Further netlink implementation. Not working yet.
[l2tpns.git] / l2tpns.c
index af19f9a..8230945 100644 (file)
--- a/l2tpns.c
+++ b/l2tpns.c
@@ -40,6 +40,7 @@ char const *cvs_id_l2tpns = "$Id: l2tpns.c,v 1.176 2011/01/20 12:48:40 bodea Exp
 #include <sys/sysinfo.h>
 #include <libcli.h>
 #include <linux/netlink.h>
+#include <linux/rtnetlink.h>
 
 #include "md5.h"
 #include "l2tpns.h"
@@ -65,8 +66,6 @@ int clifd = -1;                       // Socket listening for CLI connections.
 int daefd = -1;                        // Socket listening for DAE connections.
 int snoopfd = -1;              // UDP file handle for sending out intercept data
 int *radfds = NULL;            // RADIUS requests file handles
-int ifrfd = -1;                        // File descriptor for routing, etc
-int ifr6fd = -1;               // File descriptor for IPv6 routing, etc
 int rand_fd = -1;              // Random data source
 int cluster_sockfd = -1;       // Intra-cluster communications socket.
 int epollfd = -1;              // event polling
@@ -525,86 +524,101 @@ static void initnetlink(void)
 
        nlfd = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
        if (nlfd < 0)
-               return; // don't use netlink
+       {
+               LOG(0, 0, 0, "Can't create netlink socket: %s\n", strerror(errno));
+               exit(1);
+       }
 
        memset(&nladdr, 0, sizeof(nladdr));
        nladdr.nl_family = AF_NETLINK;
        nladdr.nl_pid = getpid();
 
-       if (bind(fd, (struct sockaddr *)&nladdr, sizeof(nladdr)) < 0)
+       if (bind(nlfd, (struct sockaddr *)&nladdr, sizeof(nladdr)) < 0)
        {
-               close(fd);
-               fd = -1;
-               return; // don't use netlink
+               LOG(0, 0, 0, "Can't bind netlink socket: %s\n", strerror(errno));
+               exit(1);
        }
 }
 
-static ssize_t netlink_send(struct nlmsghdr *nh)
+static ssize_t netlink_send(struct nlmsghdr *nh, int want_ack)
 {
        struct sockaddr_nl nladdr;
        struct iovec iov;
        struct msghdr msg;
-       ssize_t len;
-       char buf[4096];
-       struct nlmsghdr *ack_nh;
 
        nh->nlmsg_pid = getpid();
        nh->nlmsg_seq = ++nlseqnum;
-       nh->nlmsg_flags |= NLM_F_ACK;
+       if (want_ack)
+               nh->nlmsg_flags |= NLM_F_ACK;
 
-       iov = { (void *)nh, nh->msg_len };
        // set kernel address
-       memset(nladdr, 0, sizeof(nladdr));
+       memset(&nladdr, 0, sizeof(nladdr));
        nladdr.nl_family = AF_NETLINK;
-       msg = { (void *)&nladdr, sizeof(nladdr), &iov, 1, NULL, 0, 0 };
 
-       if ((len = sendmsg(nlfd, &msg, 0)) < 0)
-               return len; // return error code
+       iov = (struct iovec){ (void *)nh, nh->nlmsg_len };
+       msg = (struct msghdr){ (void *)&nladdr, sizeof(nladdr), &iov, 1, NULL, 0, 0 };
 
-       // expect ack
-       iov = { buf, sizeof(buf) };
-       msg = { (void *)&nladdr, sizeof(nladdr), &iov, 1, NULL, 0, 0 };
-       if ((len = recvmsg(nlfd, &msg, 0)) < 0)
-               return len; // return error code
+       return sendmsg(nlfd, &msg, 0);
+}
+
+static ssize_t netlink_recv(char *buf, ssize_t len)
+{
+       struct sockaddr_nl nladdr;
+       struct iovec iov;
+       struct msghdr msg;
+
+       // set kernel address
+       memset(&nladdr, 0, sizeof(nladdr));
+       nladdr.nl_family = AF_NETLINK;
+
+       iov = (struct iovec){ buf, len };
+       msg = (struct msghdr){ (void *)&nladdr, sizeof(nladdr), &iov, 1, NULL, 0, 0 };
 
-       for (ack_nh = (struct nlmsghdr *)buf; NLMSG_OK (ack_nh, len);
-                       ack_nh = NLMSG_NEXT (ack_nh, len))
+       return recvmsg(nlfd, &msg, 0);
+}
+
+static ssize_t netlink_check_ack(struct nlmsghdr *ack_nh, int msg_count)
+{
+       ssize_t len;
+       uint32_t seqnum;
+
+       // expect ack
+       for (; NLMSG_OK (ack_nh, len); ack_nh = NLMSG_NEXT (ack_nh, len))
        {
                if (ack_nh->nlmsg_type == NLMSG_DONE)
-                       return 1; // didn't get the ack
+                       return msg_count != 0 || seqnum != nlseqnum;
+                       // 1 => lost an ack
 
                if (ack_nh->nlmsg_type == NLMSG_ERROR)
                {
                        struct nlmsgerr *errmsg = NLMSG_DATA(ack_nh);
                        if (errmsg->error)
                                return errmsg->error; // got an error back
-                       if (errmsg->msg.nlmsg_seq == nh->nlmsg_seq)
-                               return 0; // ack received
                        else
-                               return 2; // wrong seq number?!
+                       {
+                               // ack received
+                               --msg_count;
+                               if (errmsg->msg.nlmsg_seq > seqnum);
+                                       seqnum = errmsg->msg.nlmsg_seq;
+                       }
                }
-               // unknown message
-               LOG(3, 0, 0, "Got an unknown netlink message: type %d\n", ack_nh->nlmsg_type);
+               else
+                       // unknown message
+                       LOG(3, 0, 0, "Got an unknown netlink message: type %d\n", ack_nh->nlmsg_type);
        }
 
-       return 3; // no end message?!
+       return 1; // malformed message?!
 }
 
-// defined in linux/ipv6.h, but tricky to include from user-space
-// TODO: move routing to use netlink rather than ioctl
-struct in6_ifreq {
-       struct in6_addr ifr6_addr;
-       __u32 ifr6_prefixlen;
-       unsigned int ifr6_ifindex;
-};
-
 //
 // Set up TUN interface
 static void inittun(void)
 {
+       struct ifinfomsg ifinfo;
        struct ifreq ifr;
-       struct in6_ifreq ifr6;
-       struct sockaddr_in sin = {0};
+       char buf[4096];
+       struct nlmsghdr resp_nh;
+
        memset(&ifr, 0, sizeof(ifr));
        ifr.ifr_flags = IFF_TUN;
 
@@ -625,79 +639,184 @@ static void inittun(void)
        }
        assert(strlen(ifr.ifr_name) < sizeof(config->tundevice));
        strncpy(config->tundevice, ifr.ifr_name, sizeof(config->tundevice) - 1);
+
        {
-               struct nlmsghdr nlh;
+               // get the interface index
+               ssize_t len;
+               struct {
+                       struct nlmsghdr nh;
+                       struct ifinfomsg ifinfo __attribute__ ((aligned(NLMSG_ALIGNTO)));
+                       struct rtattr ifname_rta __attribute__ ((aligned(RTA_ALIGNTO)));
+                       char ifname[IFNAMSIZ];
+               } req;
+
+               req.nh.nlmsg_type = RTM_GETLINK;
+               req.nh.nlmsg_flags = NLM_F_REQUEST;
 
+               req.ifinfo.ifi_family = AF_UNSPEC; // as the man says
+
+               req.ifname_rta.rta_len = RTA_LENGTH(strlen(config->tundevice)+1);
+               req.ifname_rta.rta_type = IFLA_IFNAME;
+               strncpy(req.ifname, config->tundevice, IFNAMSIZ-1);
+
+               if(netlink_send(&req.nh, 0) < 0 || (len = netlink_recv(buf, sizeof(buf))) < 0)
                {
-                       LOG(0, 0, 0, "Error setting up tun device: %s", strerror(errno));
+                       LOG(0, 0, 0, "Error getting tun ifindex: %s\n", strerror(errno));
                        exit(1);
                }
-       }
-       ifrfd = socket(PF_INET, SOCK_DGRAM, IPPROTO_IP);
 
-       sin.sin_family = AF_INET;
-       sin.sin_addr.s_addr = config->bind_address ? config->bind_address : 0x01010101; // 1.1.1.1
-       memcpy(&ifr.ifr_addr, &sin, sizeof(struct sockaddr));
-
-       if (ioctl(ifrfd, SIOCSIFADDR, (void *) &ifr) < 0)
-       {
-               LOG(0, 0, 0, "Error setting tun address: %s\n", strerror(errno));
-               exit(1);
-       }
-       /* Bump up the qlen to deal with bursts from the network */
-       ifr.ifr_qlen = 1000;
-       if (ioctl(ifrfd, SIOCSIFTXQLEN, (void *) &ifr) < 0)
-       {
-               LOG(0, 0, 0, "Error setting tun queue length: %s\n", strerror(errno));
-               exit(1);
-       }
-       /* set MTU to modem MRU */
-       ifr.ifr_mtu = MRU;
-       if (ioctl(ifrfd, SIOCSIFMTU, (void *) &ifr) < 0)
-       {
-               LOG(0, 0, 0, "Error setting tun MTU: %s\n", strerror(errno));
-               exit(1);
-       }
-       ifr.ifr_flags = IFF_UP;
-       if (ioctl(ifrfd, SIOCSIFFLAGS, (void *) &ifr) < 0)
-       {
-               LOG(0, 0, 0, "Error setting tun flags: %s\n", strerror(errno));
-               exit(1);
+               resp_nh = (struct nlmsghdr *)buf;
+               if (NLMSG_OK (resp_nh, len) && resp_nh.nlmsg_type == RTM_GETLINK)
+               {
+                       memcpy(&ifinfo, NLMSG_DATA(resp_nh), sizeof(ifinfo));
+                       // got index
+                       tunidx = ifinfo.ifi_index;
+               }
+               else
+               {
+                       LOG(0, 0, 0, "Malformed answer getting tun ifindex\n");
+                       exit(1);
+               }
        }
-       if (ioctl(ifrfd, SIOCGIFINDEX, (void *) &ifr) < 0)
        {
-               LOG(0, 0, 0, "Error getting tun ifindex: %s\n", strerror(errno));
-               exit(1);
-       }
-       tunidx = ifr.ifr_ifindex;
+               ssize_t err;
+               struct {
+                       // interface setting
+                       struct nlmsghdr nh;
+                       struct ifinfomsg ifinfo __attribute__ ((aligned(NLMSG_ALIGNTO)));
+                       struct rtattr txqlen_rta __attribute__ ((aligned(RTA_ALIGNTO)));
+                       uint32_t txqlen;
+                       struct rtattr mtu_rta __attribute__ ((aligned(RTA_ALIGNTO)));
+                       uint32_t mtu;
+
+                       // IPv4 address
+                       struct nlmsghdr addr_nh __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct ifaddrmsg ifaddr __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct rtattr addr_rta __attribute __ ((aligned(RTA_ALIGNTO)));
+                       struct in_addr addr;
+
+                       // IPv6 LL address
+                       struct nlmsghdr lladdr6_nh __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct ifaddrmsg llifaddr6 __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct rtattr lladdr6_rta __attribute __ ((aligned(RTA_ALIGNTO)));
+                       struct in6_addr lladdr6;
+
+                       // IPv6 global address
+                       struct nlmsghdr addr6_nh __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct ifaddrmsg ifaddr6 __attribute __ ((aligned(NLMSG_ALIGNTO)));
+                       struct rtattr addr6_rta __attribute __ ((aligned(RTA_ALIGNTO)));
+                       struct in6_addr addr6;
+
+                       // end header
+                       struct nlmsghdr end_nh __attribute __ ((aligned(NLMSG_ALIGNTO)));
+               } req;
+               char buf[4096];
+
+               memset(&req, 0, sizeof(req));
+
+               req.nh.nlmsg_type = RTM_NEWLINK;
+               req.nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_REPLACE | NLM_F_MULTI;
+
+               req.ifinfo = ifinfo;
+               req.ifinfo.ifi_flags |= IFF_UP; // set interface up
+               req.ifinfo.ifi_change = IFF_UP; // only change this flag
+
+               req.txqlen_rta.rta_len = RTA_LENGTH(sizeof(req.txqlen));
+               req.txqlen_rta.rta_type = IFLA_TXQLEN;
+               /* Bump up the qlen to deal with bursts from the network */
+               req.txqlen = 1000;
+
+               req.mtu_rta.rta_len = RTA_LENGTH(sizeof(req.txqlen));
+               req.mtu_rta.rta_type = IFLA_MTU;
+               /* set MTU to modem MRU */
+               req.mtu = MRU;
+
+               req.nh.nlmsg_len = NLMSG_LENGTH(sizeof(req.ifinfo)
+                               + req.txqlen_rla.rta_len
+                               + req.mtu_rta.rta_len);
+
+               req.addr_nh.nlmsg_type = RTM_NEWADDR;
+               req.addr_nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE | NLM_F_MULTI;
+
+               req.ifaddr.ifa_family = AF_INET;
+               req.ifaddr.ifa_prefixlen = 32;
+               req.ifaddr.ifa_index = ifinfo.ifi_index;
+
+               req.addr_rta.rta_len = RTA_LENGTH(sizeof(req.addr));
+               req.addr_rta.rta_type = IFA_ADDRESS;
+               req.addr.s_addr = config->bind_address ?
+                       config->bind_address : 0x01010101; // 1.1.1.1
+
+               req.addr_nh.nlmsg_len = NLMSG_LENGTH(sizeof(req.ifaddr)
+                               + req.addr_rta.rta_len);
+
+               // Only setup IPv6 on the tun device if we have a configured prefix
+               if (config->ipv6_prefix.s6_addr[0]) {
+                       req.lladdr6_nh.nlmsg_type = RTM_NEWADDR;
+                       req.lladdr6_nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE | NLM_F_MULTI;
+
+                       req.llifaddr6.ifa_family = AF_INET6;
+                       req.llifaddr6.ifa_prefixlen = 64;
+                       req.llifaddr6.ifa_scope = RT_SCOPE_LINK;
+                       req.llifaddr6.ifa_index = ifinfo.ifi_index;
+
+                       req.lladdr6_rta.rta_len = RTA_LENGTH(sizeof(req.lladdr6));
+                       req.lladdr6_rta.rta_type = IFA_LOCAL;
+                       // Link local address is FE80::1
+                       memset(&req.lladdr6, 0, sizeof(req.lladdr6));
+                       req.lladdr6.s6_addr[0] = 0xFE;
+                       req.lladdr6.s6_addr[1] = 0x80;
+                       req.lladdr6.s6_addr[15] = 1;
+
+                       req.lladdr6_nh.nlmsg_len = NLMSG_LENGTH(sizeof(req.ifaddr)
+                                       + req.lladdr6_rta.rta_len);
+
+                       req.addr6_nh.nlmsg_type = RTM_NEWADDR;
+                       req.addr6_nh.nlmsg_flags = NLM_F_REQUEST | NLM_F_CREATE | NLM_F_MULTI;
+
+                       req.ifaddr6.ifa_family = AF_INET6;
+                       req.ifaddr6.ifa_prefixlen = 64;
+                       req.ifaddr6.ifa_scope = RT_SCOPE_UNIVERSE;
+                       req.ifaddr6.ifa_index = ifinfo.ifi_index;
+
+                       req.addr6_rta.rta_len = RTA_LENGTH(sizeof(req.addr6));
+                       req.addr6_rta.rta_type = IFA_LOCAL;
+                       // Global address is prefix::1
+                       req.addr6 = config->ipv6_prefix;
+                       req.addr6.s6_addr[15] = 1;
+
+                       req.lladdr6_nh.nlmsg_len = NLMSG_LENGTH(sizeof(req.ifaddr)
+                                       + req.lladdr6_rta.rta_len);
+
+                       req.end_nh.nlmsg_type = NLMSG_DONE;
+                       req.end_nh.nlmsg_len = NLMSG_LENGTH(0);
+               }
+               else
+               {
+                       // we terminate at the beginning of IPv6 stuff
+                       req.lladdr6_nh.nlmsg_type = NLMSG_DONE;
+                       req.lladdr6_nh.nlmsg_len = NLMSG_LENGTH(0);
+               }
 
-       // Only setup IPv6 on the tun device if we have a configured prefix
-       if (config->ipv6_prefix.s6_addr[0]) {
-               ifr6fd = socket(PF_INET6, SOCK_DGRAM, 0);
+               if (netlink_send(&req.nh, 1) < 0)
+                       LOG(0, 0, 0, "Error setting up tun device: %s\n", strerror(errno));
+                       exit(1);
+               }
 
-               // Link local address is FE80::1
-               memset(&ifr6.ifr6_addr, 0, sizeof(ifr6.ifr6_addr));
-               ifr6.ifr6_addr.s6_addr[0] = 0xFE;
-               ifr6.ifr6_addr.s6_addr[1] = 0x80;
-               ifr6.ifr6_addr.s6_addr[15] = 1;
-               ifr6.ifr6_prefixlen = 64;
-               ifr6.ifr6_ifindex = ifr.ifr_ifindex;
-               if (ioctl(ifr6fd, SIOCSIFADDR, (void *) &ifr6) < 0)
+               if (netlink_recv(buf, sizeof(buf)) < 0)
                {
-                       LOG(0, 0, 0, "Error setting tun IPv6 link local address:"
-                               " %s\n", strerror(errno));
+                       LOG(0, 0, 0, "Error waiting ack for setting up tun device: %s\n", strerror(errno));
+                       exit(1);
                }
 
-               // Global address is prefix::1
-               memset(&ifr6.ifr6_addr, 0, sizeof(ifr6.ifr6_addr));
-               ifr6.ifr6_addr = config->ipv6_prefix;
-               ifr6.ifr6_addr.s6_addr[15] = 1;
-               ifr6.ifr6_prefixlen = 64;
-               ifr6.ifr6_ifindex = ifr.ifr_ifindex;
-               if (ioctl(ifr6fd, SIOCSIFADDR, (void *) &ifr6) < 0)
+               resp_nh = (struct nlmsghdr *)buf;
+               if ((err = netlink_check_ack(resp_nh, 4)))
                {
-                       LOG(0, 0, 0, "Error setting tun IPv6 global address: %s\n",
-                               strerror(errno));
+                       if (err < 0)
+                               LOG(0, 0, 0, "Error while receiving tun device ack: %s\n", strerror(errno));
+                       else
+                               LOG(0, 0, 0, "Error while receiving tun device ack\n", strerror(errno));
+                       exit(1);
                }
        }
 }