amt: add control plane of amt interface
authorTaehee Yoo <ap420073@gmail.com>
Sun, 31 Oct 2021 16:00:02 +0000 (16:00 +0000)
committerDavid S. Miller <davem@davemloft.net>
Mon, 1 Nov 2021 13:36:08 +0000 (13:36 +0000)
It adds definitions and control plane code for AMT.
this is very similar to udp tunneling interfaces such as gtp, vxlan, etc.
In the next patch, data plane code will be added.

Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MAINTAINERS
drivers/net/Kconfig
drivers/net/Makefile
drivers/net/amt.c [new file with mode: 0644]
include/net/amt.h [new file with mode: 0644]
include/uapi/linux/amt.h [new file with mode: 0644]

index 3b85f03..917f360 100644 (file)
@@ -1020,6 +1020,14 @@ S:       Maintained
 F:     Documentation/devicetree/bindings/iio/light/ams,as73211.yaml
 F:     drivers/iio/light/as73211.c
 
+AMT (Automatic Multicast Tunneling)
+M:     Taehee Yoo <ap420073@gmail.com>
+L:     netdev@vger.kernel.org
+S:     Maintained
+T:     git git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git
+T:     git git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git
+F:     drivers/net/amt.c
+
 ANALOG DEVICES INC AD7192 DRIVER
 M:     Alexandru Tachici <alexandru.tachici@analog.com>
 L:     linux-iio@vger.kernel.org
index dd335ae..034dbd4 100644 (file)
@@ -291,6 +291,22 @@ config GTP
          To compile this drivers as a module, choose M here: the module
          will be called gtp.
 
+config AMT
+       tristate "Automatic Multicast Tunneling (AMT)"
+       depends on INET && IP_MULTICAST
+       select NET_UDP_TUNNEL
+       help
+         This allows one to create AMT(Automatic Multicast Tunneling)
+         virtual interfaces that provide multicast tunneling.
+         There are two roles, Gateway, and Relay.
+         Gateway Encapsulates IGMP/MLD traffic from listeners to the Relay.
+         Gateway Decapsulates multicast traffic from the Relay to Listeners.
+         Relay Encapsulates multicast traffic from Sources to Gateway.
+         Relay Decapsulates IGMP/MLD traffic from Gateway.
+
+         To compile this drivers as a module, choose M here: the module
+         will be called amt.
+
 config MACSEC
        tristate "IEEE 802.1AE MAC-level encryption (MACsec)"
        select CRYPTO
index 7398386..50b23e7 100644 (file)
@@ -14,6 +14,7 @@ obj-$(CONFIG_WIREGUARD) += wireguard/
 obj-$(CONFIG_EQUALIZER) += eql.o
 obj-$(CONFIG_IFB) += ifb.o
 obj-$(CONFIG_MACSEC) += macsec.o
+obj-$(CONFIG_AMT) += amt.o
 obj-$(CONFIG_MACVLAN) += macvlan.o
 obj-$(CONFIG_MACVTAP) += macvtap.o
 obj-$(CONFIG_MII) += mii.o
diff --git a/drivers/net/amt.c b/drivers/net/amt.c
new file mode 100644 (file)
index 0000000..addab3b
--- /dev/null
@@ -0,0 +1,493 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Copyright (c) 2021 Taehee Yoo <ap420073@gmail.com> */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/module.h>
+#include <linux/skbuff.h>
+#include <linux/udp.h>
+#include <linux/jhash.h>
+#include <linux/if_tunnel.h>
+#include <linux/net.h>
+#include <linux/igmp.h>
+#include <linux/workqueue.h>
+#include <net/net_namespace.h>
+#include <net/protocol.h>
+#include <net/ip.h>
+#include <net/udp.h>
+#include <net/udp_tunnel.h>
+#include <net/icmp.h>
+#include <net/mld.h>
+#include <net/amt.h>
+#include <uapi/linux/amt.h>
+#include <linux/security.h>
+#include <net/gro_cells.h>
+#include <net/ipv6.h>
+#include <net/protocol.h>
+#include <net/if_inet6.h>
+#include <net/ndisc.h>
+#include <net/addrconf.h>
+#include <net/ip6_route.h>
+#include <net/inet_common.h>
+
+static struct workqueue_struct *amt_wq;
+
+static struct socket *amt_create_sock(struct net *net, __be16 port)
+{
+       struct udp_port_cfg udp_conf;
+       struct socket *sock;
+       int err;
+
+       memset(&udp_conf, 0, sizeof(udp_conf));
+       udp_conf.family = AF_INET;
+       udp_conf.local_ip.s_addr = htonl(INADDR_ANY);
+
+       udp_conf.local_udp_port = port;
+
+       err = udp_sock_create(net, &udp_conf, &sock);
+       if (err < 0)
+               return ERR_PTR(err);
+
+       return sock;
+}
+
+static int amt_socket_create(struct amt_dev *amt)
+{
+       struct udp_tunnel_sock_cfg tunnel_cfg;
+       struct socket *sock;
+
+       sock = amt_create_sock(amt->net, amt->relay_port);
+       if (IS_ERR(sock))
+               return PTR_ERR(sock);
+
+       /* Mark socket as an encapsulation socket */
+       memset(&tunnel_cfg, 0, sizeof(tunnel_cfg));
+       tunnel_cfg.sk_user_data = amt;
+       tunnel_cfg.encap_type = 1;
+       tunnel_cfg.encap_destroy = NULL;
+       setup_udp_tunnel_sock(amt->net, sock, &tunnel_cfg);
+
+       rcu_assign_pointer(amt->sock, sock);
+       return 0;
+}
+
+static int amt_dev_open(struct net_device *dev)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+       int err;
+
+       amt->ready4 = false;
+       amt->ready6 = false;
+
+       err = amt_socket_create(amt);
+       if (err)
+               return err;
+
+       amt->req_cnt = 0;
+       amt->remote_ip = 0;
+       get_random_bytes(&amt->key, sizeof(siphash_key_t));
+
+       amt->status = AMT_STATUS_INIT;
+       return err;
+}
+
+static int amt_dev_stop(struct net_device *dev)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+       struct socket *sock;
+
+       /* shutdown */
+       sock = rtnl_dereference(amt->sock);
+       RCU_INIT_POINTER(amt->sock, NULL);
+       synchronize_net();
+       if (sock)
+               udp_tunnel_sock_release(sock);
+
+       amt->ready4 = false;
+       amt->ready6 = false;
+       amt->req_cnt = 0;
+       amt->remote_ip = 0;
+
+       return 0;
+}
+
+static const struct device_type amt_type = {
+       .name = "amt",
+};
+
+static int amt_dev_init(struct net_device *dev)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+       int err;
+
+       amt->dev = dev;
+       dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
+       if (!dev->tstats)
+               return -ENOMEM;
+
+       err = gro_cells_init(&amt->gro_cells, dev);
+       if (err) {
+               free_percpu(dev->tstats);
+               return err;
+       }
+
+       return 0;
+}
+
+static void amt_dev_uninit(struct net_device *dev)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+
+       gro_cells_destroy(&amt->gro_cells);
+       free_percpu(dev->tstats);
+}
+
+static const struct net_device_ops amt_netdev_ops = {
+       .ndo_init               = amt_dev_init,
+       .ndo_uninit             = amt_dev_uninit,
+       .ndo_open               = amt_dev_open,
+       .ndo_stop               = amt_dev_stop,
+       .ndo_get_stats64        = dev_get_tstats64,
+};
+
+static void amt_link_setup(struct net_device *dev)
+{
+       dev->netdev_ops         = &amt_netdev_ops;
+       dev->needs_free_netdev  = true;
+       SET_NETDEV_DEVTYPE(dev, &amt_type);
+       dev->min_mtu            = ETH_MIN_MTU;
+       dev->max_mtu            = ETH_MAX_MTU;
+       dev->type               = ARPHRD_NONE;
+       dev->flags              = IFF_POINTOPOINT | IFF_NOARP | IFF_MULTICAST;
+       dev->hard_header_len    = 0;
+       dev->addr_len           = 0;
+       dev->priv_flags         |= IFF_NO_QUEUE;
+       dev->features           |= NETIF_F_LLTX;
+       dev->features           |= NETIF_F_GSO_SOFTWARE;
+       dev->features           |= NETIF_F_NETNS_LOCAL;
+       dev->hw_features        |= NETIF_F_SG | NETIF_F_HW_CSUM;
+       dev->hw_features        |= NETIF_F_FRAGLIST | NETIF_F_RXCSUM;
+       dev->hw_features        |= NETIF_F_GSO_SOFTWARE;
+       eth_hw_addr_random(dev);
+       eth_zero_addr(dev->broadcast);
+       ether_setup(dev);
+}
+
+static const struct nla_policy amt_policy[IFLA_AMT_MAX + 1] = {
+       [IFLA_AMT_MODE]         = { .type = NLA_U32 },
+       [IFLA_AMT_RELAY_PORT]   = { .type = NLA_U16 },
+       [IFLA_AMT_GATEWAY_PORT] = { .type = NLA_U16 },
+       [IFLA_AMT_LINK]         = { .type = NLA_U32 },
+       [IFLA_AMT_LOCAL_IP]     = { .len = sizeof_field(struct iphdr, daddr) },
+       [IFLA_AMT_REMOTE_IP]    = { .len = sizeof_field(struct iphdr, daddr) },
+       [IFLA_AMT_DISCOVERY_IP] = { .len = sizeof_field(struct iphdr, daddr) },
+       [IFLA_AMT_MAX_TUNNELS]  = { .type = NLA_U32 },
+};
+
+static int amt_validate(struct nlattr *tb[], struct nlattr *data[],
+                       struct netlink_ext_ack *extack)
+{
+       if (!data)
+               return -EINVAL;
+
+       if (!data[IFLA_AMT_LINK]) {
+               NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_LINK],
+                                   "Link attribute is required");
+               return -EINVAL;
+       }
+
+       if (!data[IFLA_AMT_MODE]) {
+               NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_MODE],
+                                   "Mode attribute is required");
+               return -EINVAL;
+       }
+
+       if (nla_get_u32(data[IFLA_AMT_MODE]) > AMT_MODE_MAX) {
+               NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_MODE],
+                                   "Mode attribute is not valid");
+               return -EINVAL;
+       }
+
+       if (!data[IFLA_AMT_LOCAL_IP]) {
+               NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_DISCOVERY_IP],
+                                   "Local attribute is required");
+               return -EINVAL;
+       }
+
+       if (!data[IFLA_AMT_DISCOVERY_IP] &&
+           nla_get_u32(data[IFLA_AMT_MODE]) == AMT_MODE_GATEWAY) {
+               NL_SET_ERR_MSG_ATTR(extack, data[IFLA_AMT_LOCAL_IP],
+                                   "Discovery attribute is required");
+               return -EINVAL;
+       }
+
+       return 0;
+}
+
+static int amt_newlink(struct net *net, struct net_device *dev,
+                      struct nlattr *tb[], struct nlattr *data[],
+                      struct netlink_ext_ack *extack)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+       int err = -EINVAL;
+
+       amt->net = net;
+       amt->mode = nla_get_u32(data[IFLA_AMT_MODE]);
+
+       if (data[IFLA_AMT_MAX_TUNNELS])
+               amt->max_tunnels = nla_get_u32(data[IFLA_AMT_MAX_TUNNELS]);
+       else
+               amt->max_tunnels = AMT_MAX_TUNNELS;
+
+       spin_lock_init(&amt->lock);
+       amt->max_groups = AMT_MAX_GROUP;
+       amt->max_sources = AMT_MAX_SOURCE;
+       amt->hash_buckets = AMT_HSIZE;
+       amt->nr_tunnels = 0;
+       get_random_bytes(&amt->hash_seed, sizeof(amt->hash_seed));
+       amt->stream_dev = dev_get_by_index(net,
+                                          nla_get_u32(data[IFLA_AMT_LINK]));
+       if (!amt->stream_dev) {
+               NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_LINK],
+                                   "Can't find stream device");
+               return -ENODEV;
+       }
+
+       if (amt->stream_dev->type != ARPHRD_ETHER) {
+               NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_LINK],
+                                   "Invalid stream device type");
+               goto err;
+       }
+
+       amt->local_ip = nla_get_in_addr(data[IFLA_AMT_LOCAL_IP]);
+       if (ipv4_is_loopback(amt->local_ip) ||
+           ipv4_is_zeronet(amt->local_ip) ||
+           ipv4_is_multicast(amt->local_ip)) {
+               NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_LOCAL_IP],
+                                   "Invalid Local address");
+               goto err;
+       }
+
+       if (data[IFLA_AMT_RELAY_PORT])
+               amt->relay_port = nla_get_be16(data[IFLA_AMT_RELAY_PORT]);
+       else
+               amt->relay_port = htons(IANA_AMT_UDP_PORT);
+
+       if (data[IFLA_AMT_GATEWAY_PORT])
+               amt->gw_port = nla_get_be16(data[IFLA_AMT_GATEWAY_PORT]);
+       else
+               amt->gw_port = htons(IANA_AMT_UDP_PORT);
+
+       if (!amt->relay_port) {
+               NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
+                                   "relay port must not be 0");
+               goto err;
+       }
+       if (amt->mode == AMT_MODE_RELAY) {
+               amt->qrv = amt->net->ipv4.sysctl_igmp_qrv;
+               amt->qri = 10;
+               dev->needed_headroom = amt->stream_dev->needed_headroom +
+                                      AMT_RELAY_HLEN;
+               dev->mtu = amt->stream_dev->mtu - AMT_RELAY_HLEN;
+               dev->max_mtu = dev->mtu;
+               dev->min_mtu = ETH_MIN_MTU + AMT_RELAY_HLEN;
+       } else {
+               if (!data[IFLA_AMT_DISCOVERY_IP]) {
+                       NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
+                                           "discovery must be set in gateway mode");
+                       goto err;
+               }
+               if (!amt->gw_port) {
+                       NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
+                                           "gateway port must not be 0");
+                       goto err;
+               }
+               amt->remote_ip = 0;
+               amt->discovery_ip = nla_get_in_addr(data[IFLA_AMT_DISCOVERY_IP]);
+               if (ipv4_is_loopback(amt->discovery_ip) ||
+                   ipv4_is_zeronet(amt->discovery_ip) ||
+                   ipv4_is_multicast(amt->discovery_ip)) {
+                       NL_SET_ERR_MSG_ATTR(extack, tb[IFLA_AMT_DISCOVERY_IP],
+                                           "discovery must be unicast");
+                       goto err;
+               }
+
+               dev->needed_headroom = amt->stream_dev->needed_headroom +
+                                      AMT_GW_HLEN;
+               dev->mtu = amt->stream_dev->mtu - AMT_GW_HLEN;
+               dev->max_mtu = dev->mtu;
+               dev->min_mtu = ETH_MIN_MTU + AMT_GW_HLEN;
+       }
+       amt->qi = AMT_INIT_QUERY_INTERVAL;
+
+       err = register_netdevice(dev);
+       if (err < 0) {
+               netdev_dbg(dev, "failed to register new netdev %d\n", err);
+               goto err;
+       }
+
+       err = netdev_upper_dev_link(amt->stream_dev, dev, extack);
+       if (err < 0) {
+               unregister_netdevice(dev);
+               goto err;
+       }
+
+       return 0;
+err:
+       dev_put(amt->stream_dev);
+       return err;
+}
+
+static void amt_dellink(struct net_device *dev, struct list_head *head)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+
+       unregister_netdevice_queue(dev, head);
+       netdev_upper_dev_unlink(amt->stream_dev, dev);
+       dev_put(amt->stream_dev);
+}
+
+static size_t amt_get_size(const struct net_device *dev)
+{
+       return nla_total_size(sizeof(__u32)) + /* IFLA_AMT_MODE */
+              nla_total_size(sizeof(__u16)) + /* IFLA_AMT_RELAY_PORT */
+              nla_total_size(sizeof(__u16)) + /* IFLA_AMT_GATEWAY_PORT */
+              nla_total_size(sizeof(__u32)) + /* IFLA_AMT_LINK */
+              nla_total_size(sizeof(__u32)) + /* IFLA_MAX_TUNNELS */
+              nla_total_size(sizeof(struct iphdr)) + /* IFLA_AMT_DISCOVERY_IP */
+              nla_total_size(sizeof(struct iphdr)) + /* IFLA_AMT_REMOTE_IP */
+              nla_total_size(sizeof(struct iphdr)); /* IFLA_AMT_LOCAL_IP */
+}
+
+static int amt_fill_info(struct sk_buff *skb, const struct net_device *dev)
+{
+       struct amt_dev *amt = netdev_priv(dev);
+
+       if (nla_put_u32(skb, IFLA_AMT_MODE, amt->mode))
+               goto nla_put_failure;
+       if (nla_put_be16(skb, IFLA_AMT_RELAY_PORT, amt->relay_port))
+               goto nla_put_failure;
+       if (nla_put_be16(skb, IFLA_AMT_GATEWAY_PORT, amt->gw_port))
+               goto nla_put_failure;
+       if (nla_put_u32(skb, IFLA_AMT_LINK, amt->stream_dev->ifindex))
+               goto nla_put_failure;
+       if (nla_put_in_addr(skb, IFLA_AMT_LOCAL_IP, amt->local_ip))
+               goto nla_put_failure;
+       if (nla_put_in_addr(skb, IFLA_AMT_DISCOVERY_IP, amt->discovery_ip))
+               goto nla_put_failure;
+       if (amt->remote_ip)
+               if (nla_put_in_addr(skb, IFLA_AMT_REMOTE_IP, amt->remote_ip))
+                       goto nla_put_failure;
+       if (nla_put_u32(skb, IFLA_AMT_MAX_TUNNELS, amt->max_tunnels))
+               goto nla_put_failure;
+
+       return 0;
+
+nla_put_failure:
+       return -EMSGSIZE;
+}
+
+static struct rtnl_link_ops amt_link_ops __read_mostly = {
+       .kind           = "amt",
+       .maxtype        = IFLA_AMT_MAX,
+       .policy         = amt_policy,
+       .priv_size      = sizeof(struct amt_dev),
+       .setup          = amt_link_setup,
+       .validate       = amt_validate,
+       .newlink        = amt_newlink,
+       .dellink        = amt_dellink,
+       .get_size       = amt_get_size,
+       .fill_info      = amt_fill_info,
+};
+
+static struct net_device *amt_lookup_upper_dev(struct net_device *dev)
+{
+       struct net_device *upper_dev;
+       struct amt_dev *amt;
+
+       for_each_netdev(dev_net(dev), upper_dev) {
+               if (netif_is_amt(upper_dev)) {
+                       amt = netdev_priv(upper_dev);
+                       if (amt->stream_dev == dev)
+                               return upper_dev;
+               }
+       }
+
+       return NULL;
+}
+
+static int amt_device_event(struct notifier_block *unused,
+                           unsigned long event, void *ptr)
+{
+       struct net_device *dev = netdev_notifier_info_to_dev(ptr);
+       struct net_device *upper_dev;
+       struct amt_dev *amt;
+       LIST_HEAD(list);
+       int new_mtu;
+
+       upper_dev = amt_lookup_upper_dev(dev);
+       if (!upper_dev)
+               return NOTIFY_DONE;
+       amt = netdev_priv(upper_dev);
+
+       switch (event) {
+       case NETDEV_UNREGISTER:
+               amt_dellink(amt->dev, &list);
+               unregister_netdevice_many(&list);
+               break;
+       case NETDEV_CHANGEMTU:
+               if (amt->mode == AMT_MODE_RELAY)
+                       new_mtu = dev->mtu - AMT_RELAY_HLEN;
+               else
+                       new_mtu = dev->mtu - AMT_GW_HLEN;
+
+               dev_set_mtu(amt->dev, new_mtu);
+               break;
+       }
+
+       return NOTIFY_DONE;
+}
+
+static struct notifier_block amt_notifier_block __read_mostly = {
+       .notifier_call = amt_device_event,
+};
+
+static int __init amt_init(void)
+{
+       int err;
+
+       err = register_netdevice_notifier(&amt_notifier_block);
+       if (err < 0)
+               goto err;
+
+       err = rtnl_link_register(&amt_link_ops);
+       if (err < 0)
+               goto unregister_notifier;
+
+       amt_wq = alloc_workqueue("amt", WQ_UNBOUND, 1);
+       if (!amt_wq)
+               goto rtnl_unregister;
+
+       return 0;
+
+rtnl_unregister:
+       rtnl_link_unregister(&amt_link_ops);
+unregister_notifier:
+       unregister_netdevice_notifier(&amt_notifier_block);
+err:
+       pr_err("error loading AMT module loaded\n");
+       return err;
+}
+late_initcall(amt_init);
+
+static void __exit amt_fini(void)
+{
+       rtnl_link_unregister(&amt_link_ops);
+       unregister_netdevice_notifier(&amt_notifier_block);
+       destroy_workqueue(amt_wq);
+}
+module_exit(amt_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Taehee Yoo <ap420073@gmail.com>");
+MODULE_ALIAS_RTNL_LINK("amt");
diff --git a/include/net/amt.h b/include/net/amt.h
new file mode 100644 (file)
index 0000000..ce24ff8
--- /dev/null
@@ -0,0 +1,235 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (c) 2021 Taehee Yoo <ap420073@gmail.com>
+ */
+#ifndef _NET_AMT_H_
+#define _NET_AMT_H_
+
+#include <linux/siphash.h>
+#include <linux/jhash.h>
+
+enum amt_msg_type {
+       AMT_MSG_DISCOVERY = 1,
+       AMT_MSG_ADVERTISEMENT,
+       AMT_MSG_REQUEST,
+       AMT_MSG_MEMBERSHIP_QUERY,
+       AMT_MSG_MEMBERSHIP_UPDATE,
+       AMT_MSG_MULTICAST_DATA,
+       AMT_MSG_TEARDOWM,
+       __AMT_MSG_MAX,
+};
+
+#define AMT_MSG_MAX (__AMT_MSG_MAX - 1)
+
+enum amt_status {
+       AMT_STATUS_INIT,
+       AMT_STATUS_SENT_DISCOVERY,
+       AMT_STATUS_RECEIVED_DISCOVERY,
+       AMT_STATUS_SENT_ADVERTISEMENT,
+       AMT_STATUS_RECEIVED_ADVERTISEMENT,
+       AMT_STATUS_SENT_REQUEST,
+       AMT_STATUS_RECEIVED_REQUEST,
+       AMT_STATUS_SENT_QUERY,
+       AMT_STATUS_RECEIVED_QUERY,
+       AMT_STATUS_SENT_UPDATE,
+       AMT_STATUS_RECEIVED_UPDATE,
+       __AMT_STATUS_MAX,
+};
+
+#define AMT_STATUS_MAX (__AMT_STATUS_MAX - 1)
+
+struct amt_header_discovery {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u32     type:4,
+               version:4,
+               reserved:24;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u32     version:4,
+               type:4,
+               reserved:24;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+       __be32  nonce;
+} __packed;
+
+struct amt_header_advertisement {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u32     type:4,
+               version:4,
+               reserved:24;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u32     version:4,
+               type:4,
+               reserved:24;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+       __be32  nonce;
+       __be32  ip4;
+} __packed;
+
+struct amt_header_request {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u32     type:4,
+               version:4,
+               reserved1:7,
+               p:1,
+               reserved2:16;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u32     version:4,
+               type:4,
+               p:1,
+               reserved1:7,
+               reserved2:16;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+       __be32  nonce;
+} __packed;
+
+struct amt_header_membership_query {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u64     type:4,
+               version:4,
+               reserved:6,
+               l:1,
+               g:1,
+               response_mac:48;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u64     version:4,
+               type:4,
+               g:1,
+               l:1,
+               reserved:6,
+               response_mac:48;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+       __be32  nonce;
+} __packed;
+
+struct amt_header_membership_update {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u64     type:4,
+               version:4,
+               reserved:8,
+               response_mac:48;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u64     version:4,
+               type:4,
+               reserved:8,
+               response_mac:48;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+       __be32  nonce;
+} __packed;
+
+struct amt_header_mcast_data {
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+       u16     type:4,
+               version:4,
+               reserved:8;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+       u16     version:4,
+               type:4,
+               reserved:8;
+#else
+#error  "Please fix <asm/byteorder.h>"
+#endif
+} __packed;
+
+struct amt_gw_headers {
+       union {
+               struct amt_header_discovery discovery;
+               struct amt_header_request request;
+               struct amt_header_membership_update update;
+       };
+} __packed;
+
+struct amt_relay_headers {
+       union {
+               struct amt_header_advertisement advertisement;
+               struct amt_header_membership_query query;
+               struct amt_header_mcast_data data;
+       };
+} __packed;
+
+struct amt_dev {
+       struct net_device       *dev;
+       struct net_device       *stream_dev;
+       struct net              *net;
+       /* Global lock for amt device */
+       spinlock_t              lock;
+       /* Used only in relay mode */
+       struct list_head        tunnel_list;
+       struct gro_cells        gro_cells;
+
+       /* Protected by RTNL */
+       struct delayed_work     discovery_wq;
+       /* Protected by RTNL */
+       struct delayed_work     req_wq;
+       /* Protected by RTNL */
+       struct delayed_work     secret_wq;
+       /* AMT status */
+       enum amt_status         status;
+       /* Generated key */
+       siphash_key_t           key;
+       struct socket     __rcu *sock;
+       u32                     max_groups;
+       u32                     max_sources;
+       u32                     hash_buckets;
+       u32                     hash_seed;
+       /* Default 128 */
+       u32                     max_tunnels;
+       /* Default 128 */
+       u32                     nr_tunnels;
+       /* Gateway or Relay mode */
+       u32                     mode;
+       /* Default 2268 */
+       __be16                  relay_port;
+       /* Default 2268 */
+       __be16                  gw_port;
+       /* Outer local ip */
+       __be32                  local_ip;
+       /* Outer remote ip */
+       __be32                  remote_ip;
+       /* Outer discovery ip */
+       __be32                  discovery_ip;
+       /* Only used in gateway mode */
+       __be32                  nonce;
+       /* Gateway sent request and received query */
+       bool                    ready4;
+       bool                    ready6;
+       u8                      req_cnt;
+       u8                      qi;
+       u64                     qrv;
+       u64                     qri;
+       /* Used only in gateway mode */
+       u64                     mac:48,
+                               reserved:16;
+};
+
+#define AMT_MAX_GROUP          32
+#define AMT_MAX_SOURCE         128
+#define AMT_HSIZE_SHIFT                8
+#define AMT_HSIZE              (1 << AMT_HSIZE_SHIFT)
+
+#define AMT_INIT_QUERY_INTERVAL        125
+#define IANA_AMT_UDP_PORT      2268
+#define AMT_MAX_TUNNELS         128
+#define AMT_MAX_REQS           128
+#define AMT_GW_HLEN (sizeof(struct iphdr) + \
+                    sizeof(struct udphdr) + \
+                    sizeof(struct amt_gw_headers))
+#define AMT_RELAY_HLEN (sizeof(struct iphdr) + \
+                    sizeof(struct udphdr) + \
+                    sizeof(struct amt_relay_headers))
+
+static inline bool netif_is_amt(const struct net_device *dev)
+{
+       return dev->rtnl_link_ops && !strcmp(dev->rtnl_link_ops->kind, "amt");
+}
+
+#endif /* _NET_AMT_H_ */
diff --git a/include/uapi/linux/amt.h b/include/uapi/linux/amt.h
new file mode 100644 (file)
index 0000000..2dccff4
--- /dev/null
@@ -0,0 +1,62 @@
+/* SPDX-License-Identifier: GPL-2.0-only WITH Linux-syscall-note */
+/*
+ * Copyright (c) 2021 Taehee Yoo <ap420073@gmail.com>
+ */
+#ifndef _UAPI_AMT_H_
+#define _UAPI_AMT_H_
+
+enum ifla_amt_mode {
+       /* AMT interface works as Gateway mode.
+        * The Gateway mode encapsulates IGMP/MLD traffic and decapsulates
+        * multicast traffic.
+        */
+       AMT_MODE_GATEWAY = 0,
+       /* AMT interface works as Relay mode.
+        * The Relay mode encapsulates multicast traffic and decapsulates
+        * IGMP/MLD traffic.
+        */
+       AMT_MODE_RELAY,
+       __AMT_MODE_MAX,
+};
+
+#define AMT_MODE_MAX (__AMT_MODE_MAX - 1)
+
+enum {
+       IFLA_AMT_UNSPEC,
+       /* This attribute specify mode etier Gateway or Relay. */
+       IFLA_AMT_MODE,
+       /* This attribute specify Relay port.
+        * AMT interface is created as Gateway mode, this attribute is used
+        * to specify relay(remote) port.
+        * AMT interface is created as Relay mode, this attribute is used
+        * as local port.
+        */
+       IFLA_AMT_RELAY_PORT,
+       /* This attribute specify Gateway port.
+        * AMT interface is created as Gateway mode, this attribute is used
+        * as local port.
+        * AMT interface is created as Relay mode, this attribute is not used.
+        */
+       IFLA_AMT_GATEWAY_PORT,
+       /* This attribute specify physical device */
+       IFLA_AMT_LINK,
+       /* This attribute specify local ip address */
+       IFLA_AMT_LOCAL_IP,
+       /* This attribute specify Relay ip address.
+        * So, this is not used by Relay.
+        */
+       IFLA_AMT_REMOTE_IP,
+       /* This attribute specify Discovery ip address.
+        * When Gateway get started, it send discovery message to find the
+        * Relay's ip address.
+        * So, this is not used by Relay.
+        */
+       IFLA_AMT_DISCOVERY_IP,
+       /* This attribute specify number of maximum tunnel. */
+       IFLA_AMT_MAX_TUNNELS,
+       __IFLA_AMT_MAX,
+};
+
+#define IFLA_AMT_MAX (__IFLA_AMT_MAX - 1)
+
+#endif /* _UAPI_AMT_H_ */