488,13 @@ static int vxlan_mdb_get_parse(struct net_device *dev, struct nlattr *tb[],
 {
 	struct br_mdb_entry *entry = nla_data(tb[MDBA_GET_ENTRY]);
 	struct nlattr *mdbe_attrs[MDBE_ATTR_MAX + 1];
+	const struct vxlan_config *cfg;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	int err;
 
+	cfg = rtnl_dereference(vxlan->cfg);
 	memset(group, 0, sizeof(*group));
-	group->vni = vxlan->default_dst.remote_vni;
+	group->vni = cfg->vni;
 
 	if (!tb[MDBA_GET_ENTRY_ATTRS]) {
 		vxlan_mdb_group_set(group, entry, NULL);
@@ -1621,6 +1628,7 @@ int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,
 }
 
 struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
+						const struct vxlan_config *cfg,
 						struct sk_buff *skb,
 						__be32 src_vni)
 {
@@ -1634,8 +1642,8 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
 	/* When not in collect metadata mode, 'src_vni' is zero, but MDB
 	 * entries are stored with the VNI of the VXLAN device.
 	 */
-	if (!(vxlan->cfg.flags & VXLAN_F_COLLECT_METADATA))
-		src_vni = vxlan->default_dst.remote_vni;
+	if (!(cfg->flags & VXLAN_F_COLLECT_METADATA))
+		src_vni = cfg->vni;
 
 	memset(&group, 0, sizeof(group));
 	group.vni = src_vni;
@@ -1700,6 +1708,7 @@ struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
 }
 
 netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg,
 			   const struct vxlan_mdb_entry *mdb_entry,
 			   struct sk_buff *skb)
 {
@@ -1721,12 +1730,12 @@ netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
 
 		skb1 = skb_clone(skb, GFP_ATOMIC);
 		if (skb1)
-			vxlan_xmit_one(skb1, vxlan->dev, src_vni,
+			vxlan_xmit_one(skb1, vxlan->dev, cfg, src_vni,
 				       rcu_dereference(remote->rd), false);
 	}
 
 	if (fremote)
-		vxlan_xmit_one(skb, vxlan->dev, src_vni,
+		vxlan_xmit_one(skb, vxlan->dev, cfg, src_vni,
 			       rcu_dereference(fremote->rd), false);
 	else
 		kfree_skb_reason(skb, SKB_DROP_REASON_NO_TX_TARGET);
@@ -1757,7 +1766,7 @@ void vxlan_mdb_fini(struct vxlan_dev *vxlan)
 	struct vxlan_mdb_flush_desc desc = {};
 
 	vxlan_mdb_flush(vxlan, &desc);
-	WARN_ON_ONCE(vxlan->cfg.flags & VXLAN_F_MDB);
+	WARN_ON_ONCE(test_bit(VXLAN_DEV_F_MDB, &vxlan->flags));
 	rhashtable_free_and_destroy(&vxlan->mdb_tbl, vxlan_mdb_check_empty,
 				    NULL);
 }
diff --git a/drivers/net/vxlan/vxlan_multicast.c b/drivers/net/vxlan/vxlan_multicast.c
index 3b75b48dc726d..a5ef6f66f00d4 100644
--- a/drivers/net/vxlan/vxlan_multicast.c
+++ b/drivers/net/vxlan/vxlan_multicast.c
@@ -14,11 +14,12 @@
 /* Update multicast group membership when first VNI on
  * multicast address is brought up
  */
-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		    int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &vxlan->default_dst.remote_ip);
-	int ifindex = (rifindex ? : vxlan->default_dst.remote_ifindex);
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *ip = (rip ? : &cfg->remote_ip);
+	int ifindex = (rifindex ? : cfg->remote_ifindex);
 	int ret = -EINVAL;
 	struct sock *sk;
 
@@ -47,11 +48,12 @@ int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
 	return ret;
 }
 
-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		     int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &vxlan->default_dst.remote_ip);
-	int ifindex = (rifindex ? : vxlan->default_dst.remote_ifindex);
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *ip = (rip ? : &cfg->remote_ip);
+	int ifindex = (rifindex ? : cfg->remote_ifindex);
 	int ret = -EINVAL;
 	struct sock *sk;
 
@@ -80,8 +82,8 @@ int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
 	return ret;
 }
 
-static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,
-				   union vxlan_addr *rip, int rifindex)
+static bool vxlan_group_used_match(const union vxlan_addr *ip, int ifindex,
+				   const union vxlan_addr *rip, int rifindex)
 {
 	if (!vxlan_addr_multicast(rip))
 		return false;
@@ -96,14 +98,16 @@ static bool vxlan_group_used_match(union vxlan_addr *ip, int ifindex,
 }
 
 static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
-					  union vxlan_addr *ip, int ifindex)
+					  const struct vxlan_config *cfg,
+					  const union vxlan_addr *ip,
+					  int ifindex)
 {
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
 	struct vxlan_vni_node *v, *tmp;
 
 	if (vxlan_group_used_match(ip, ifindex,
-				   &vxlan->default_dst.remote_ip,
-				   vxlan->default_dst.remote_ifindex))
+				   &cfg->remote_ip,
+				   cfg->remote_ifindex))
 		return true;
 
 	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
@@ -112,7 +116,7 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
 
 		if (vxlan_group_used_match(ip, ifindex,
 					   &v->remote_ip,
-					   vxlan->default_dst.remote_ifindex))
+					   cfg->remote_ifindex))
 			return true;
 	}
 
@@ -121,16 +125,17 @@ static bool vxlan_group_used_by_vnifilter(struct vxlan_dev *vxlan,
 
 /* See if multicast group is already in use by other ID */
 bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
-		      __be32 vni, union vxlan_addr *rip, int rifindex)
+		      __be32 vni, const union vxlan_addr *rip, int rifindex)
 {
-	union vxlan_addr *ip = (rip ? : &dev->default_dst.remote_ip);
-	int ifindex = (rifindex ? : dev->default_dst.remote_ifindex);
+	const struct vxlan_config *dev_cfg = rtnl_dereference(dev->cfg);
+	const union vxlan_addr *ip = (rip ? : &dev_cfg->remote_ip);
+	int ifindex = (rifindex ? : dev_cfg->remote_ifindex);
 	struct vxlan_dev *vxlan;
 	struct vxlan_sock *sock4;
 #if IS_ENABLED(CONFIG_IPV6)
 	struct vxlan_sock *sock6;
 #endif
-	unsigned short family = dev->default_dst.remote_ip.sa.sa_family;
+	unsigned short family = dev_cfg->remote_ip.sa.sa_family;
 
 	sock4 = rtnl_dereference(dev->vn4_sock);
 
@@ -147,6 +152,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 #endif
 
 	list_for_each_entry(vxlan, &vn->vxlan_list, next) {
+		const struct vxlan_config *cfg;
+
 		if (!netif_running(vxlan->dev) || vxlan == dev)
 			continue;
 
@@ -158,13 +165,15 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 		    rtnl_dereference(vxlan->vn6_sock) != sock6)
 			continue;
 #endif
-		if (vxlan->cfg.flags & VXLAN_F_VNIFILTER) {
-			if (!vxlan_group_used_by_vnifilter(vxlan, ip, ifindex))
+		cfg = rtnl_dereference(vxlan->cfg);
+
+		if (cfg->flags & VXLAN_F_VNIFILTER) {
+			if (!vxlan_group_used_by_vnifilter(vxlan, cfg, ip, ifindex))
 				continue;
 		} else {
 			if (!vxlan_group_used_match(ip, ifindex,
-						    &vxlan->default_dst.remote_ip,
-						    vxlan->default_dst.remote_ifindex))
+						    &cfg->remote_ip,
+						    cfg->remote_ifindex))
 				continue;
 		}
 
@@ -174,7 +183,8 @@ bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
 	return false;
 }
 
-static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
+static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan,
+				       const struct vxlan_config *cfg)
 {
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
 	struct vxlan_vni_node *v, *tmp, *vgood = NULL;
@@ -185,7 +195,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 			continue;
 		/* skip if address is same as default address */
 		if (vxlan_addr_equal(&v->remote_ip,
-				     &vxlan->default_dst.remote_ip))
+				     &cfg->remote_ip))
 			continue;
 		ret = vxlan_igmp_join(vxlan, &v->remote_ip, 0);
 		if (ret == -EADDRINUSE)
@@ -200,7 +210,7 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 			if (!vxlan_addr_multicast(&v->remote_ip))
 				continue;
 			if (vxlan_addr_equal(&v->remote_ip,
-					     &vxlan->default_dst.remote_ip))
+					     &cfg->remote_ip))
 				continue;
 			vxlan_igmp_leave(vxlan, &v->remote_ip, 0);
 			if (v == vgood)
@@ -211,7 +221,8 @@ static int vxlan_multicast_join_vnigrp(struct vxlan_dev *vxlan)
 	return ret;
 }
 
-static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
+static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan,
+					const struct vxlan_config *cfg)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
 	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
@@ -219,10 +230,17 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
 	int last_err = 0, ret;
 
 	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
-		if (vxlan_addr_multicast(&v->remote_ip) &&
-		    !vxlan_group_used(vn, vxlan, v->vni, &v->remote_ip,
+		if (!vxlan_addr_multicast(&v->remote_ip))
+			continue;
+		/* skip if address is same as default address */
+		if (vxlan_addr_equal(&v->remote_ip,
+				     &cfg->remote_ip))
+			continue;
+		if (!vxlan_group_used(vn, vxlan, v->vni, &v->remote_ip,
 				      0)) {
 			ret = vxlan_igmp_leave(vxlan, &v->remote_ip, 0);
+			if (ret == -EADDRNOTAVAIL)
+				ret = 0;
 			if (ret)
 				last_err = ret;
 		}
@@ -233,38 +251,42 @@ static int vxlan_multicast_leave_vnigrp(struct vxlan_dev *vxlan)
 
 int vxlan_multicast_join(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	int ret = 0;
 
-	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip)) {
-		ret = vxlan_igmp_join(vxlan, &vxlan->default_dst.remote_ip,
-				      vxlan->default_dst.remote_ifindex);
+	if (vxlan_addr_multicast(&cfg->remote_ip)) {
+		ret = vxlan_igmp_join(vxlan, &cfg->remote_ip,
+				      cfg->remote_ifindex);
 		if (ret == -EADDRINUSE)
 			ret = 0;
 		if (ret)
 			return ret;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
-		return vxlan_multicast_join_vnigrp(vxlan);
+	if (cfg->flags & VXLAN_F_VNIFILTER)
+		return vxlan_multicast_join_vnigrp(vxlan, cfg);
 
 	return 0;
 }
 
 int vxlan_multicast_leave(struct vxlan_dev *vxlan)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
 	int ret = 0;
 
-	if (vxlan_addr_multicast(&vxlan->default_dst.remote_ip) &&
+	if (vxlan_addr_multicast(&cfg->remote_ip) &&
 	    !vxlan_group_used(vn, vxlan, 0, NULL, 0)) {
-		ret = vxlan_igmp_leave(vxlan, &vxlan->default_dst.remote_ip,
-				       vxlan->default_dst.remote_ifindex);
+		ret = vxlan_igmp_leave(vxlan, &cfg->remote_ip,
+				       cfg->remote_ifindex);
+		if (ret == -EADDRNOTAVAIL)
+			ret = 0;
 		if (ret)
 			return ret;
 	}
 
-	if (vxlan->cfg.flags & VXLAN_F_VNIFILTER)
-		return vxlan_multicast_leave_vnigrp(vxlan);
+	if (cfg->flags & VXLAN_F_VNIFILTER)
+		return vxlan_multicast_leave_vnigrp(vxlan, cfg);
 
 	return 0;
 }
diff --git a/drivers/net/vxlan/vxlan_private.h b/drivers/net/vxlan/vxlan_private.h
index b1eec22163608..f0ee85f83732a 100644
--- a/drivers/net/vxlan/vxlan_private.h
+++ b/drivers/net/vxlan/vxlan_private.h
@@ -22,6 +22,8 @@ struct vxlan_net {
 	/* sock_list is protected by rtnl lock */
 	struct hlist_head sock_list[PORT_HASH_SIZE];
 	struct notifier_block nexthop_notifier_block;
+	/* Generation counter for RTM_GETTUNNEL dumps */
+	atomic_t vnifilter_seq;
 };
 
 struct vxlan_fdb_key {
@@ -177,6 +179,13 @@ vxlan_vnifilter_lookup(struct vxlan_dev *vxlan, __be32 vni)
 				      vxlan_vni_rht_params);
 }
 
+static inline void vxlan_vnifilter_seq_inc(const struct net *net)
+{
+	struct vxlan_net *vn = net_generic(net, vxlan_net_id);
+
+	atomic_inc(&vn->vnifilter_seq);
+}
+
 /* vxlan_core.c */
 int vxlan_fdb_create(struct vxlan_dev *vxlan,
 		     const u8 *mac, union vxlan_addr *ip,
@@ -195,9 +204,10 @@ int vxlan_fdb_update(struct vxlan_dev *vxlan,
 		     __u32 ifindex, __u16 ndm_flags, u32 nhid,
 		     bool swdev_notify, struct netlink_ext_ack *extack);
 void vxlan_xmit_one(struct sk_buff *skb, struct net_device *dev,
+		    const struct vxlan_config *cfg,
 		    __be32 default_vni, struct vxlan_rdst *rdst, bool did_rsc);
 int vxlan_vni_in_use(struct net *src_net, struct vxlan_dev *vxlan,
-		     struct vxlan_config *conf, __be32 vni);
+		     const struct vxlan_config *conf, __be32 vni);
 
 /* vxlan_vnifilter.c */
 int vxlan_vnigroup_init(struct vxlan_dev *vxlan);
@@ -205,7 +215,8 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan);
 
 int vxlan_vnifilter_init(void);
 void vxlan_vnifilter_uninit(void);
-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,
+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg, __be32 vni,
 			   struct vxlan_vni_node *vninode,
 			   int type, unsigned int len);
 
@@ -213,9 +224,16 @@ void vxlan_vs_add_vnigrp(struct vxlan_dev *vxlan,
 			 struct vxlan_sock *vs,
 			 bool ipv6);
 void vxlan_vs_del_vnigrp(struct vxlan_dev *vxlan);
+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan);
+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
+				   const union vxlan_addr *old_remote_ip,
+				   const union vxlan_addr *remote_ip,
+				   u32 old_ifindex, u32 new_ifindex,
+				   struct netlink_ext_ack *extack);
 int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
-			       union vxlan_addr *old_remote_ip,
-			       union vxlan_addr *new_remote_ip,
+			       const union vxlan_addr *old_remote_ip,
+			       const union vxlan_addr *new_remote_ip,
+			       u32 old_ifindex, u32 new_ifindex,
 			       struct netlink_ext_ack *extack);
 
 
@@ -223,10 +241,10 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
 int vxlan_multicast_join(struct vxlan_dev *vxlan);
 int vxlan_multicast_leave(struct vxlan_dev *vxlan);
 bool vxlan_group_used(struct vxlan_net *vn, struct vxlan_dev *dev,
-		      __be32 vni, union vxlan_addr *rip, int rifindex);
-int vxlan_igmp_join(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+		      __be32 vni, const union vxlan_addr *rip, int rifindex);
+int vxlan_igmp_join(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		    int rifindex);
-int vxlan_igmp_leave(struct vxlan_dev *vxlan, union vxlan_addr *rip,
+int vxlan_igmp_leave(struct vxlan_dev *vxlan, const union vxlan_addr *rip,
 		     int rifindex);
 
 /* vxlan_mdb.c */
@@ -241,9 +259,11 @@ int vxlan_mdb_del_bulk(struct net_device *dev, struct nlattr *tb[],
 int vxlan_mdb_get(struct net_device *dev, struct nlattr *tb[], u32 portid,
 		  u32 seq, struct netlink_ext_ack *extack);
 struct vxlan_mdb_entry *vxlan_mdb_entry_skb_get(struct vxlan_dev *vxlan,
+						const struct vxlan_config *cfg,
 						struct sk_buff *skb,
 						__be32 src_vni);
 netdev_tx_t vxlan_mdb_xmit(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg,
 			   const struct vxlan_mdb_entry *mdb_entry,
 			   struct sk_buff *skb);
 int vxlan_mdb_init(struct vxlan_dev *vxlan);
diff --git a/drivers/net/vxlan/vxlan_vnifilter.c b/drivers/net/vxlan/vxlan_vnifilter.c
index dd94085e08865..41095096b698d 100644
--- a/drivers/net/vxlan/vxlan_vnifilter.c
+++ b/drivers/net/vxlan/vxlan_vnifilter.c
@@ -171,13 +171,14 @@ static void vxlan_vnifilter_stats_add(struct vxlan_vni_node *vninode,
 	u64_stats_update_end(&pstats->syncp);
 }
 
-void vxlan_vnifilter_count(struct vxlan_dev *vxlan, __be32 vni,
+void vxlan_vnifilter_count(struct vxlan_dev *vxlan,
+			   const struct vxlan_config *cfg, __be32 vni,
 			   struct vxlan_vni_node *vninode,
 			   int type, unsigned int len)
 {
 	struct vxlan_vni_node *vnode;
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (!(cfg->flags & VXLAN_F_VNIFILTER))
 		return;
 
 	if (vninode) {
@@ -333,22 +334,36 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 				    struct sk_buff *skb,
 				    struct netlink_callback *cb)
 {
-	struct vxlan_vni_node *tmp, *v, *vbegin = NULL, *vend = NULL;
+	struct vxlan_vni_node *v, *vbegin = NULL, *vend = NULL;
 	struct vxlan_dev *vxlan = netdev_priv(dev);
 	struct tunnel_msg *new_tmsg, *tmsg;
-	int idx = 0, s_idx = cb->args[1];
+	const struct vxlan_config *cfg;
 	struct vxlan_vni_group *vg;
 	struct nlmsghdr *nlh;
+	int idx = 0, s_idx;
 	bool dump_stats;
 	int err = 0;
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (cb->args[2] != dev->ifindex) {
+		cb->args[1] = 0;
+		cb->args[2] = dev->ifindex;
+	}
+	s_idx = cb->args[1];
+
+	cfg = rcu_dereference(vxlan->cfg);
+	if (!(cfg->flags & VXLAN_F_VNIFILTER)) {
+		cb->args[1] = 0;
+		cb->args[2] = 0;
 		return -EINVAL;
+	}
 
 	/* RCU needed because of the vni locking rules (rcu || rtnl) */
 	vg = rcu_dereference(vxlan->vnigrp);
-	if (!vg || !vg->num_vnis)
+	if (!vg || !vg->num_vnis) {
+		cb->args[1] = 0;
+		cb->args[2] = 0;
 		return 0;
+	}
 
 	tmsg = nlmsg_data(cb->nlh);
 	dump_stats = !!(tmsg->flags & TUNNEL_MSG_FLAG_STATS);
@@ -362,7 +377,7 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 	new_tmsg->family = PF_BRIDGE;
 	new_tmsg->ifindex = dev->ifindex;
 
-	list_for_each_entry_safe(v, tmp, &vg->vni_list, vlist) {
+	list_for_each_entry_rcu(v, &vg->vni_list, vlist) {
 		if (idx < s_idx) {
 			idx++;
 			continue;
@@ -394,12 +409,26 @@ static int vxlan_vnifilter_dump_dev(const struct net_device *dev,
 	}
 
 	cb->args[1] = err ? idx : 0;
+	cb->args[2] = err ? dev->ifindex : 0;
 
 	nlmsg_end(skb, nlh);
 
+	nl_dump_check_consistent(cb, nlh);
+
 	return err;
 }
 
+static u32 vxlan_vnifilter_base_seq(const struct net *net)
+{
+	const struct vxlan_net *vn = net_generic(net, vxlan_net_id);
+	u32 res = atomic_read(&vn->vnifilter_seq);
+
+	/* Must not return 0 (see nl_dump_check_consistent()) */
+	if (!res)
+		res = 0x80000000;
+	return res;
+}
+
 static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	int idx = 0, err = 0, s_idx = cb->args[0];
@@ -419,6 +448,9 @@ static int vxlan_vnifilter_dump(struct sk_buff *skb, struct netlink_callback *cb
 	}
 
 	rcu_read_lock();
+
+	cb->seq = vxlan_vnifilter_base_seq(net);
+
 	if (tmsg->ifindex) {
 		dev = dev_get_by_index_rcu(net, tmsg->ifindex);
 		if (!dev) {
@@ -470,26 +502,50 @@ static const struct nla_policy vni_filter_policy[VXLAN_VNIFILTER_MAX + 1] = {
 	[VXLAN_VNIFILTER_ENTRY] = { .type = NLA_NESTED },
 };
 
-static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
-					  union vxlan_addr *old_remote_ip,
-					  union vxlan_addr *remote_ip,
-					  struct netlink_ext_ack *extack)
+bool vxlan_vnifilter_has_multicast(const struct vxlan_dev *vxlan)
 {
-	struct vxlan_rdst *dst = &vxlan->default_dst;
+	struct vxlan_vni_group *vg = rtnl_dereference(vxlan->vnigrp);
+	struct vxlan_vni_node *v;
+
+	if (!vg)
+		return false;
+
+	list_for_each_entry(v, &vg->vni_list, vlist) {
+		if (vxlan_addr_multicast(&v->remote_ip))
+			return true;
+	}
+
+	return false;
+}
+
+int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
+				   const union vxlan_addr *old_remote_ip,
+				   const union vxlan_addr *remote_ip,
+				   u32 old_ifindex, u32 new_ifindex,
+				   struct netlink_ext_ack *extack)
+{
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	int err = 0;
 
+	if (old_remote_ip && remote_ip &&
+	    vxlan_addr_equal(old_remote_ip, remote_ip) &&
+	    old_ifindex == new_ifindex)
+		return 0;
+
 	spin_lock_bh(&vxlan->hash_lock);
 	if (remote_ip && !vxlan_addr_any(remote_ip)) {
+		union vxlan_addr rip = *remote_ip;
+
 		err = vxlan_fdb_update(vxlan, all_zeros_mac,
-				       remote_ip,
+				       &rip,
 				       NUD_REACHABLE | NUD_PERMANENT,
 				       NLM_F_APPEND | NLM_F_CREATE,
-				       vxlan->cfg.dst_port,
+				       cfg->dst_port,
 				       vni,
 				       vni,
-				       dst->remote_ifindex,
+				       new_ifindex,
 				       NTF_SELF, 0, true, extack);
-		if (err) {
+		if (err && extack) {
 			spin_unlock_bh(&vxlan->hash_lock);
 			return err;
 		}
@@ -498,9 +554,9 @@ static int vxlan_update_default_fdb_entry(struct vxlan_dev *vxlan, __be32 vni,
 	if (old_remote_ip && !vxlan_addr_any(old_remote_ip)) {
 		__vxlan_fdb_delete(vxlan, all_zeros_mac,
 				   *old_remote_ip,
-				   vxlan->cfg.dst_port,
+				   cfg->dst_port,
 				   vni, vni,
-				   dst->remote_ifindex,
+				   old_ifindex,
 				   true);
 	}
 	spin_unlock_bh(&vxlan->hash_lock);
@@ -515,8 +571,8 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 				  struct netlink_ext_ack *extack)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
-	union vxlan_addr *newrip = NULL, *oldrip = NULL;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
+	const union vxlan_addr *newrip = NULL, *oldrip = NULL;
 	union vxlan_addr old_remote_ip;
 	int ret = 0;
 
@@ -528,15 +584,16 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 	if (group && !vxlan_addr_any(group)) {
 		newrip = group;
 	} else {
-		if (!vxlan_addr_any(&dst->remote_ip))
-			newrip = &dst->remote_ip;
+		if (!vxlan_addr_any(&cfg->remote_ip))
+			newrip = &cfg->remote_ip;
 	}
 
-	/* if old rip exists, and no newrip,
-	 * explicitly delete old rip
-	 */
-	if (!newrip && !vxlan_addr_any(&old_remote_ip))
-		oldrip = &old_remote_ip;
+	if (!create) {
+		if (!vxlan_addr_any(&old_remote_ip))
+			oldrip = &old_remote_ip;
+		else if (!vxlan_addr_any(&cfg->remote_ip))
+			oldrip = &cfg->remote_ip;
+	}
 
 	if (!newrip && !oldrip)
 		return 0;
@@ -546,20 +603,27 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 
 	ret = vxlan_update_default_fdb_entry(vxlan, vninode->vni,
 					     oldrip, newrip,
+					     cfg->remote_ifindex,
+					     cfg->remote_ifindex,
 					     extack);
 	if (ret)
 		goto out;
 
-	if (group)
+	if (group) {
 		memcpy(&vninode->remote_ip, group, sizeof(vninode->remote_ip));
+		if (!create)
+			vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
+	}
 
 	if (vxlan->dev->flags & IFF_UP) {
 		if (vxlan_addr_multicast(&old_remote_ip) &&
 		    !vxlan_group_used(vn, vxlan, vninode->vni,
 				      &old_remote_ip,
-				      vxlan->default_dst.remote_ifindex)) {
+				      cfg->remote_ifindex)) {
 			ret = vxlan_igmp_leave(vxlan, &old_remote_ip,
 					       0);
+			if (ret == -EADDRNOTAVAIL)
+				ret = 0;
 			if (ret)
 				goto out;
 		}
@@ -581,10 +645,12 @@ static int vxlan_vni_update_group(struct vxlan_dev *vxlan,
 }
 
 int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
-			       union vxlan_addr *old_remote_ip,
-			       union vxlan_addr *new_remote_ip,
+			       const union vxlan_addr *old_remote_ip,
+			       const union vxlan_addr *new_remote_ip,
+			       u32 old_ifindex, u32 new_ifindex,
 			       struct netlink_ext_ack *extack)
 {
+	const union vxlan_addr *oldrip, *newrip;
 	struct list_head *headp, *hpos;
 	struct vxlan_vni_group *vg;
 	struct vxlan_vni_node *vent;
@@ -595,37 +661,66 @@ int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
 	headp = &vg->vni_list;
 	list_for_each_prev(hpos, headp) {
 		vent = list_entry(hpos, struct vxlan_vni_node, vlist);
+
 		if (vxlan_addr_any(&vent->remote_ip)) {
-			ret = vxlan_update_default_fdb_entry(vxlan, vent->vni,
-							     old_remote_ip,
-							     new_remote_ip,
-							     extack);
-			if (ret)
-				return ret;
+			oldrip = old_remote_ip;
+			newrip = new_remote_ip;
+		} else {
+			/* A vni with its own group keeps it, but its fdb entry
+			 * is still keyed on the device remote_ifindex.
+			 */
+			oldrip = &vent->remote_ip;
+			newrip = &vent->remote_ip;
 		}
+
+		ret = vxlan_update_default_fdb_entry(vxlan, vent->vni,
+						     oldrip, newrip,
+						     old_ifindex, new_ifindex,
+						     extack);
+		if (ret)
+			goto err_unwind;
 	}
 
 	return 0;
+
+err_unwind:
+	list_for_each_continue(hpos, headp) {
+		vent = list_entry(hpos, struct vxlan_vni_node, vlist);
+
+		if (vxlan_addr_any(&vent->remote_ip)) {
+			oldrip = old_remote_ip;
+			newrip = new_remote_ip;
+		} else {
+			oldrip = &vent->remote_ip;
+			newrip = &vent->remote_ip;
+		}
+
+		vxlan_update_default_fdb_entry(vxlan, vent->vni,
+					       newrip, oldrip,
+					       new_ifindex, old_ifindex,
+					       NULL);
+	}
+	return ret;
 }
 
 static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,
 				   struct vxlan_vni_node *vninode)
 {
 	struct vxlan_net *vn = net_generic(vxlan->net, vxlan_net_id);
-	struct vxlan_rdst *dst = &vxlan->default_dst;
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 
 	/* if per vni remote_ip not present, delete the
 	 * default dst remote_ip previously added for this vni
 	 */
 	if (!vxlan_addr_any(&vninode->remote_ip) ||
-	    !vxlan_addr_any(&dst->remote_ip)) {
+	    !vxlan_addr_any(&cfg->remote_ip)) {
 		spin_lock_bh(&vxlan->hash_lock);
 		__vxlan_fdb_delete(vxlan, all_zeros_mac,
 				   (vxlan_addr_any(&vninode->remote_ip) ?
-				   dst->remote_ip : vninode->remote_ip),
-				   vxlan->cfg.dst_port,
+				   cfg->remote_ip : vninode->remote_ip),
+				   cfg->dst_port,
 				   vninode->vni, vninode->vni,
-				   dst->remote_ifindex,
+				   cfg->remote_ifindex,
 				   true);
 		spin_unlock_bh(&vxlan->hash_lock);
 	}
@@ -634,7 +729,7 @@ static void vxlan_vni_delete_group(struct vxlan_dev *vxlan,
 		if (vxlan_addr_multicast(&vninode->remote_ip) &&
 		    !vxlan_group_used(vn, vxlan, vninode->vni,
 				      &vninode->remote_ip,
-				      dst->remote_ifindex)) {
+				      cfg->remote_ifindex)) {
 			vxlan_igmp_leave(vxlan, &vninode->remote_ip, 0);
 		}
 	}
@@ -665,7 +760,8 @@ static int vxlan_vni_update(struct vxlan_dev *vxlan,
 	return 0;
 }
 
-static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,
+static void __vxlan_vni_add_list(struct vxlan_dev *vxlan,
+				 struct vxlan_vni_group *vg,
 				 struct vxlan_vni_node *v)
 {
 	struct list_head *headp, *hpos;
@@ -681,13 +777,16 @@ static void __vxlan_vni_add_list(struct vxlan_vni_group *vg,
 	}
 	list_add_rcu(&v->vlist, hpos);
 	vg->num_vnis++;
+	vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
 }
 
-static void __vxlan_vni_del_list(struct vxlan_vni_group *vg,
+static void __vxlan_vni_del_list(struct vxlan_dev *vxlan,
+				 struct vxlan_vni_group *vg,
 				 struct vxlan_vni_node *v)
 {
 	list_del_rcu(&v->vlist);
 	vg->num_vnis--;
+	vxlan_vnifilter_seq_inc(dev_net(vxlan->dev));
 }
 
 static struct vxlan_vni_node *vxlan_vni_alloc(struct vxlan_dev *vxlan,
@@ -723,6 +822,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 			 u32 vni, union vxlan_addr *group,
 			 struct netlink_ext_ack *extack)
 {
+	const struct vxlan_config *cfg = rtnl_dereference(vxlan->cfg);
 	struct vxlan_vni_node *vninode;
 	__be32 v = cpu_to_be32(vni);
 	bool changed = false;
@@ -731,7 +831,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 	if (vxlan_vnifilter_lookup(vxlan, v))
 		return vxlan_vni_update(vxlan, vg, v, group, &changed, extack);
 
-	err = vxlan_vni_in_use(vxlan->net, vxlan, &vxlan->cfg, v);
+	err = vxlan_vni_in_use(vxlan->net, vxlan, cfg, v);
 	if (err) {
 		NL_SET_ERR_MSG(extack, "VNI in use");
 		return err;
@@ -749,7 +849,7 @@ static int vxlan_vni_add(struct vxlan_dev *vxlan,
 		return err;
 	}
 
-	__vxlan_vni_add_list(vg, vninode);
+	__vxlan_vni_add_list(vxlan, vg, vninode);
 
 	if (vxlan->dev->flags & IFF_UP)
 		vxlan_vs_add_del_vninode(vxlan, vninode, false);
@@ -795,7 +895,7 @@ static int vxlan_vni_del(struct vxlan_dev *vxlan,
 	if (err)
 		goto out;
 
-	__vxlan_vni_del_list(vg, vninode);
+	__vxlan_vni_del_list(vxlan, vg, vninode);
 
 	vxlan_vnifilter_notify(vxlan, vninode, RTM_DELTUNNEL);
 
@@ -844,6 +944,7 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,
 				    int cmd, struct netlink_ext_ack *extack)
 {
 	struct nlattr *vattrs[VXLAN_VNIFILTER_ENTRY_MAX + 1];
+	const struct vxlan_config *cfg;
 	u32 vni_start = 0, vni_end = 0;
 	union vxlan_addr group;
 	int err;
@@ -881,7 +982,8 @@ static int vxlan_process_vni_filter(struct vxlan_dev *vxlan,
 		memset(&group, 0, sizeof(group));
 	}
 
-	if (vxlan_addr_multicast(&group) && !vxlan->default_dst.remote_ifindex) {
+	cfg = rtnl_dereference(vxlan->cfg);
+	if (vxlan_addr_multicast(&group) && !cfg->remote_ifindex) {
 		NL_SET_ERR_MSG(extack,
 			       "Local interface required for multicast remote group");
 
@@ -909,7 +1011,7 @@ void vxlan_vnigroup_uninit(struct vxlan_dev *vxlan)
 #if IS_ENABLED(CONFIG_IPV6)
 		hlist_del_init_rcu(&v->hlist6.hlist);
 #endif
-		__vxlan_vni_del_list(vg, v);
+		__vxlan_vni_del_list(vxlan, vg, v);
 		vxlan_vnifilter_notify(vxlan, v, RTM_DELTUNNEL);
 		call_rcu(&v->rcu, vxlan_vni_node_rcu_free);
 	}
@@ -940,6 +1042,7 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,
 				   struct netlink_ext_ack *extack)
 {
 	struct net *net = sock_net(skb->sk);
+	const struct vxlan_config *cfg;
 	struct tunnel_msg *tmsg;
 	struct vxlan_dev *vxlan;
 	struct net_device *dev;
@@ -964,8 +1067,9 @@ static int vxlan_vnifilter_process(struct sk_buff *skb, struct nlmsghdr *nlh,
 	}
 
 	vxlan = netdev_priv(dev);
+	cfg = rtnl_dereference(vxlan->cfg);
 
-	if (!(vxlan->cfg.flags & VXLAN_F_VNIFILTER))
+	if (!(cfg->flags & VXLAN_F_VNIFILTER))
 		return -EOPNOTSUPP;
 
 	nlmsg_for_each_attr_type(attr, VXLAN_VNIFILTER_ENTRY, nlh,
diff --git a/include/net/vxlan.h b/include/net/vxlan.h
index 7b82075055237..c3c9f2ccc3d1b 100644
--- a/include/net/vxlan.h
+++ b/include/net/vxlan.h
@@ -204,7 +204,6 @@ struct vxlan_rdst {
 	u8			 offloaded:1;
 	__be32			 remote_vni;
 	u32			 remote_ifindex;
-	struct net_device	 *remote_dev;
 	struct list_head	 list;
 	struct rcu_head		 rcu;
 	struct dst_cache	 dst_cache;
@@ -229,6 +228,7 @@ struct vxlan_config {
 	bool				no_share;
 	enum ifla_vxlan_df		df;
 	struct vxlanhdr			reserved_bits;
+	struct rcu_head			rcu;
 };
 
 enum {
@@ -294,14 +294,15 @@ struct vxlan_dev {
 #endif
 	struct net_device *dev;
 	struct net	  *net;		/* netns for packet i/o */
-	struct vxlan_rdst default_dst;	/* default destination */
+	struct net_device *lowerdev;
 
 	struct timer_list age_timer;
 	spinlock_t	  hash_lock;
 	unsigned int	  addrcnt;
 	struct gro_cells  gro_cells;
+	unsigned long	  flags;
 
-	struct vxlan_config	cfg;
+	struct vxlan_config __rcu	*cfg;
 
 	struct vxlan_vni_group  __rcu *vnigrp;
 
@@ -313,6 +314,10 @@ struct vxlan_dev {
 	unsigned int mdb_seq;
 };
 
+enum vxlan_dev_flags {
+	VXLAN_DEV_F_MDB,
+};
+
 #define VXLAN_F_LEARN			0x01
 #define VXLAN_F_PROXY			0x02
 #define VXLAN_F_RSC			0x04
@@ -331,7 +336,6 @@ struct vxlan_dev {
 #define VXLAN_F_IPV6_LINKLOCAL		0x8000
 #define VXLAN_F_TTL_INHERIT		0x10000
 #define VXLAN_F_VNIFILTER               0x20000
-#define VXLAN_F_MDB			0x40000
 #define VXLAN_F_LOCALBYPASS		0x80000
 #define VXLAN_F_MC_ROUTE		0x100000
 

13/3 2026/09/25 02:07 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 49454
							output: 20
							thoughts: 549
						

Thoughts:
**Analyzing the VXLAN Driver Patch**

Okay, so this patch is clearly making some significant changes to the VXLAN driver, specifically touching the core structure and interactions with Mellanox drivers (`mlx5`, `mlxsw`). My initial impression is that they're moving towards a more dynamic configuration model, likely to handle changes without requiring a full device teardown and rebuild.

The first thing I notice is the change of `struct vxlan_config cfg;` to `struct vxlan_config __rcu *cfg;` within `struct vxlan_dev`. This is a big clue – it's going from a statically allocated structure to a pointer, suggesting dynamic allocation and RCU (Read-Copy-Update) synchronization. The removal or restructuring of `default_dst` is also interesting; I'll need to see how they've replaced that functionality, likely involving `vxlan->lowerdev` and accessing fields within the now dynamically allocated `cfg`.

The dynamic allocation with `kmemdup()` in `vxlan_dev_configure` and `vxlan_changelink`, and the associated `kfree_rcu()` and `kfree()` calls, confirm my suspicion about the dynamic allocation and the use of RCU to ensure safe concurrent access. The move of `VXLAN_F_MDB` to `test_bit()` and the pervasive use of RCU dereferencing are consistent with this approach. The updates to the `mlx5` and `mlxsw` access patterns are further evidence of this architectural shift.

Now, this is where I get serious. With dynamic allocation and RCU, there's always a risk of introducing memory corruption issues, particularly if the initialization, copying, or freeing of the configuration data isn't handled carefully. So, I need to methodically check for potential uninitialized memory issues. We can't have those in production.

First, I need to focus on where the `vxlan_config` structure is allocated and initialized. Let's look at `vxlan_config_apply` which calls `vxlan_dev_configure`. Inside `vxlan_dev_configure`, I see `kmemdup()` being used to allocate and copy a `conf` struct. I need to trace back where the `conf` comes from. Looking up where it comes from, it seems it gets passed in from `vxlan_dev_create`. Now I must figure out how `conf` is initialized in the *caller* of `vxlan_dev_create`. I'm going to run grepper to quickly search the codebase for the calls to `vxlan_dev_create` and see how they initialize `conf`. I must find all the code paths.



14/3 2026/09/25 02:08 tool grepper
0m Args:
{
  "Expression": "vxlan_dev_create"
}

Results:
{
  "Output": "drivers/net/vxlan/vxlan_core.c=4078=static int vxlan_dev_configure(struct net *src_net, struct net_device *dev,\n--\ndrivers/net/vxlan/vxlan_core.c-4099-\ndrivers/net/vxlan/vxlan_core.c:4100:static int vxlan_dev_create(struct net *net, struct net_device *dev,\ndrivers/net/vxlan/vxlan_core.c-4101-\t\t\t    struct vxlan_config *conf,\n--\ndrivers/net/vxlan/vxlan_core.c=4545=static int vxlan_newlink(struct net_device *dev,\n--\ndrivers/net/vxlan/vxlan_core.c-4558-\ndrivers/net/vxlan/vxlan_core.c:4559:\treturn vxlan_dev_create(link_net, dev, \u0026conf, extack);\ndrivers/net/vxlan/vxlan_core.c-4560-}\n"
}

15/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 50857
							output: 19
							thoughts: 21
						

Thoughts:
Here's my attempt at summarizing the thought process, tailored for an expert audience:

**Diving into VXLAN Configuration**

Okay, time to get my hands dirty. I need to take a look at the core of VXLAN configuration, specifically the `vxlan_newlink` and `vxlan_nl2conf` functions. My immediate goal is to understand how these pieces interact within the kernel to set up and validate a new VXLAN interface. I'm assuming, of course, that these functions are responsible for the low-level Netlink interaction and the conversion of Netlink attributes to a kernel-internal data structure (likely `struct vxlan_dev` or similar). I'll be looking for potential data races, incorrect attribute handling, and any gotchas related to multicast group joining and ARP handling. I'll need to confirm that error handling is robust, particularly around Netlink failures, memory allocation, and parameter validation. Furthermore, I'll be paying close attention to the sequencing – does the creation and configuration process adhere to the expected order to prevent inconsistent states? Are there any obvious security vulnerabilities lurking in this configuration path, especially related to untrusted input? It's a critical area, so I'll be thorough.



16/3 2026/09/25 02:08 tool grepper
0m Args:
{
  "Expression": "vxlan_newlink"
}

Results:
{
  "Output": "drivers/net/vxlan/vxlan_core.c=4201=static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],\n--\ndrivers/net/vxlan/vxlan_core.c-4544-\ndrivers/net/vxlan/vxlan_core.c:4545:static int vxlan_newlink(struct net_device *dev,\ndrivers/net/vxlan/vxlan_core.c-4546-\t\t\t struct rtnl_newlink_params *params,\n--\ndrivers/net/vxlan/vxlan_core.c=4866=static struct rtnl_link_ops vxlan_link_ops __read_mostly = {\n--\ndrivers/net/vxlan/vxlan_core.c-4872-\t.validate\t= vxlan_validate,\ndrivers/net/vxlan/vxlan_core.c:4873:\t.newlink\t= vxlan_newlink,\ndrivers/net/vxlan/vxlan_core.c-4874-\t.changelink\t= vxlan_changelink,\n"
}

17/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 51433
							output: 42
							thoughts: 36
						

Thoughts:
Here's my analysis of the provided text, rendered in the first person, assuming an expert user:

**Diving into the VXLAN Core**

Alright, let's get down to brass tacks. I need to understand the VXLAN setup process better, so I'm going straight to the source. Time to dig into the heart of the VXLAN implementation in the Linux kernel. I'm targeting `drivers/net/vxlan/vxlan_core.c` and specifically, I want to drill down into how new VXLAN interfaces are created and configured. I'm focusing on `vxlan_newlink` – that's likely where the interface initialization magic happens, the creation of the virtual tunnel itself. Then, immediately after, I want to see how `vxlan_nl2conf` is used. That probably translates the Netlink messages that are being sent into a configuration that's going to drive the behavior of that VXLAN interface. By examining those two functions in sequence, I should be able to reconstruct the interface creation lifecycle. That will provide a solid foundation for understanding the control plane interactions and the overall data path. Let's see what interesting techniques are used there!



18/3 2026/09/25 02:08 tool read-file
0m Args:
{
  "File": "drivers/net/vxlan/vxlan_core.c",
  "FirstLine": 4200,
  "LineCount": 50
}

Results:
{
  "Contents": "4200:\t\n4201:\tstatic int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],\n4202:\t\t\t\t struct net_device *dev, struct vxlan_config *conf,\n4203:\t\t\t\t bool changelink, struct netlink_ext_ack *extack)\n4204:\t{\n4205:\t\tstruct vxlanhdr used_bits = {\n4206:\t\t\t.vx_flags = VXLAN_HF_VNI,\n4207:\t\t\t.vx_vni = VXLAN_VNI_MASK,\n4208:\t\t};\n4209:\t\tstruct vxlan_dev *vxlan = netdev_priv(dev);\n4210:\t\tint err = 0;\n4211:\t\n4212:\t\tmemset(conf, 0, sizeof(*conf));\n4213:\t\n4214:\t\t/* if changelink operation, start with old existing cfg */\n4215:\t\tif (changelink) {\n4216:\t\t\tconst struct vxlan_config *cfg = rtnl_dereference(vxlan-\u003ecfg);\n4217:\t\n4218:\t\t\tif (cfg)\n4219:\t\t\t\tmemcpy(conf, cfg, sizeof(*conf));\n4220:\t\t}\n4221:\t\n4222:\t\tif (data[IFLA_VXLAN_ID]) {\n4223:\t\t\t__be32 vni = cpu_to_be32(nla_get_u32(data[IFLA_VXLAN_ID]));\n4224:\t\n4225:\t\t\tif (changelink \u0026\u0026 (vni != conf-\u003evni)) {\n4226:\t\t\t\tNL_SET_ERR_MSG_ATTR(extack, tb[IFLA_VXLAN_ID], \"Cannot change VNI\");\n4227:\t\t\t\treturn -EOPNOTSUPP;\n4228:\t\t\t}\n4229:\t\t\tconf-\u003evni = vni;\n4230:\t\t}\n4231:\t\n4232:\t\tif (data[IFLA_VXLAN_GROUP]) {\n4233:\t\t\tif (changelink \u0026\u0026 (conf-\u003eremote_ip.sa.sa_family != AF_INET)) {\n4234:\t\t\t\tNL_SET_ERR_MSG_ATTR(extack, tb[IFLA_VXLAN_GROUP], \"New group address family does not match old group\");\n4235:\t\t\t\treturn -EOPNOTSUPP;\n4236:\t\t\t}\n4237:\t\n4238:\t\t\tconf-\u003eremote_ip.sin.sin_addr.s_addr = nla_get_in_addr(data[IFLA_VXLAN_GROUP]);\n4239:\t\t\tconf-\u003eremote_ip.sa.sa_family = AF_INET;\n4240:\t\t} else if (data[IFLA_VXLAN_GROUP6]) {\n4241:\t\t\tif (!IS_ENABLED(CONFIG_IPV6)) {\n4242:\t\t\t\tNL_SET_ERR_MSG_ATTR(extack, tb[IFLA_VXLAN_GROUP6], \"IPv6 support not enabled in the kernel\");\n4243:\t\t\t\treturn -EPFNOSUPPORT;\n4244:\t\t\t}\n4245:\t\n4246:\t\t\tif (changelink \u0026\u0026 (conf-\u003eremote_ip.sa.sa_family != AF_INET6)) {\n4247:\t\t\t\tNL_SET_ERR_MSG_ATTR(extack, tb[IFLA_VXLAN_GROUP6], \"New group address family does not match old group\");\n4248:\t\t\t\treturn -EOPNOTSUPP;\n4249:\t\t\t}\n"
}

19/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 52676
							output: 24
							thoughts: 575
						

Thoughts:
**Analyzing VXLAN Configuration and Memory Management**

Okay, let's break this down. Line 4212 immediately catches my eye: `memset(conf, 0, sizeof(*conf));`.  The `conf` structure is being completely zeroed out *before* any population, which is generally good practice to prevent unpredictable values, especially in networking code.

Then I see it in `vxlan_changelink`:  `conf` is a local variable, and it's being zeroed with `memset`. The code then *conditionally* copies data from `cfg` into `conf` using `memcpy`, and finally, `kmemdup` is used to create a kernel-managed copy.  The important point here is that even before the potentially non-zero contents are copied, everything is zeroed. This means that, after `memset`, even the `rcu` field within `struct vxlan_config` and any padding are initialized to zero.

Now, let's shift focus to `vxlan_fill_info`.  I see the code is fetching `cfg` using `rcu_dereference`, which is expected for concurrent access. However, I’m paying close attention to `struct ifla_vxlan_port_range ports;` since I've seen this pattern before, and I know it has the potential to bite me. Initially, this looks correct, using `htons` to handle endianness.

**The critical part is here:** the structure `struct ifla_vxlan_port_range` is no longer initialized inline, but has the fields assigned after declaration.  Crucially, it is declared but not initialized. The original code provided direct inline initialization and it has been removed, but if `struct ifla_vxlan_port_range` has any padding then we're in trouble. The current version will read uninitialized values and transmit them.

This change is a potential vulnerability because it's possible that the structure has padding, and if we're not careful, uninitialized memory could be exposed when using `nla_put`. It would be a bug, especially when dealing with network configurations that might be used by external systems. I need to go back and check the definition of `struct ifla_vxlan_port_range` *immediately* to confirm if padding exists and ensure that this hasn't been missed. This is something that must be checked for safety.



20/3 2026/09/25 02:08 tool grepper
0m Args:
{
  "Expression": "struct ifla_vxlan_port_range"
}

Results:
{
  "Output": "drivers/net/vxlan/vxlan_core.c=3527=static const struct nla_policy vxlan_policy[IFLA_VXLAN_MAX + 1] = {\n--\ndrivers/net/vxlan/vxlan_core.c-3540-\t[IFLA_VXLAN_LIMIT]\t= { .type = NLA_U32 },\ndrivers/net/vxlan/vxlan_core.c:3541:\t[IFLA_VXLAN_PORT_RANGE] = { .len  = sizeof(struct ifla_vxlan_port_range) },\ndrivers/net/vxlan/vxlan_core.c-3542-\t[IFLA_VXLAN_PROXY]\t= { .type = NLA_U8 },\n--\ndrivers/net/vxlan/vxlan_core.c=3565=static int vxlan_validate(struct nlattr *tb[], struct nlattr *data[],\n--\ndrivers/net/vxlan/vxlan_core.c-3608-\tif (data[IFLA_VXLAN_PORT_RANGE]) {\ndrivers/net/vxlan/vxlan_core.c:3609:\t\tconst struct ifla_vxlan_port_range *p\ndrivers/net/vxlan/vxlan_core.c-3610-\t\t\t= nla_data(data[IFLA_VXLAN_PORT_RANGE]);\n--\ndrivers/net/vxlan/vxlan_core.c=4201=static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],\n--\ndrivers/net/vxlan/vxlan_core.c-4367-\t\tif (!changelink) {\ndrivers/net/vxlan/vxlan_core.c:4368:\t\t\tconst struct ifla_vxlan_port_range *p\ndrivers/net/vxlan/vxlan_core.c-4369-\t\t\t\t= nla_data(data[IFLA_VXLAN_PORT_RANGE]);\n--\ndrivers/net/vxlan/vxlan_core.c=4707=static size_t vxlan_get_size(const struct net_device *dev)\n--\ndrivers/net/vxlan/vxlan_core.c-4734-\t\t/* IFLA_VXLAN_PORT_RANGE */\ndrivers/net/vxlan/vxlan_core.c:4735:\t\tnla_total_size(sizeof(struct ifla_vxlan_port_range)) +\ndrivers/net/vxlan/vxlan_core.c-4736-\t\tnla_total_size(0) + /* IFLA_VXLAN_GBP */\n--\ndrivers/net/vxlan/vxlan_core.c=4745=static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)\n--\ndrivers/net/vxlan/vxlan_core.c-4747-\tconst struct vxlan_dev *vxlan = netdev_priv(dev);\ndrivers/net/vxlan/vxlan_core.c:4748:\tstruct ifla_vxlan_port_range ports;\ndrivers/net/vxlan/vxlan_core.c-4749-\tconst struct vxlan_config *cfg;\n--\ninclude/uapi/linux/if_link.h=1431=enum {\n--\ninclude/uapi/linux/if_link.h-1470-\ninclude/uapi/linux/if_link.h:1471:struct ifla_vxlan_port_range {\ninclude/uapi/linux/if_link.h-1472-\t__be16\tlow;\n--\ntools/include/uapi/linux/if_link.h=1365=enum {\n--\ntools/include/uapi/linux/if_link.h-1402-\ntools/include/uapi/linux/if_link.h:1403:struct ifla_vxlan_port_range {\ntools/include/uapi/linux/if_link.h-1404-\t__be16\tlow;\n"
}

21/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 54797
							output: 41
							thoughts: 35
						

Thoughts:
Here's my attempt at summarizing the thought process, tailored for an expert audience:

**Diving into the VXLAN Port Range Definition**

Okay, time to get specific. I need to understand how the VXLAN port range is defined within the kernel's network interface configuration. Let's start with the authoritative source: the `include/uapi/linux/if_link.h` header file. Specifically, I'm going to look up the `struct ifla_vxlan_port_range` definition. I'm hoping this will lay bare how port ranges are represented at the user-space/kernel API level. Understanding the struct's fields – their types, sizes, and meanings – is crucial for any work involving VXLAN configuration manipulation. I need to know the canonical representation, so I can ensure my tools and analyses are aligned with the kernel's expectations. This is where the core truth resides.



22/3 2026/09/25 02:08 tool read-file
0m Args:
{
  "File": "include/uapi/linux/if_link.h",
  "FirstLine": 1468,
  "LineCount": 10
}

Results:
{
  "Contents": "1468:\t};\n1469:\t#define IFLA_VXLAN_MAX\t(__IFLA_VXLAN_MAX - 1)\n1470:\t\n1471:\tstruct ifla_vxlan_port_range {\n1472:\t\t__be16\tlow;\n1473:\t\t__be16\thigh;\n1474:\t};\n1475:\t\n1476:\tenum ifla_vxlan_df {\n1477:\t\tVXLAN_DF_UNSET = 0,\n"
}

23/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 55218
							output: 25
							thoughts: 804
						

Thoughts:
**Unraveling Potential Initialization Issues in VXLAN Configuration**

Okay, so I'm looking at this `struct ifla_vxlan_port_range`. It's straightforward enough – just two `__be16` fields, `low` and `high`, totaling 4 bytes with no padding. The initialization of those ports with `htons` seems correct, filling all 4 bytes.

But then, I jumped into this `mlxsw_sp_bridge_vlan_aware_vxlan_join` function. I see the code snippet in `drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c`. Initially, it looks like `params` is fully initialized in a single line. Now, it's just declared as `struct mlxsw_sp_nve_params params;`, leaving it uninitialized. The individual fields are then set later.

This immediately flags a potential issue. Before, the struct was *explicitly* zeroed through initialization, preventing garbage data. Now, since it's uninitialized, it could contain whatever was on the stack at that memory location. Then I see the individual fields are later assigned.

I need to understand what `struct mlxsw_sp_nve_params` *actually* contains. What are its members? What could be affected by this change? I have to figure out what `mlxsw_sp_fid_8021q_lookup` (or whatever other function it's passed to) does with the `params` struct. Is it reading uninitialized data? Is this a potential security vulnerability? I need to analyze the `struct mlxsw_sp_nve_params` definition itself!



24/3 2026/09/25 02:08 tool grepper
0m Args:
{
  "Expression": "struct mlxsw_sp_nve_params"
}

Results:
{
  "Output": "drivers/net/ethernet/mellanox/mlxsw/spectrum.h=1352=extern const struct mlxsw_sp_mr_tcam_ops mlxsw_sp2_mr_tcam_ops;\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h-1354-/* spectrum_nve.c */\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h:1355:struct mlxsw_sp_nve_params {\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h-1356-\tenum mlxsw_sp_nve_type type;\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h=1387=int mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_fid *fid,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h:1388:\t\t\t    struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h-1389-\t\t\t    struct netlink_ext_ack *extack);\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c=935=int mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_fid *fid,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c:936:\t\t\t    struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c-937-\t\t\t    struct netlink_ext_ack *extack)\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h=36=struct mlxsw_sp_nve_ops {\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h-38-\tbool (*can_offload)(const struct mlxsw_sp_nve *nve,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h:39:\t\t\t    const struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h-40-\t\t\t    struct netlink_ext_ack *extack);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h-41-\tvoid (*nve_config)(const struct mlxsw_sp_nve *nve,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h:42:\t\t\t   const struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.h-43-\t\t\t   struct mlxsw_sp_nve_config *config);\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c=58=static bool mlxsw_sp_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c:59:\t\t\t\t\t   const struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c-60-\t\t\t\t\t   struct netlink_ext_ack *extack)\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c=127=static bool mlxsw_sp1_nve_vxlan_can_offload(const struct mlxsw_sp_nve *nve,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c:128:\t\t\t\t\t    const struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c-129-\t\t\t\t\t    struct netlink_ext_ack *extack)\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c=155=static void mlxsw_sp_nve_vxlan_config(const struct mlxsw_sp_nve *nve,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c:156:\t\t\t\t      const struct mlxsw_sp_nve_params *params,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve_vxlan.c-157-\t\t\t\t      struct mlxsw_sp_nve_config *config)\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c=2509=mlxsw_sp_bridge_vlan_aware_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2515-\tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c:2516:\tstruct mlxsw_sp_nve_params params;\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2517-\tconst struct vxlan_config *cfg;\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c=2703=mlxsw_sp_bridge_8021d_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2708-\tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c:2709:\tstruct mlxsw_sp_nve_params params;\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2710-\tconst struct vxlan_config *cfg;\n"
}

25/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 57999
							output: 46
							thoughts: 31
						

Thoughts:
Okay, here's my interpretation of that thought process, tailored for an expert audience, and written in the first person:

**Diving into the Spectrum Driver**

Alright, time to dig into the Mellanox driver, specifically the `mlxsw/spectrum.h` header file. My interest is piqued around line 1355. I'm probably looking for something specific, likely related to data structures or function definitions. Could be exploring an issue I'm debugging, reviewing a potential feature implementation, or just refreshing my memory on a particular aspect of the driver's internals. That line number suggests it's likely a mid-level area, not the very top-level API or the deepest hardware access code. I expect to find some key enums or structs that define the interaction between the higher-level networking stack and the underlying Spectrum switch hardware. Perhaps it involves flow control, queueing, or packet classification configuration. I'm mentally preparing to navigate a potentially complex set of nested structures and macro definitions. Let's see what gems are hidden in there!



26/3 2026/09/25 02:08 tool read-file
0m Args:
{
  "File": "drivers/net/ethernet/mellanox/mlxsw/spectrum.h",
  "FirstLine": 1350,
  "LineCount": 25
}

Results:
{
  "Contents": "1350:\t\n1351:\t/* spectrum2_mr_tcam.c */\n1352:\textern const struct mlxsw_sp_mr_tcam_ops mlxsw_sp2_mr_tcam_ops;\n1353:\t\n1354:\t/* spectrum_nve.c */\n1355:\tstruct mlxsw_sp_nve_params {\n1356:\t\tenum mlxsw_sp_nve_type type;\n1357:\t\t__be32 vni;\n1358:\t\tconst struct net_device *dev;\n1359:\t\tu16 ethertype;\n1360:\t};\n1361:\t\n1362:\textern const struct mlxsw_sp_nve_ops *mlxsw_sp1_nve_ops_arr[];\n1363:\textern const struct mlxsw_sp_nve_ops *mlxsw_sp2_nve_ops_arr[];\n1364:\t\n1365:\tint mlxsw_sp_nve_learned_ip_resolve(struct mlxsw_sp *mlxsw_sp, u32 uip,\n1366:\t\t\t\t\t    enum mlxsw_sp_l3proto proto,\n1367:\t\t\t\t\t    union mlxsw_sp_l3addr *addr);\n1368:\tint mlxsw_sp_nve_flood_ip_add(struct mlxsw_sp *mlxsw_sp,\n1369:\t\t\t\t      struct mlxsw_sp_fid *fid,\n1370:\t\t\t\t      enum mlxsw_sp_l3proto proto,\n1371:\t\t\t\t      union mlxsw_sp_l3addr *addr);\n1372:\tvoid mlxsw_sp_nve_flood_ip_del(struct mlxsw_sp *mlxsw_sp,\n1373:\t\t\t\t       struct mlxsw_sp_fid *fid,\n1374:\t\t\t\t       enum mlxsw_sp_l3proto proto,\n"
}

27/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 58817
							output: 49
							thoughts: 321
						

Thoughts:
**Analyzing the `mlxsw_sp_nve_params` Structure for Potential Padding**

Okay, let's dissect this `struct mlxsw_sp_nve_params`. I need to ensure there are no subtle issues lurking in this structure that could bite us later. I see the definition:

```c
struct mlxsw_sp_nve_params {
	enum mlxsw_sp_nve_type type;
	__be32 vni;
	const struct net_device *dev;
	u16 ethertype;
};
```

Right away, I see every member is initialized, like so:

```c
params.type = MLXSW_SP_NVE_TYPE_VXLAN;
params.vni = cfg->vni;
params.dev = vxlan_dev;
params.ethertype = ethertype;
```

My first thought is always, "Where's the padding?". Let's break down the sizes: `enum mlxsw_sp_nve_type type` is 4 bytes. `__be32 vni` is also 4 bytes. `const struct net_device *dev` is 8 bytes. `u16 ethertype` is 2 bytes. That's a total of 18 bytes. This means there's a good chance there's 6 bytes of padding at the end of the struct. I need to be certain.

The important question is: How is `params` used? Is this structure copied to userspace? Is it used in `memcmp` calls? These are the typical situations where padding can become problematic. I need to track down all the usages. Let's see... I'll check `mlxsw_sp_bridge_vlan_aware_vxlan_join` for clues. Let's start with a peek at `drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c` around lines 2513 and 2704. Hopefully, I'll see if `params` is being passed in a way that could reveal information about the internal padding.



28/3 2026/09/25 02:08 tool read-file
0m Args:
{
  "File": "drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c",
  "FirstLine": 2513,
  "LineCount": 40
}

Results:
{
  "Contents": "2513:\t{\n2514:\t\tstruct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_device-\u003edev);\n2515:\t\tstruct vxlan_dev *vxlan = netdev_priv(vxlan_dev);\n2516:\t\tstruct mlxsw_sp_nve_params params;\n2517:\t\tconst struct vxlan_config *cfg;\n2518:\t\tstruct mlxsw_sp_fid *fid;\n2519:\t\tint err;\n2520:\t\n2521:\t\tcfg = rtnl_dereference(vxlan-\u003ecfg);\n2522:\t\tparams.type = MLXSW_SP_NVE_TYPE_VXLAN;\n2523:\t\tparams.vni = cfg-\u003evni;\n2524:\t\tparams.dev = vxlan_dev;\n2525:\t\tparams.ethertype = ethertype;\n2526:\t\n2527:\t\t/* If the VLAN is 0, we need to find the VLAN that is configured as\n2528:\t\t * PVID and egress untagged on the bridge port of the VxLAN device.\n2529:\t\t * It is possible no such VLAN exists\n2530:\t\t */\n2531:\t\tif (!vid) {\n2532:\t\t\terr = mlxsw_sp_vxlan_mapped_vid(vxlan_dev, \u0026vid);\n2533:\t\t\tif (err || !vid)\n2534:\t\t\t\treturn err;\n2535:\t\t}\n2536:\t\n2537:\t\tfid = mlxsw_sp_fid_8021q_get(mlxsw_sp, vid);\n2538:\t\tif (IS_ERR(fid)) {\n2539:\t\t\tNL_SET_ERR_MSG_MOD(extack, \"Failed to create 802.1Q FID\");\n2540:\t\t\treturn PTR_ERR(fid);\n2541:\t\t}\n2542:\t\n2543:\t\tif (mlxsw_sp_fid_vni_is_set(fid)) {\n2544:\t\t\tNL_SET_ERR_MSG_MOD(extack, \"VNI is already set on FID\");\n2545:\t\t\terr = -EINVAL;\n2546:\t\t\tgoto err_vni_exists;\n2547:\t\t}\n2548:\t\n2549:\t\terr = mlxsw_sp_nve_fid_enable(mlxsw_sp, fid, \u0026params, extack);\n2550:\t\tif (err)\n2551:\t\t\tgoto err_nve_fid_enable;\n2552:\t\n"
}

29/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 60338
							output: 25
							thoughts: 38
						

Thoughts:
Okay, here's my interpretation of that thought process, tailored for an expert audience, and written in the first person:

**Analyzing the mlxsw_sp_nve_fid_enable Function Call**

Alright, so I'm seeing `mlxsw_sp_nve_fid_enable` being called, and the input parameter is, predictably, `&params`. Naturally, my immediate instinct is to dissect what's happening within that function call. I need to understand the implications of enabling a Fabric ID (FID) within the NVE context. Given the naming convention, I'm assuming this function is central to establishing or enabling something related to Network Virtualization Edge (NVE) and how FIDs are handled. Let's dig into the implementation of `mlxsw_sp_nve_fid_enable`. My first thoughts immediately go to understanding the context of that parameter. Is it a structure? If so, what members within that structure are critical here? Does the enable operation directly program hardware registers, or is it more of a configuration change that's deferred? What are the implications of the function's return code? I need to analyze how it handles the parameter, as this is the starting point for whatever changes are being made.



30/3 2026/09/25 02:08 tool grepper
0m Args:
{
  "Expression": "mlxsw_sp_nve_fid_enable"
}

Results:
{
  "Output": "drivers/net/ethernet/mellanox/mlxsw/spectrum.h=1385=void mlxsw_sp_nve_ipv6_addr_map_del(struct mlxsw_sp *mlxsw_sp, const char *mac,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h-1386-\t\t\t\t    u16 fid_index);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h:1387:int mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_fid *fid,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum.h-1388-\t\t\t    struct mlxsw_sp_nve_params *params,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c=919=static void mlxsw_sp_nve_ipv6_addr_flush_by_fid(struct mlxsw_sp *mlxsw_sp,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c-934-\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c:935:int mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_fid *fid,\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c-936-\t\t\t    struct mlxsw_sp_nve_params *params,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c=2509=mlxsw_sp_bridge_vlan_aware_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2548-\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c:2549:\terr = mlxsw_sp_nve_fid_enable(mlxsw_sp, fid, \u0026params, extack);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2550-\tif (err)\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c=2703=mlxsw_sp_bridge_8021d_vxlan_join(struct mlxsw_sp_bridge_device *bridge_device,\n--\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2731-\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c:2732:\terr = mlxsw_sp_nve_fid_enable(mlxsw_sp, fid, \u0026params, extack);\ndrivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c-2733-\tif (err)\n"
}

31/3 2026/09/25 02:08 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 61420
							output: 48
							thoughts: 17
						

Thoughts:
Okay, here's my interpretation of "Let's read `mlxsw_sp_nve_fid_enable`," framed as an expert's thought process:

**Diving into `mlxsw_sp_nve_fid_enable`**

Alright, let's get down to brass tacks. The naming convention strongly suggests we're dealing with a function within the MLXSW (Mellanox Switch) driver, specifically related to NVE (Network Virtualization Edge) and FIDs (Flow IDs). My initial hunch is that this function will be responsible for *enabling* some aspect of FID management, likely within the context of NVE functionality. Thinking about it, it's probably related to programming hardware registers or control structures within the switch ASIC to configure how the switch handles and maps flows associated with the NVE functionality.

I'm expecting to see operations related to:

*   **FID Allocation/Deallocation:** It might check if the FIDs are available to be used by the NVE.
*   **Hardware Programming:** It will probably configure specific hardware tables or memory structures within the switch chip that are used for forwarding based on NVE headers and flow identifiers.
*   **Error Handling:** It would be good to understand its error-handling and failure management logic in the event of an allocation problem.

I'll be looking to understand the specific implications of "enable." Does it just turn on the *capability* or does it actually activate NVE based flow using FIDs for a specific NVE interface or tunnel? I’m interested in the scope: is this global, per-port, or per-tunnel? I hope to see good comments or documentation for this, otherwise I will be left guessing.

Time to dig in and see how right (or wrong!) I am.



32/3 2026/09/25 02:09 tool read-file
0m Args:
{
  "File": "drivers/net/ethernet/mellanox/mlxsw/spectrum_nve.c",
  "FirstLine": 935,
  "LineCount": 40
}

Results:
{
  "Contents": " 935:\tint mlxsw_sp_nve_fid_enable(struct mlxsw_sp *mlxsw_sp, struct mlxsw_sp_fid *fid,\n 936:\t\t\t\t    struct mlxsw_sp_nve_params *params,\n 937:\t\t\t\t    struct netlink_ext_ack *extack)\n 938:\t{\n 939:\t\tstruct mlxsw_sp_nve *nve = mlxsw_sp-\u003enve;\n 940:\t\tconst struct mlxsw_sp_nve_ops *ops;\n 941:\t\tstruct mlxsw_sp_nve_config config;\n 942:\t\tint err;\n 943:\t\n 944:\t\tops = nve-\u003enve_ops_arr[params-\u003etype];\n 945:\t\n 946:\t\tif (!ops-\u003ecan_offload(nve, params, extack))\n 947:\t\t\treturn -EINVAL;\n 948:\t\n 949:\t\tmemset(\u0026config, 0, sizeof(config));\n 950:\t\tops-\u003enve_config(nve, params, \u0026config);\n 951:\t\tif (nve-\u003enum_nve_tunnels \u0026\u0026\n 952:\t\t    memcmp(\u0026config, \u0026nve-\u003econfig, sizeof(config))) {\n 953:\t\t\tNL_SET_ERR_MSG_MOD(extack, \"Conflicting NVE tunnels configuration\");\n 954:\t\t\treturn -EINVAL;\n 955:\t\t}\n 956:\t\n 957:\t\terr = mlxsw_sp_nve_tunnel_init(mlxsw_sp, \u0026config);\n 958:\t\tif (err) {\n 959:\t\t\tNL_SET_ERR_MSG_MOD(extack, \"Failed to initialize NVE tunnel\");\n 960:\t\t\treturn err;\n 961:\t\t}\n 962:\t\n 963:\t\terr = mlxsw_sp_fid_vni_set(fid, params-\u003etype, params-\u003evni,\n 964:\t\t\t\t\t   params-\u003edev-\u003eifindex);\n 965:\t\tif (err) {\n 966:\t\t\tNL_SET_ERR_MSG_MOD(extack, \"Failed to set VNI on FID\");\n 967:\t\t\tgoto err_fid_vni_set;\n 968:\t\t}\n 969:\t\n 970:\t\terr = ops-\u003efdb_replay(params-\u003edev, params-\u003evni, extack);\n 971:\t\tif (err)\n 972:\t\t\tgoto err_fdb_replay;\n 973:\t\n 974:\t\treturn 0;\n"
}

33/3 2026/09/25 02:09 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 62559
							output: 17
							thoughts: 1543
						

Thoughts:
**Analyzing Potential Uninitialized Memory in VXLAN Code**

Okay, so I'm digging into this VXLAN code, trying to see if there are any lurking uninitialized memory issues. My first thought was about `params`, but it seems that's only used within internal functions and the four fields initialized are not user-facing. So that's probably safe.

Now, let's look at `mlx5e_tc_tun_get_remote_ifindex`. It initializes `ifindex` to zero and then reads `cfg->remote_ifindex` under an RCU read lock. That seems fine, as there are no possible issues from this function.

Next, I'm checking the `vxlan_fdb_update` and related functions because I'm seeing this `union vxlan_addr rip = cfg->remote_ip;`. My initial question is why make a copy of `cfg->remote_ip` when the parameter type is different? Then, I realized that `cfg->remote_ip` is `const` because `cfg` itself is `const`. Makes sense. But I need to confirm if `rip` could potentially contain uninitialized memory, however, `cfg->remote_ip` is fully initialized, so the copy should be safe.

I'm also checking `vxlan_update_default_fdb_entry` and seeing `union vxlan_addr rip = *remote_ip;`. That's just a copy, same as before, so no uninitialized data should be involved.

Alright, let's move on to `vxlan_fdb_info`. I see that it uses `rcu_dereference_rtnl(vxlan->cfg)` and puts netlink attributes into an `skb`. It seems like the code explicitly copies the necessary data into the `skb` so again, I don't see any issues here either.

Now, let's look at `vxlan_fill_info`. This one is interesting: `if (nla_put_u32(skb, IFLA_VXLAN_ID, be32_to_cpu(cfg->vni)))`. Now I'm thinking, what happens if `cfg` is NULL? Could that ever happen?

I'm going to track how `vxlan->cfg` is handled. The destructor `vxlan_free_dev` sets `vxlan->cfg` to `NULL`, but, it's called using `dev->priv_destructor = vxlan_free_dev;` in `vxlan_setup`. And in `vxlan_dev_configure`, `new_cfg = kmemdup(conf, sizeof(*conf), GFP_KERNEL);` and the copy is applied using `vxlan_config_apply(dev, new_cfg, lowerdev, src_net, false, true);`. The `vxlan_dev_create` function also calls `vxlan_free_dev` if an error occurs. That implies that `vxlan->cfg` should never be NULL while the device is registered, preventing a NULL pointer dereference. KASAN should be able to catch the NULL pointer dereference.

However, back in `vxlan_fill_info`, I see `nla_put(skb, IFLA_VXLAN_RESERVED_BITS, sizeof(cfg->reserved_bits), &cfg->reserved_bits)`. Here comes the most important question: Where is `cfg->reserved_bits` initialized? I need to check how `reserved_bits` is handled in `vxlan_nl2conf` and `vxlan_config`.



34/3 2026/09/25 02:09 tool grepper
0m Args:
{
  "Expression": "reserved_bits"
}

Results:
{
  "Output": "arch/x86/include/asm/kvm_host.h=570=struct kvm_pmu {\n--\narch/x86/include/asm/kvm_host.h-582-\tu64 global_status_rsvd;\narch/x86/include/asm/kvm_host.h:583:\tu64 reserved_bits;\narch/x86/include/asm/kvm_host.h-584-\tu64 raw_event_mask;\n--\narch/x86/kvm/cpuid.c=376=void kvm_vcpu_after_set_cpuid(struct kvm_vcpu *vcpu)\n--\narch/x86/kvm/cpuid.c-453-#define __kvm_cpu_cap_has(UNUSED_, f) kvm_cpu_cap_has(f)\narch/x86/kvm/cpuid.c:454:\tvcpu-\u003earch.cr4_guest_rsvd_bits = __cr4_reserved_bits(__kvm_cpu_cap_has, UNUSED_) |\narch/x86/kvm/cpuid.c:455:\t\t\t\t\t __cr4_reserved_bits(guest_cpu_cap_has, vcpu);\narch/x86/kvm/cpuid.c-456-#undef __kvm_cpu_cap_has\n--\narch/x86/kvm/lapic.c=2849=int kvm_apic_set_base(struct kvm_vcpu *vcpu, u64 value, bool host_initiated)\n--\narch/x86/kvm/lapic.c-2856-\narch/x86/kvm/lapic.c:2857:\tu64 reserved_bits = kvm_vcpu_reserved_gpa_bits_raw(vcpu) | 0x2ff |\narch/x86/kvm/lapic.c-2858-\t\t(guest_cpu_cap_has(vcpu, X86_FEATURE_X2APIC) ? 0 : X2APIC_ENABLE);\narch/x86/kvm/lapic.c-2859-\narch/x86/kvm/lapic.c:2860:\tif ((value \u0026 reserved_bits) != 0 || new_mode == LAPIC_MODE_INVALID)\narch/x86/kvm/lapic.c-2861-\t\treturn 1;\n--\narch/x86/kvm/msrs.c=1442=int kvm_set_msr_common(struct kvm_vcpu *vcpu, struct msr_data *msr_info)\n--\narch/x86/kvm/msrs.c-1497-\tcase MSR_IA32_PRED_CMD: {\narch/x86/kvm/msrs.c:1498:\t\tu64 reserved_bits = ~(PRED_CMD_IBPB | PRED_CMD_SBPB);\narch/x86/kvm/msrs.c-1499-\n--\narch/x86/kvm/msrs.c-1505-\t\t\t    !guest_cpu_cap_has(vcpu, X86_FEATURE_AMD_IBPB))\narch/x86/kvm/msrs.c:1506:\t\t\t\treserved_bits |= PRED_CMD_IBPB;\narch/x86/kvm/msrs.c-1507-\narch/x86/kvm/msrs.c-1508-\t\t\tif (!guest_cpu_cap_has(vcpu, X86_FEATURE_SBPB))\narch/x86/kvm/msrs.c:1509:\t\t\t\treserved_bits |= PRED_CMD_SBPB;\narch/x86/kvm/msrs.c-1510-\t\t}\n--\narch/x86/kvm/msrs.c-1512-\t\tif (!boot_cpu_has(X86_FEATURE_IBPB))\narch/x86/kvm/msrs.c:1513:\t\t\treserved_bits |= PRED_CMD_IBPB;\narch/x86/kvm/msrs.c-1514-\narch/x86/kvm/msrs.c-1515-\t\tif (!boot_cpu_has(X86_FEATURE_SBPB))\narch/x86/kvm/msrs.c:1516:\t\t\treserved_bits |= PRED_CMD_SBPB;\narch/x86/kvm/msrs.c-1517-\narch/x86/kvm/msrs.c:1518:\t\tif (data \u0026 reserved_bits)\narch/x86/kvm/msrs.c-1519-\t\t\treturn 1;\n--\narch/x86/kvm/pmu.c=974=void kvm_pmu_refresh(struct kvm_vcpu *vcpu)\n--\narch/x86/kvm/pmu.c-991-\tpmu-\u003ecounter_bitmask[KVM_PMC_FIXED] = 0;\narch/x86/kvm/pmu.c:992:\tpmu-\u003ereserved_bits = 0xffffffff00200000ull;\narch/x86/kvm/pmu.c-993-\tpmu-\u003eraw_event_mask = X86_RAW_EVENT_MASK;\n--\narch/x86/kvm/regs.h=379=static inline bool __kvm_is_valid_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)\n--\narch/x86/kvm/regs.h-383-\narch/x86/kvm/regs.h:384:#define __cr4_reserved_bits(__cpu_has, __c)             \\\narch/x86/kvm/regs.h-385-({                                                      \\\narch/x86/kvm/regs.h:386:\tu64 __reserved_bits = CR4_RESERVED_BITS;        \\\narch/x86/kvm/regs.h-387-                                                        \\\narch/x86/kvm/regs.h-388-\tif (!__cpu_has(__c, X86_FEATURE_XSAVE))         \\\narch/x86/kvm/regs.h:389:\t\t__reserved_bits |= X86_CR4_OSXSAVE;     \\\narch/x86/kvm/regs.h-390-\tif (!__cpu_has(__c, X86_FEATURE_SMEP))          \\\narch/x86/kvm/regs.h:391:\t\t__reserved_bits |= X86_CR4_SMEP;        \\\narch/x86/kvm/regs.h-392-\tif (!__cpu_has(__c, X86_FEATURE_SMAP))          \\\narch/x86/kvm/regs.h:393:\t\t__reserved_bits |= X86_CR4_SMAP;        \\\narch/x86/kvm/regs.h-394-\tif (!__cpu_has(__c, X86_FEATURE_FSGSBASE))      \\\narch/x86/kvm/regs.h:395:\t\t__reserved_bits |= X86_CR4_FSGSBASE;    \\\narch/x86/kvm/regs.h-396-\tif (!__cpu_has(__c, X86_FEATURE_PKU))           \\\narch/x86/kvm/regs.h:397:\t\t__reserved_bits |= X86_CR4_PKE;         \\\narch/x86/kvm/regs.h-398-\tif (!__cpu_has(__c, X86_FEATURE_LA57))          \\\narch/x86/kvm/regs.h:399:\t\t__reserved_bits |= X86_CR4_LA57;        \\\narch/x86/kvm/regs.h-400-\tif (!__cpu_has(__c, X86_FEATURE_UMIP))          \\\narch/x86/kvm/regs.h:401:\t\t__reserved_bits |= X86_CR4_UMIP;        \\\narch/x86/kvm/regs.h-402-\tif (!__cpu_has(__c, X86_FEATURE_VMX))           \\\narch/x86/kvm/regs.h:403:\t\t__reserved_bits |= X86_CR4_VMXE;        \\\narch/x86/kvm/regs.h-404-\tif (!__cpu_has(__c, X86_FEATURE_PCID))          \\\narch/x86/kvm/regs.h:405:\t\t__reserved_bits |= X86_CR4_PCIDE;       \\\narch/x86/kvm/regs.h-406-\tif (!__cpu_has(__c, X86_FEATURE_LAM))           \\\narch/x86/kvm/regs.h:407:\t\t__reserved_bits |= X86_CR4_LAM_SUP;     \\\narch/x86/kvm/regs.h-408-\tif (!__cpu_has(__c, X86_FEATURE_SHSTK) \u0026\u0026       \\\narch/x86/kvm/regs.h-409-\t    !__cpu_has(__c, X86_FEATURE_IBT))           \\\narch/x86/kvm/regs.h:410:\t\t__reserved_bits |= X86_CR4_CET;         \\\narch/x86/kvm/regs.h:411:\t__reserved_bits;                                \\\narch/x86/kvm/regs.h-412-})\n--\narch/x86/kvm/svm/pmu.c=150=static int amd_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)\n--\narch/x86/kvm/svm/pmu.c-165-\tif (pmc) {\narch/x86/kvm/svm/pmu.c:166:\t\tdata \u0026= ~pmu-\u003ereserved_bits;\narch/x86/kvm/svm/pmu.c-167-\t\tif (data != pmc-\u003eeventsel) {\n--\narch/x86/kvm/svm/pmu.c=185=static void amd_pmu_refresh(struct kvm_vcpu *vcpu)\n--\narch/x86/kvm/svm/pmu.c-216-\narch/x86/kvm/svm/pmu.c:217:\tpmu-\u003ereserved_bits = 0xfffffff000280000ull;\narch/x86/kvm/svm/pmu.c-218-\tif (guest_cpu_cap_has(vcpu, X86_FEATURE_SVM) \u0026\u0026 kvm_vcpu_has_mediated_pmu(vcpu))\narch/x86/kvm/svm/pmu.c:219:\t\tpmu-\u003ereserved_bits \u0026= ~AMD64_EVENTSEL_HOST_GUEST_MASK;\narch/x86/kvm/svm/pmu.c-220-\n--\narch/x86/kvm/vmx/nested.c=1354=static int vmx_restore_vmx_basic(struct vcpu_vmx *vmx, u64 data)\n--\narch/x86/kvm/vmx/nested.c-1360-\narch/x86/kvm/vmx/nested.c:1361:\tconst u64 reserved_bits = GENMASK_ULL(63, 57) |\narch/x86/kvm/vmx/nested.c-1362-\t\t\t\t  GENMASK_ULL(47, 45) |\n--\narch/x86/kvm/vmx/nested.c-1366-\narch/x86/kvm/vmx/nested.c:1367:\tBUILD_BUG_ON(feature_bits \u0026 reserved_bits);\narch/x86/kvm/vmx/nested.c-1368-\n--\narch/x86/kvm/vmx/nested.c-1374-\t */\narch/x86/kvm/vmx/nested.c:1375:\tif (!is_bitwise_subset(vmx_basic, data, feature_bits | reserved_bits))\narch/x86/kvm/vmx/nested.c-1376-\t\treturn -EINVAL;\n--\narch/x86/kvm/vmx/nested.c=1449=static int vmx_restore_vmx_misc(struct vcpu_vmx *vmx, u64 data)\n--\narch/x86/kvm/vmx/nested.c-1460-\narch/x86/kvm/vmx/nested.c:1461:\tconst u64 reserved_bits = BIT_ULL(31) | GENMASK_ULL(13, 9);\narch/x86/kvm/vmx/nested.c-1462-\n--\narch/x86/kvm/vmx/nested.c-1465-\narch/x86/kvm/vmx/nested.c:1466:\tBUILD_BUG_ON(feature_bits \u0026 reserved_bits);\narch/x86/kvm/vmx/nested.c-1467-\n--\narch/x86/kvm/vmx/nested.c-1472-\t */\narch/x86/kvm/vmx/nested.c:1473:\tif (!is_bitwise_subset(vmx_misc, data, feature_bits | reserved_bits))\narch/x86/kvm/vmx/nested.c-1474-\t\treturn -EINVAL;\n--\narch/x86/kvm/vmx/pmu_intel.c=382=static int intel_pmu_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)\n--\narch/x86/kvm/vmx/pmu_intel.c-387-\tu64 data = msr_info-\u003edata;\narch/x86/kvm/vmx/pmu_intel.c:388:\tu64 reserved_bits, diff;\narch/x86/kvm/vmx/pmu_intel.c-389-\n--\narch/x86/kvm/vmx/pmu_intel.c-435-\t\t} else if ((pmc = get_gp_pmc(pmu, msr, MSR_P6_EVNTSEL0))) {\narch/x86/kvm/vmx/pmu_intel.c:436:\t\t\treserved_bits = pmu-\u003ereserved_bits;\narch/x86/kvm/vmx/pmu_intel.c-437-\t\t\tif ((pmc-\u003eidx == 2) \u0026\u0026\narch/x86/kvm/vmx/pmu_intel.c-438-\t\t\t    (pmu-\u003eraw_event_mask \u0026 HSW_IN_TX_CHECKPOINTED))\narch/x86/kvm/vmx/pmu_intel.c:439:\t\t\t\treserved_bits ^= HSW_IN_TX_CHECKPOINTED;\narch/x86/kvm/vmx/pmu_intel.c:440:\t\t\tif (data \u0026 reserved_bits)\narch/x86/kvm/vmx/pmu_intel.c-441-\t\t\t\treturn 1;\n--\narch/x86/kvm/vmx/pmu_intel.c=501=static void intel_pmu_refresh(struct kvm_vcpu *vcpu)\n--\narch/x86/kvm/vmx/pmu_intel.c-547-\t    (entry-\u003eebx \u0026 (X86_FEATURE_HLE|X86_FEATURE_RTM))) {\narch/x86/kvm/vmx/pmu_intel.c:548:\t\tpmu-\u003ereserved_bits ^= HSW_IN_TX;\narch/x86/kvm/vmx/pmu_intel.c-549-\t\tpmu-\u003eraw_event_mask |= (HSW_IN_TX|HSW_IN_TX_CHECKPOINTED);\n--\narch/x86/kvm/vmx/pmu_intel.c-593-\t\t\tpmu-\u003epebs_enable_rsvd = counter_rsvd;\narch/x86/kvm/vmx/pmu_intel.c:594:\t\t\tpmu-\u003ereserved_bits \u0026= ~ICL_EVENTSEL_ADAPTIVE;\narch/x86/kvm/vmx/pmu_intel.c-595-\t\t\tpmu-\u003epebs_data_cfg_rsvd = ~0xff00000full;\n--\narch/x86/kvm/x86.c=6974=static int kvm_x86_check_processor_compatibility(void)\n--\narch/x86/kvm/x86.c-6986-\narch/x86/kvm/x86.c:6987:\tif (__cr4_reserved_bits(cpu_has, c) !=\narch/x86/kvm/x86.c:6988:\t    __cr4_reserved_bits(cpu_has, \u0026boot_cpu_data))\narch/x86/kvm/x86.c-6989-\t\treturn -EIO;\n--\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h=1269=union dmub_shared_state_ips_fw_signals {\n--\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h-1275-\t\tuint32_t ips1z8_commit: 1; /**\u003c 1 if in IPS1 Z8 Retention */\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h:1276:\t\tuint32_t reserved_bits : 27; /**\u003c Reversed */\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h-1277-\t} bits;\n--\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h=1284=union dmub_shared_state_ips_driver_signals {\n--\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h-1295-\t\tuint32_t allow_dynamic_ips1_z8: 1; /**\u003c 1 if IPS1 z8 ret is allowed in dynamic use cases such as VPB */\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h:1296:\t\tuint32_t reserved_bits : 22; /**\u003c Reversed bits */\ndrivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h-1297-\t} bits;\n--\ndrivers/gpu/drm/arm/display/include/malidp_product.h=25=union komeda_config_id {\n--\ndrivers/gpu/drm/arm/display/include/malidp_product.h-31-\t\t\tn_richs:3, /* number of rich layers per pipeline */\ndrivers/gpu/drm/arm/display/include/malidp_product.h:32:\t\t\treserved_bits:6;\ndrivers/gpu/drm/arm/display/include/malidp_product.h-33-\t};\n--\ndrivers/infiniband/ulp/rtrs/rtrs-pri.h=186=struct rtrs_msg_conn_req {\n--\ndrivers/infiniband/ulp/rtrs/rtrs-pri.h-205-\t/* private: */\ndrivers/infiniband/ulp/rtrs/rtrs-pri.h:206:\tu8\t\treserved_bits : 7;\ndrivers/infiniband/ulp/rtrs/rtrs-pri.h-207-\tu8\t\treserved[11];\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-h264.c=18=struct rkvdec_sps {\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-h264.c-38-\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-h264.c:39:\tu16 reserved_bits:\t\t\t\t12;\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-h264.c-40-\tu16 reserved[11];\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c=54=struct rkvdec_hevc_pps {\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c-89-\tu16 sps_pps_mode\t\t\t\t: 4;\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c:90:\tu16 reserved_bits\t\t\t\t: 14;\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c-91-\tu16 reserved;\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c=124=static void assemble_hw_pps(struct rkvdec_ctx *ctx,\n--\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c-257-\thw_ps-\u003epps.sps_pps_mode = 0;\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c:258:\thw_ps-\u003epps.reserved_bits = 0x3fff;\ndrivers/media/platform/rockchip/rkvdec/rkvdec-vdpu381-hevc.c-259-\thw_ps-\u003epps.reserved = 0xffff;\n--\ndrivers/net/vxlan/vxlan_core.c=1675=static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)\n--\ndrivers/net/vxlan/vxlan_core.c-1721-\ndrivers/net/vxlan/vxlan_core.c:1722:\tif (vh-\u003evx_flags \u0026 cfg-\u003ereserved_bits.vx_flags ||\ndrivers/net/vxlan/vxlan_core.c:1723:\t    vh-\u003evx_vni \u0026 cfg-\u003ereserved_bits.vx_vni) {\ndrivers/net/vxlan/vxlan_core.c-1724-\t\t/* If the header uses bits besides those enabled by the\n--\ndrivers/net/vxlan/vxlan_core.c=4201=static int vxlan_nl2conf(struct nlattr *tb[], struct nlattr *data[],\n--\ndrivers/net/vxlan/vxlan_core.c-4461-\tif (data[IFLA_VXLAN_RESERVED_BITS]) {\ndrivers/net/vxlan/vxlan_core.c:4462:\t\tstruct vxlanhdr reserved_bits;\ndrivers/net/vxlan/vxlan_core.c-4463-\n--\ndrivers/net/vxlan/vxlan_core.c-4466-\t\t\t\t\t    data[IFLA_VXLAN_RESERVED_BITS],\ndrivers/net/vxlan/vxlan_core.c:4467:\t\t\t\t\t    \"Cannot change reserved_bits\");\ndrivers/net/vxlan/vxlan_core.c-4468-\t\t\treturn -EOPNOTSUPP;\n--\ndrivers/net/vxlan/vxlan_core.c-4470-\ndrivers/net/vxlan/vxlan_core.c:4471:\t\tnla_memcpy(\u0026reserved_bits, data[IFLA_VXLAN_RESERVED_BITS],\ndrivers/net/vxlan/vxlan_core.c:4472:\t\t\t   sizeof(reserved_bits));\ndrivers/net/vxlan/vxlan_core.c:4473:\t\tif (used_bits.vx_flags \u0026 reserved_bits.vx_flags ||\ndrivers/net/vxlan/vxlan_core.c:4474:\t\t    used_bits.vx_vni \u0026 reserved_bits.vx_vni) {\ndrivers/net/vxlan/vxlan_core.c-4475-\t\t\t__be64 ub_be64, rb_be64;\n--\ndrivers/net/vxlan/vxlan_core.c-4477-\t\t\tmemcpy(\u0026ub_be64, \u0026used_bits, sizeof(ub_be64));\ndrivers/net/vxlan/vxlan_core.c:4478:\t\t\tmemcpy(\u0026rb_be64, \u0026reserved_bits, sizeof(rb_be64));\ndrivers/net/vxlan/vxlan_core.c-4479-\n--\ndrivers/net/vxlan/vxlan_core.c-4487-\ndrivers/net/vxlan/vxlan_core.c:4488:\t\tconf-\u003ereserved_bits = reserved_bits;\ndrivers/net/vxlan/vxlan_core.c-4489-\t} else {\n--\ndrivers/net/vxlan/vxlan_core.c-4492-\t\t */\ndrivers/net/vxlan/vxlan_core.c:4493:\t\tconf-\u003ereserved_bits = (struct vxlanhdr) {\ndrivers/net/vxlan/vxlan_core.c-4494-\t\t\t.vx_flags = ~used_bits.vx_flags,\n--\ndrivers/net/vxlan/vxlan_core.c=4745=static int vxlan_fill_info(struct sk_buff *skb, const struct net_device *dev)\n--\ndrivers/net/vxlan/vxlan_core.c-4846-\tif (nla_put(skb, IFLA_VXLAN_RESERVED_BITS,\ndrivers/net/vxlan/vxlan_core.c:4847:\t\t    sizeof(cfg-\u003ereserved_bits),\ndrivers/net/vxlan/vxlan_core.c:4848:\t\t    \u0026cfg-\u003ereserved_bits))\ndrivers/net/vxlan/vxlan_core.c-4849-\t\tgoto nla_put_failure;\n--\ninclude/net/vxlan.h=212=struct vxlan_config {\n--\ninclude/net/vxlan.h-229-\tenum ifla_vxlan_df\t\tdf;\ninclude/net/vxlan.h:230:\tstruct vxlanhdr\t\t\treserved_bits;\ninclude/net/vxlan.h-231-\tstruct rcu_head\t\t\trcu;\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh=159=ecn_decap_test()\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh-189-\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh:190:reserved_bits_payload_get()\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh-191-{\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh=260=decap_error_test()\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh-266-\tcorrupted_packet_test \"Decap error: Reserved bits in use\" \\\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh:267:\t\t\"reserved_bits_payload_get\"\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan.sh-268-\tcorrupted_packet_test \"Decap error: Too short inner packet\" \\\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh=163=ecn_decap_test()\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh-194-\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh:195:reserved_bits_payload_get()\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh-196-{\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh=269=decap_error_test()\n--\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh-275-\tcorrupted_packet_test \"Decap error: Reserved bits in use\" \\\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh:276:\t\t\"reserved_bits_payload_get\"\ntools/testing/selftests/drivers/net/mlxsw/devlink_trap_tunnel_vxlan_ipv6.sh-277-\tcorrupted_packet_test \"Decap error: Too short inner packet\" \\\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh=153=vxlan_ping_do()\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-160-\tlocal vni=$1; shift\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:161:\tlocal reserved_bits=$1; shift\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-162-\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:163:\tlocal vxlan_header=$(vxlan_header_bytes $vni $reserved_bits)\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-164-\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh=196=vxlan_device_add()\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-203-\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:204:vxlan_all_reserved_bits()\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-205-{\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh=221=vxlan_ping_reserved()\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-222-{\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:223:\tfor bit in $(vxlan_all_reserved_bits); do\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-224-\t\tvxlan_ping_do 1 $rp2 $(mac_get $rp1) \\\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh=287=plain_test_do()\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-288-{\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:289:\tvxlan_device_add reserved_bits 0xf7ffffff000000ff\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:290:\t__default_test_do 0 \"reserved_bits 0xf7ffffff000000ff\"\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-291-}\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh=299=reserved_test()\n--\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-304-\tlocal reserved_bytes=$(neg_bytes $allowed_bytes)\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:305:\tlocal reserved_bits=${reserved_bytes//:/}\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-306-\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:307:\tvxlan_device_add reserved_bits 0x$reserved_bits\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh:308:\t__default_test_do 1 \"reserved_bits 0x$reserved_bits\"\ntools/testing/selftests/net/forwarding/vxlan_reserved.sh-309-}\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/.gitignore=18=invalid_event_code_test\ntools/testing/selftests/powerpc/pmu/event_code_tests/.gitignore:19:reserved_bits_mmcra_sample_elig_mode_test\ntools/testing/selftests/powerpc/pmu/event_code_tests/.gitignore:20:reserved_bits_mmcra_thresh_ctl_test\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/Makefile=2=TEST_GEN_PROGS := group_constraint_pmc56_test group_pmc56_exclude_constraints_test group_constraint_pmc_count_test \\\ntools/testing/selftests/powerpc/pmu/event_code_tests/Makefile:3:\tgroup_constraint_repeat_test group_constraint_radix_scope_qual_test reserved_bits_mmcra_sample_elig_mode_test \\\ntools/testing/selftests/powerpc/pmu/event_code_tests/Makefile:4:\tgroup_constraint_mmcra_sample_test invalid_event_code_test reserved_bits_mmcra_thresh_ctl_test \\\ntools/testing/selftests/powerpc/pmu/event_code_tests/Makefile-5-\tblacklisted_events_test event_alternatives_tests_p9 event_alternatives_tests_p10 generic_events_valid_test \\\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c-20-\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c:21:static int reserved_bits_mmcra_sample_elig_mode(void)\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c-22-{\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c=74=int main(void)\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c-75-{\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c:76:\treturn test_harness(reserved_bits_mmcra_sample_elig_mode,\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c:77:\t\t\t    \"reserved_bits_mmcra_sample_elig_mode\");\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_sample_elig_mode_test.c-78-}\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c-16-\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c:17:static int reserved_bits_mmcra_thresh_ctl(void)\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c-18-{\n--\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c=41=int main(void)\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c-42-{\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c:43:\treturn test_harness(reserved_bits_mmcra_thresh_ctl, \"reserved_bits_mmcra_thresh_ctl\");\ntools/testing/selftests/powerpc/pmu/event_code_tests/reserved_bits_mmcra_thresh_ctl_test.c-44-}\n--\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c=371=FIXTURE_VARIANT_ADD(v_csr_invalid, new_vlenb)\n--\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c-383-/* invalid reserved bits in vcsr */\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c:384:FIXTURE_VARIANT_ADD(v_csr_invalid, vcsr_invalid_reserved_bits)\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c-385-{\n--\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c-396-/* invalid reserved bits in vtype */\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c:397:FIXTURE_VARIANT_ADD(v_csr_invalid, vtype_invalid_reserved_bits)\ntools/testing/selftests/riscv/vector/validate_v_ptrace.c-398-{\n"
}

35/3 2026/09/25 02:09 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 74654
							output: 44
							thoughts: 732
						

Thoughts:
**Analyzing VXLAN Configuration and VNI Filter Dump**

Okay, so I'm looking at this VXLAN code, specifically the changes related to the VNI filter. It seems the configuration structure, `conf`, is being initialized with `memset` to zero. Then, depending on the scenario (changelink or other operations), either a `memcpy` or a direct assignment (using `nla_memcpy` or setting `reserved_bits`) is performed. The `sizeof(struct vxlanhdr)` is explicitly 8 bytes, so we know the size of the header. The code ensures full initialization, which is good practice.

Now, let's dive into the `vxlan_vnifilter.c` changes. I'm seeing a netlink dump function, `vxlan_vnifilter_dump_dev`. The key here is the handling of the netlink callback (`cb`). It appears that there's a check for `VXLAN_F_VNIFILTER` flag in the configuration. Also, the code now includes some logic to track the state and index for the dump operation, and checks for `dev->ifindex`. This makes sure that the dump operation is consistent across multiple calls.

Most importantly, there's a new sequence number (`cb->seq`) calculated using `vxlan_vnifilter_base_seq()`. It checks for consistency during the netlink dump with `nl_dump_check_consistent()`. This is crucial for reliable netlink communication, especially when dealing with data that might be changing concurrently. The dump is producing `tunnel_msg` structures.



36/3 2026/09/25 02:09 tool read-file
0m Args:
{
  "File": "drivers/net/vxlan/vxlan_vnifilter.c",
  "FirstLine": 355,
  "LineCount": 45
}

Results:
{
  "Contents": " 355:\t\t\tcb-\u003eargs[1] = 0;\n 356:\t\t\tcb-\u003eargs[2] = 0;\n 357:\t\t\treturn -EINVAL;\n 358:\t\t}\n 359:\t\n 360:\t\t/* RCU needed because of the vni locking rules (rcu || rtnl) */\n 361:\t\tvg = rcu_dereference(vxlan-\u003evnigrp);\n 362:\t\tif (!vg || !vg-\u003enum_vnis) {\n 363:\t\t\tcb-\u003eargs[1] = 0;\n 364:\t\t\tcb-\u003eargs[2] = 0;\n 365:\t\t\treturn 0;\n 366:\t\t}\n 367:\t\n 368:\t\ttmsg = nlmsg_data(cb-\u003enlh);\n 369:\t\tdump_stats = !!(tmsg-\u003eflags \u0026 TUNNEL_MSG_FLAG_STATS);\n 370:\t\n 371:\t\tnlh = nlmsg_put(skb, NETLINK_CB(cb-\u003eskb).portid, cb-\u003enlh-\u003enlmsg_seq,\n 372:\t\t\t\tRTM_NEWTUNNEL, sizeof(*new_tmsg), NLM_F_MULTI);\n 373:\t\tif (!nlh)\n 374:\t\t\treturn -EMSGSIZE;\n 375:\t\tnew_tmsg = nlmsg_data(nlh);\n 376:\t\tmemset(new_tmsg, 0, sizeof(*new_tmsg));\n 377:\t\tnew_tmsg-\u003efamily = PF_BRIDGE;\n 378:\t\tnew_tmsg-\u003eifindex = dev-\u003eifindex;\n 379:\t\n 380:\t\tlist_for_each_entry_rcu(v, \u0026vg-\u003evni_list, vlist) {\n 381:\t\t\tif (idx \u003c s_idx) {\n 382:\t\t\t\tidx++;\n 383:\t\t\t\tcontinue;\n 384:\t\t\t}\n 385:\t\t\tif (!vbegin) {\n 386:\t\t\t\tvbegin = v;\n 387:\t\t\t\tvend = v;\n 388:\t\t\t\tcontinue;\n 389:\t\t\t}\n 390:\t\t\tif (!dump_stats \u0026\u0026 vnirange(vend, v) == 1 \u0026\u0026\n 391:\t\t\t    vxlan_addr_equal(\u0026v-\u003eremote_ip, \u0026vend-\u003eremote_ip)) {\n 392:\t\t\t\tgoto update_end;\n 393:\t\t\t} else {\n 394:\t\t\t\tif (!vxlan_fill_vni_filter_entry(skb, vbegin, vend,\n 395:\t\t\t\t\t\t\t\t dump_stats)) {\n 396:\t\t\t\t\terr = -EMSGSIZE;\n 397:\t\t\t\t\tbreak;\n 398:\t\t\t\t}\n 399:\t\t\t\tidx += vnirange(vbegin, vend) + 1;\n"
}

37/3 2026/09/25 02:09 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 76504
							output: 44
							thoughts: 160
						

Thoughts:
Okay, let's dive into this.

**Examining the `vxlan_vnilist_update_group` Function**

Right, I need to understand this `vxlan_vnilist_update_group` function within the VXLAN driver. Specifically, it's defined in `drivers/net/vxlan/vxlan_vnifilter.c`. I'm going to quickly reference the function signature to establish context. It looks like this:

```c
int vxlan_vnilist_update_group(struct vxlan_dev *vxlan,
			       const union vxlan_addr *old_remote_ip,
			       const union vxlan_addr *new_remote_ip,
			       u32 old_ifindex, u32 new_ifindex,
			       struct netlink_ext_ack *extack)
```

Looks like it handles updates to the VNILIST group, likely involving changes to remote IP addresses and interface indices. I'll need to pay close attention to how these changes impact VXLAN traffic. Now, let's jump straight to lines 645 through 700 to understand the core logic. That's where the real work will be done. I'll be looking for the crucial bits of this function.



38/3 2026/09/25 02:09 tool read-file
0m Args:
{
  "File": "drivers/net/vxlan/vxlan_vnifilter.c",
  "FirstLine": 645,
  "LineCount": 60
}

Results:
{
  "Contents": " 645:\t}\n 646:\t\n 647:\tint vxlan_vnilist_update_group(struct vxlan_dev *vxlan,\n 648:\t\t\t\t       const union vxlan_addr *old_remote_ip,\n 649:\t\t\t\t       const union vxlan_addr *new_remote_ip,\n 650:\t\t\t\t       u32 old_ifindex, u32 new_ifindex,\n 651:\t\t\t\t       struct netlink_ext_ack *extack)\n 652:\t{\n 653:\t\tconst union vxlan_addr *oldrip, *newrip;\n 654:\t\tstruct list_head *headp, *hpos;\n 655:\t\tstruct vxlan_vni_group *vg;\n 656:\t\tstruct vxlan_vni_node *vent;\n 657:\t\tint ret;\n 658:\t\n 659:\t\tvg = rtnl_dereference(vxlan-\u003evnigrp);\n 660:\t\n 661:\t\theadp = \u0026vg-\u003evni_list;\n 662:\t\tlist_for_each_prev(hpos, headp) {\n 663:\t\t\tvent = list_entry(hpos, struct vxlan_vni_node, vlist);\n 664:\t\n 665:\t\t\tif (vxlan_addr_any(\u0026vent-\u003eremote_ip)) {\n 666:\t\t\t\toldrip = old_remote_ip;\n 667:\t\t\t\tnewrip = new_remote_ip;\n 668:\t\t\t} else {\n 669:\t\t\t\t/* A vni with its own group keeps it, but its fdb entry\n 670:\t\t\t\t * is still keyed on the device remote_ifindex.\n 671:\t\t\t\t */\n 672:\t\t\t\toldrip = \u0026vent-\u003eremote_ip;\n 673:\t\t\t\tnewrip = \u0026vent-\u003eremote_ip;\n 674:\t\t\t}\n 675:\t\n 676:\t\t\tret = vxlan_update_default_fdb_entry(vxlan, vent-\u003evni,\n 677:\t\t\t\t\t\t\t     oldrip, newrip,\n 678:\t\t\t\t\t\t\t     old_ifindex, new_ifindex,\n 679:\t\t\t\t\t\t\t     extack);\n 680:\t\t\tif (ret)\n 681:\t\t\t\tgoto err_unwind;\n 682:\t\t}\n 683:\t\n 684:\t\treturn 0;\n 685:\t\n 686:\terr_unwind:\n 687:\t\tlist_for_each_continue(hpos, headp) {\n 688:\t\t\tvent = list_entry(hpos, struct vxlan_vni_node, vlist);\n 689:\t\n 690:\t\t\tif (vxlan_addr_any(\u0026vent-\u003eremote_ip)) {\n 691:\t\t\t\toldrip = old_remote_ip;\n 692:\t\t\t\tnewrip = new_remote_ip;\n 693:\t\t\t} else {\n 694:\t\t\t\toldrip = \u0026vent-\u003eremote_ip;\n 695:\t\t\t\tnewrip = \u0026vent-\u003eremote_ip;\n 696:\t\t\t}\n 697:\t\n 698:\t\t\tvxlan_update_default_fdb_entry(vxlan, vent-\u003evni,\n 699:\t\t\t\t\t\t       newrip, oldrip,\n 700:\t\t\t\t\t\t       new_ifindex, old_ifindex,\n 701:\t\t\t\t\t\t       NULL);\n 702:\t\t}\n 703:\t\treturn ret;\n 704:\t}\n"
}

39/3 2026/09/25 02:09 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 77906
							output: 42
							thoughts: 1761
						

Thoughts:
**Analyzing Potential Data Races and Uninitialized Memory in VXLAN Code**

Okay, let's break this down. I'm looking at a potential issue related to data races and uninitialized memory within the VXLAN module in the Linux kernel. Specifically, I'm focusing on how the configuration (`vxlan_config`) is handled and passed around, and how various functions might access it.

First, I'm examining `vxlan_update_default_fdb_entry`. This function seems straightforward enough. It updates the default FDB entry, which stores the mapping of a MAC address to a VXLAN tunnel endpoint. The function checks if remote IP addresses and interface indexes have changed. If they have, it either adds or deletes an FDB entry. Notably, it takes pointers to `union vxlan_addr` for IP addresses which can be safely dereferenced.

Next, I'm moving on to `vxlan_changelink`. This function is responsible for handling changes to a VXLAN network device's configuration through netlink. I can see that `vxlan_changelink` calls `vxlan_nl2conf` to parse the netlink attributes and populate a `vxlan_config` structure, and then a new config is allocated using `kmemdup`. My concern at the moment is whether there might be any uninitialized memory in the copied `new_cfg`. I've checked and confirmed that the `conf` struct is correctly initialized with `memset(conf, 0, sizeof(*conf))` in `vxlan_nl2conf`. Consequently, `new_cfg` that is allocated via `kmemdup` should not have any uninitialized memory.

Finally, I'm analyzing `vxlan_xmit` and `vxlan_xmit_one`. These functions are responsible for the core VXLAN packet transmission logic. I've noted that `vxlan_xmit` retrieves the configuration using `rcu_dereference` to the `vxlan->cfg`. This is standard RCU practice for concurrent access. The configuration data is passed in as a parameter in `vxlan_xmit_one`. This appears safe and correct. However, I'm still trying to confirm if there are any potential data races.

Now, I'm particularly interested in this `struct ip_tunnel_key key;` inside `vxlan_xmit_one`. I want to confirm whether this struct is initialized before it's used. Based on the code snippet I've analyzed, `key` looks uninitialized. This could be a potential bug. My analysis will focus on how `key` is populated and used. A search on the original code may be needed to locate the point where it is used, and initialized.



40/3 2026/09/25 02:09 tool read-file
0m Args:
{
  "File": "drivers/net/vxlan/vxlan_core.c",
  "FirstLine": 2420,
  "LineCount": 40
}

Results:
{
  "Contents": "2420:\t\tint addr_family;\n2421:\t\tbool use_cache;\n2422:\t\t__be32 vni = 0;\n2423:\t\t__u8 tos, ttl;\n2424:\t\tint ifindex;\n2425:\t\tint err = 0;\n2426:\t\tbool xnet;\n2427:\t\n2428:\t\tvxlan = netdev_priv(dev);\n2429:\t\txnet = !net_eq(vxlan-\u003enet, dev_net(vxlan-\u003edev));\n2430:\t\tmd = \u0026_md;\n2431:\t\n2432:\t\tno_eth_encap = flags \u0026 VXLAN_F_GPE \u0026\u0026 skb-\u003eprotocol != htons(ETH_P_TEB);\n2433:\t\treason = skb_vlan_inet_prepare(skb, no_eth_encap);\n2434:\t\tif (reason)\n2435:\t\t\tgoto drop;\n2436:\t\n2437:\t\treason = SKB_DROP_REASON_NOT_SPECIFIED;\n2438:\t\told_iph = ip_hdr(skb);\n2439:\t\n2440:\t\tinfo = skb_tunnel_info(skb);\n2441:\t\tuse_cache = ip_tunnel_dst_cache_usable(skb, info);\n2442:\t\n2443:\t\tif (rdst) {\n2444:\t\t\tmemset(\u0026key, 0, sizeof(key));\n2445:\t\t\tpkey = \u0026key;\n2446:\t\n2447:\t\t\tif (vxlan_addr_any(\u0026rdst-\u003eremote_ip)) {\n2448:\t\t\t\tif (did_rsc) {\n2449:\t\t\t\t\t/* short-circuited back to local bridge */\n2450:\t\t\t\t\tvxlan_encap_bypass(skb, vxlan, vxlan, cfg,\n2451:\t\t\t\t\t\t\t   default_vni, true);\n2452:\t\t\t\t\treturn;\n2453:\t\t\t\t}\n2454:\t\t\t\tgoto drop;\n2455:\t\t\t}\n2456:\t\n2457:\t\t\taddr_family = cfg-\u003esaddr.sa.sa_family;\n2458:\t\t\tdst_port = rdst-\u003eremote_port ? rdst-\u003eremote_port : cfg-\u003edst_port;\n2459:\t\t\tvni = (rdst-\u003eremote_vni) ? : default_vni;\n"
}

41/3 2026/09/25 02:09 llm kmsan-evaluator
0m Model:
gemini-3.8-flash

Args:
null

Results:
null

Tokens:
							input: 80920
							output: 402
							thoughts: 3244
						

Thoughts:
**KMSAN Assessment: vxlan RCU Configuration Change**

Okay, let's break this down. Line 2444 shows a `memset` on the `key` variable, making sure it's zeroed. Good. Now, looking at `vxlan_encap_bypass`, the `loopback` address seems a bit concerning, but it's *supposed* to be initialized correctly based on the `sa_family` (AF_INET or AF_INET6) from the `dst_cfg`. Checking the original code confirms this was already present and correct. We're good.

Moving onto `vxlan_mdb_entry_info_fill` and related MDB functions in `vxlan_mdb.c`. Here, we see that the code is replacing fields within the existing configuration (`cfg`) rather than introducing new, potentially uninitialized, variables. The code is carefully comparing existing values and only updating the user-space output if values have changed. In `vxlan_mdb_config_init`, the `cfg` struct is explicitly `memset` to zero, and the code appears to correctly copy values. Similarly, `vxlan_mdb_get_parse` sets the `group` struct to zero, and the code in `vxlan_mdb_entry_get`, `vxlan_mdb_entry_put`, and `vxlan_mdb_fini` appears safe, particularly because of the `alloc_netdev_mqs` which zero-initializes the private data of net devices.

Checking `vxlan_multicast.c`, the code only passes existing, potentially initialized configuration data to functions like `vxlan_igmp_join` and `vxlan_igmp_leave`. There are no memory allocations there or uninitialized values being used. The error handling for `-EADDRNOTAVAIL` is also fine.

In `vxlan_rcv`, the `_md` struct is zeroed using `memset` before any use. This removes any possible risk of reading uninitialized memory.

Now, let's see if this patch introduces any KMSAN issues. The four criteria:

1.  **Kernel structures to user space**: We check `ports` (`struct ifla_vxlan_port_range`), `new_tmsg` (`struct tunnel_msg`), and `cfg->reserved_bits`. `ports` has two `__be16` fields with no padding, both of which are initialized. `new_tmsg` is zeroed via `memset`. `cfg->reserved_bits` has been initialized during device config in `vxlan_nl2conf` (where `conf` was `memset(conf, 0, sizeof(*conf));`).
2.  **Conditional logic**: There's no branching logic based on uninitialized variables.
3.  **Complex data structures**: No, the main changes here involve initializing structures via `kmemdup` from zeroed structures or `memset`.
4.  **Bounds checks**: No risks of accessing uninitialized memory via bounds checks.

Okay, let's summarize the patch's purpose. This commit refactors `vxlan->cfg` to an RCU-managed pointer. This changes how configuration updates happen by making those atomic via RCU. The commit also removes `vxlan->default_dst` and relies on data from `cfg`. The code also updates dereferencing of the configuration via `rtnl_dereference` and `rcu_dereference` to reflect the change.

The potential bugs we could introduce are RCU-related (use-after-free, missing locks), or NULL dereferences. These are exactly the kind of problems KASAN and PROVE_RCU/LOCKDEP are designed to catch. Also, memory leaks, which are handled by KMEMLEAK/KASAN.

There is *no* risk of uninitialized memory reads or information leaks here. All allocated structures are zero-initialized via `kmemdup` or `memset`. The structures sent to user space via netlink are carefully initialized. **Therefore, NeedsKMSAN should be false.** KMSAN is not justified here.



42/3 2026/09/25 02:09 tool set-results
0m Args:
{
  "KMSANReasoning": "The patch refactors vxlan device configuration (`vxlan-\u003ecfg`) from an embedded struct to a dynamically-allocated, RCU-protected pointer (`struct vxlan_config __rcu *cfg`) to allow atomic configuration updates under RCU. It also replaces `vxlan-\u003edefault_dst` with fields directly in `cfg` and adds RCU dereference helpers across vxlan core, mlx5, and mlxsw drivers.\n\nEvaluation of uninitialized memory risks:\n1. All dynamically allocated `struct vxlan_config` instances are created with `kmemdup` from a `struct vxlan_config conf` that is explicitly zero-initialized via `memset(conf, 0, sizeof(*conf))` in `vxlan_nl2conf()`.\n2. In `vxlan_fill_info()`, `struct ifla_vxlan_port_range ports` contains only two `__be16` fields (`low` and `high`) without padding, both of which are fully assigned before `nla_put()`.\n3. In `mlxsw_sp_bridge_vlan_aware_vxlan_join()`, `struct mlxsw_sp_nve_params params` has all its active fields explicitly assigned, and is solely consumed internally within the mlxsw driver without exposure to userspace.\n4. No uninitialized memory is read, copied to userspace, or branched upon.\n\nThe primary bug risks associated with this refactor involve RCU pointer lifetimes, use-after-free of old configuration objects during grace periods, and NULL pointer dereferences. These classes of bugs are thoroughly covered by KASAN, PROVE_RCU, and standard kernel debugging facilities, not KMSAN. Therefore, a dedicated KMSAN fuzzing session is not warranted.",
  "NeedsKMSAN": false
}

Results:
{
  "KMSANReasoning": "The patch refactors vxlan device configuration (`vxlan-\u003ecfg`) from an embedded struct to a dynamically-allocated, RCU-protected pointer (`struct vxlan_config __rcu *cfg`) to allow atomic configuration updates under RCU. It also replaces `vxlan-\u003edefault_dst` with fields directly in `cfg` and adds RCU dereference helpers across vxlan core, mlx5, and mlxsw drivers.\n\nEvaluation of uninitialized memory risks:\n1. All dynamically allocated `struct vxlan_config` instances are created with `kmemdup` from a `struct vxlan_config conf` that is explicitly zero-initialized via `memset(conf, 0, sizeof(*conf))` in `vxlan_nl2conf()`.\n2. In `vxlan_fill_info()`, `struct ifla_vxlan_port_range ports` contains only two `__be16` fields (`low` and `high`) without padding, both of which are fully assigned before `nla_put()`.\n3. In `mlxsw_sp_bridge_vlan_aware_vxlan_join()`, `struct mlxsw_sp_nve_params params` has all its active fields explicitly assigned, and is solely consumed internally within the mlxsw driver without exposure to userspace.\n4. No uninitialized memory is read, copied to userspace, or branched upon.\n\nThe primary bug risks associated with this refactor involve RCU pointer lifetimes, use-after-free of old configuration objects during grace periods, and NULL pointer dereferences. These classes of bugs are thoroughly covered by KASAN, PROVE_RCU, and standard kernel debugging facilities, not KMSAN. Therefore, a dedicated KMSAN fuzzing session is not warranted.",
  "NeedsKMSAN": false
}

LLM Calls Summary:
Total Calls Total Tokens Avg Tokens Total Duration (Seconds) Avg Duration (Seconds)
Tool Calls Summary:
Total Calls Total Duration (Seconds) Avg Duration (Seconds)