changeset 1510:4f3a5d600803

Updated version, having complete support for ip _link_, _addr_, _rule_, _route_ and _tunnel_ options.
author Ashwini Sharma <ak.ashwini1981@gmail.com>
date Sun, 28 Sep 2014 19:34:53 -0500
parents 22691dfb17b9
children df792340e3f7
files toys/pending/ip.c
diffstat 1 files changed, 1805 insertions(+), 0 deletions(-) [+]
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/toys/pending/ip.c	Sun Sep 28 19:34:53 2014 -0500
@@ -0,0 +1,1805 @@
+/* ip.c - Show / manipulate routing, devices, policy routing and tunnels.
+ *
+ * Copyright 2014 Sameer Prakash Pradhan <sameer.p.pradhan@gmail.com>
+ * Copyright 2014 Ranjan Kumar <ranjankumar.bth@gmail.com>
+ * Copyright 2014 Rajni Kant <rajnikant12345@gmail.com>
+ *
+ * No Standard.
+ *
+USE_IP(NEWTOY(ip, NULL, TOYFLAG_SBIN))
+USE_IP(OLDTOY(ipaddr, ip, NULL, TOYFLAG_SBIN))
+USE_IP(OLDTOY(iplink, ip, NULL, TOYFLAG_SBIN))
+USE_IP(OLDTOY(iproute, ip, NULL, TOYFLAG_SBIN))
+USE_IP(OLDTOY(iprule, ip, NULL, TOYFLAG_SBIN))
+USE_IP(OLDTOY(iptunnel, ip, NULL, TOYFLAG_SBIN))
+
+config IP
+  bool "ip"
+  default n
+  help
+    usage: ip [ OPTIONS ] OBJECT { COMMAND }
+
+    Show / manipulate routing, devices, policy routing and tunnels.
+
+    where OBJECT := {address | link | route | rule | tunnel}
+    OPTIONS := { -f[amily] { inet | inet6 | link } | -o[neline] }
+*/
+#define FOR_ip
+#include "toys.h"
+#include <linux/netlink.h>
+#include <linux/rtnetlink.h>
+#include <linux/if_ether.h>
+#include <linux/if_addr.h>
+#include <net/if_arp.h>
+#include <ifaddrs.h>
+#include <fnmatch.h>
+//=============================================================================
+
+// TODO List
+/*
+ * 1. At this point of time, not clear about using hashing technique
+ *    is useful to improve performance.
+ * 2. Pending decision:
+ *  a. rt_realms, rt_dsfield and rt_tables will get free or not.
+ *  b. rt_realms, rt_dsfield and rt_tables will be global or local.
+ * 3. Use atolx_range instead of strtoul.
+ * 4. substring_to_idx for multiple match cases.
+ */
+
+GLOBALS(
+  char stats;
+  int sockfd;
+  char gbuf[8192];
+  int8_t addressfamily;
+  int8_t is_addr;
+  char singleline;
+  char flush;
+  char *filter_dev;
+)
+
+struct arglist {
+  char *name;
+  int idx;
+};
+
+static struct 
+{
+  int ifindex, scope, scopemask, up, to;
+  char *label, *addr;
+} addrinfo;
+
+struct linkdata  {
+  struct linkdata *next, *prev;
+  int flags, iface_idx, mtu, txqueuelen, parent,iface_type;
+  char qdiscpline[IFNAMSIZ+1], state[IFNAMSIZ+1], type[IFNAMSIZ+1],
+       iface[IFNAMSIZ+1], laddr[64], bcast[64];
+  struct  rtnl_link_stats rt_stat;
+}*linfo;
+
+typedef int (*cmdobj)(char **argv);
+
+#define MESG_LEN 8192
+
+// For "/etc/iproute2/RPDB_tables"
+enum {
+  RPDB_rtdsfield = 1,
+  RPDB_rtrealms = 3,
+  RPDB_rtscopes = 4,
+  RPDB_rttables = 5
+};
+
+#define RPDB_ENTRIES 256
+static int8_t rttable_init;
+static int8_t rtdsfield_init;
+static int8_t rtscope_init;
+static int8_t rtrealms_init;
+
+static struct arglist default_table = {"default", RT_TABLE_DEFAULT};
+static struct arglist main_table = {"main", RT_TABLE_MAIN};
+static struct arglist local_table = {"local", RT_TABLE_LOCAL};
+static struct arglist unspec_realms = {"unspec", RT_TABLE_UNSPEC};
+
+static struct arglist *rt_dsfield[RPDB_ENTRIES] = {0};
+static struct arglist *rt_tables[RPDB_ENTRIES] = {
+  [RT_TABLE_DEFAULT] = &default_table,
+  [RT_TABLE_MAIN] = &main_table,
+  [RT_TABLE_LOCAL] = &local_table,
+};
+static struct arglist *rt_realms[RPDB_ENTRIES] = {
+  [0] = &unspec_realms,
+};
+
+// default values from /etc/iproute2/rt_scope
+static struct arglist global_table = {"global", 0};
+static struct arglist site_table = {"site", 200};
+static struct arglist link_table = {"link", 253};
+static struct arglist host_table = {"host", 254};
+static struct arglist nowhere_table = {"nowhere", 255};
+static struct arglist *rt_scope[256] = {
+  [RT_SCOPE_UNIVERSE] = &global_table,
+  [RT_SCOPE_SITE] = &site_table,
+  [RT_SCOPE_LINK] = &link_table,
+  [RT_SCOPE_HOST] = &host_table,
+  [RT_SCOPE_NOWHERE] = &nowhere_table,
+};
+
+static struct arglist rtmtypes[] = { {"none", RTN_UNSPEC},
+  {"unicast", RTN_UNICAST}, {"local", RTN_LOCAL},
+  {"broadcast", RTN_BROADCAST}, {"anycast", RTN_ANYCAST},
+  {"multicast", RTN_MULTICAST}, {"blackhole", RTN_BLACKHOLE},
+  {"unreachable", RTN_UNREACHABLE}, {"prohibit", RTN_PROHIBIT},
+  {"throw", RTN_THROW}, {"nat", RTN_NAT},
+  {"xresolve", RTN_XRESOLVE}, {NULL, -1}
+};
+
+static int filter_nlmesg(int (*fun)(struct nlmsghdr *mhdr, char **), char **);
+static int  ipaddr_print(struct linkdata *, int flg);
+
+
+// ===========================================================================
+// Common Code for IP Options (like: addr, link, route etc.)
+// ===========================================================================
+static int substring_to_idx(char *str, struct arglist *list)
+{
+  struct arglist *alist;
+  int idx = -1, mcount = 0, len;
+
+  if (!str) return -1;
+  len = strlen(str);
+
+  for (alist = list; alist->name; alist++) {
+    if (!memcmp(str, alist->name, len)) {
+      if (!*(alist->name + len))
+        return alist->idx;
+      idx = alist->idx;
+      ++mcount; // No of matches.
+    }
+  }
+  if (mcount > 1) return -1; // Multiple match found.
+  return idx;
+}
+
+static int string_to_idx(char *str, struct arglist *list)
+{
+  struct arglist *alist;
+
+  if (!str) return -1;
+  for (alist = list; alist->name; alist++)
+    if (!strcmp(str, alist->name)) return alist->idx;
+  return -1;
+}
+
+static char *idx_to_string(int idx, struct arglist *list)
+{
+  struct arglist *alist;
+
+  if (idx < 0) return NULL;
+  for (alist = list; alist->name; alist++)
+    if (idx == alist->idx) return alist->name;
+  return NULL;
+}
+
+static void iphelp(void)
+{
+  toys.exithelp = 1;
+  error_exit(NULL);
+}
+
+static void send_nlmesg(int type, int flags, int family,
+    void *buf, int blen)
+{
+  struct {
+    struct nlmsghdr nlh;
+    struct rtgenmsg g;
+  } req;
+
+  if (!buf) {
+    memset(&req, 0, sizeof(req));
+    req.nlh.nlmsg_len = sizeof(req);
+    req.nlh.nlmsg_type = type;
+    req.nlh.nlmsg_flags = flags;
+    req.g.rtgen_family = family;
+    buf = &req;
+    blen = sizeof(req);
+  }
+  if (send(TT.sockfd , (void*)buf, blen, 0) < 0)
+    perror_exit("Unable to send data on socket.");
+}
+
+// Parse /etc/iproute2/RPDB_tables and prepare list.
+static void parseRPDB(char *fname, struct arglist **list, int32_t size)
+{
+  char *line;
+  int fd = open(fname, O_RDONLY);
+
+  if (fd < 0) return;
+  for (; (line = get_line(fd)); free(line)) {
+    char *ptr = line;
+    int32_t idx;
+
+    while (*ptr == ' ' || *ptr == '\t') ptr++;
+    if (*ptr == 0 || *ptr == '#' || *ptr == '\n') continue;
+    if ((sscanf(ptr, "0x%x %s\n", &idx, toybuf) != 2) &&
+        (sscanf(ptr, "0x%x %s #", &idx, toybuf) != 2) &&
+        (sscanf(ptr, "%d %s\n", &idx, toybuf) != 2) &&
+        (sscanf(ptr, "%d %s #", &idx, toybuf) != 2)) {
+      error_msg("Corrupted '%s' file", fname);
+      xclose(fd);
+      free(line);
+      return;
+    }
+    if (idx >= 0 && idx < size) {
+      int index = idx & (size-1);
+      list[index] = xzalloc(sizeof(struct arglist));
+      list[index]->idx = idx;
+      list[index]->name = xstrdup(toybuf);
+    }
+  }
+  xclose(fd);
+}
+
+static struct arglist **getlist(u_int8_t whichDB)
+{
+  struct arglist **alist;
+
+  switch (whichDB) {
+    case RPDB_rtdsfield:
+      alist = rt_dsfield;
+      if (!rtdsfield_init) {
+        rtdsfield_init = 1;
+        parseRPDB("/etc/iproute2/rt_dsfield", alist, ARRAY_LEN(rt_dsfield));
+      }
+      break;
+    case RPDB_rtrealms:
+      alist = rt_realms;
+      if (!rtrealms_init) {
+        rtrealms_init = 1;
+        parseRPDB("/etc/iproute2/rt_realms", alist, ARRAY_LEN(rt_realms));
+      }
+      break;
+    case RPDB_rtscopes:
+      alist = rt_scope;
+      if (!rtscope_init) {
+        rtscope_init = 1;
+        parseRPDB("/etc/iproute2/rt_scopes", alist, ARRAY_LEN(rt_scope));
+      }
+      break;
+    case RPDB_rttables:
+      alist = rt_tables;
+      if (!rttable_init) {
+        rttable_init = 1;
+        parseRPDB("/etc/iproute2/rt_tables", alist, ARRAY_LEN(rt_tables));
+      }
+      break;
+    default: 
+      error_exit("wrong database");
+      break; // Unreachable code.
+  }
+  return alist;
+}
+
+/*
+ * Parse RPBD tables (if not parsed already).
+ * return RPDB table name as per idx.
+ */
+static char *namefromRPDB(int idx, u_int8_t whichDB)
+{
+  struct arglist **alist;
+
+  if (idx < RT_TABLE_UNSPEC || idx >= RPDB_ENTRIES) {
+    snprintf(toybuf, RPDB_ENTRIES, "%u", idx);
+    return toybuf;
+  }
+
+  alist = getlist(whichDB);
+
+  if (alist[idx] && alist[idx]->name) return alist[idx]->name;
+  if (whichDB == RPDB_rtdsfield) snprintf(toybuf, RPDB_ENTRIES, "0x%02x", idx);
+  else snprintf(toybuf, RPDB_ENTRIES, "%u", idx);
+
+  return toybuf;
+}
+
+static int idxfromRPDB(char *name, u_int8_t whichDB)
+{
+  struct arglist **alist;
+  int i = 0;
+
+  for (alist = getlist(whichDB); i < RPDB_ENTRIES; i++) {
+    if (!alist[i] || !alist[i]->name) continue;
+    if (!strcmp(alist[i]->name, name)) return i;
+  }
+  return (atolx_range(name, 0, 255));
+}
+
+/*
+ * TODO:
+ * rtmtype_idx2str and rtmtype_str2idx interfaces may be used in IPROUTE too.
+ */
+static char *rtmtype_idx2str(u_int8_t idx)
+{
+  char *name = idx_to_string(idx, rtmtypes);
+
+  if (!name) snprintf(toybuf, RPDB_ENTRIES, "%u", idx);
+  else strcpy(toybuf, name);
+
+  return toybuf;
+}
+
+static int rtmtype_str2idx(char *name)
+{
+  int idx = string_to_idx(name, rtmtypes);
+
+  if (idx < 0) return atolx_range(name, 0, 255);
+  return idx;
+}
+
+/*
+ * Used to get the prefix value in binary form.
+ * For IPv4: non-standard parsing used; as 10.10 will be treated as 10.10.0.0
+ * unlike inet_aton which is 10.0.0.10
+ */
+static int get_prefix(uint32_t *addr, uint8_t *af, char *name, int family)
+{
+  if (family == AF_PACKET) error_exit("'%s' may be inet prefix", name);
+  if (!memcmp(name, "default", strlen(name))
+      || !memcmp(name, "all", strlen(name))
+      || !memcmp(name, "any", strlen(name))) {
+    *af = family;
+    return 0;
+  }
+  if (strchr(name, ':')) {
+    *af = AF_INET6;
+    if (family != AF_UNSPEC && family != AF_INET6) return 1;
+    if (inet_pton(AF_INET6, name, (void *)addr) != 1) 
+      return 1;
+  } else { // for IPv4.
+    char *ptr = name;
+    uint8_t count = 0;
+
+    *af = AF_INET;
+    if (family != AF_UNSPEC && family != AF_INET) return 1;
+    while (*ptr) {
+      int val, len = 0;
+
+      if (*ptr == '.') ptr++;
+      sscanf(ptr, "%d%n", &val, &len);
+      if (!len || len > 3 || val < 0 || val > 255 || count > 3) return 1;
+      ptr += len;
+      ((uint8_t*)addr)[count++] = val;
+    }
+  }
+  return 0;
+}
+
+/*
+ * Used to calculate netmask, which can be in the form of
+ * either 255.255.255.0 or 24 or default or any or all strings.
+ */
+static int get_nmask_prefix(uint32_t *netmask, uint8_t af,
+    char *name, uint8_t family)
+{
+  char *ptr;
+  uint32_t naddr[4] = {0,};
+  uint64_t plen;
+  uint8_t naf = AF_UNSPEC;
+
+  *netmask = (af == AF_INET6) ? 128 : 32; // set default netmask
+  plen = strtoul(name, &ptr, 0);
+
+  if (!ptr || ptr == name || *ptr || !plen || plen > *netmask) {
+    if (get_prefix(naddr, &naf, name, family)) return -1;
+    if (naf == AF_INET) {
+      uint32_t mask = htonl(*naddr), host = ~mask;
+      if (host & (host + 1)) return -1;
+      for (plen = 0; mask; mask <<= 1) ++plen;
+      if (plen > 32) return -1;
+    }
+  }
+  *netmask = plen;
+  return 0;
+}
+
+/*
+ * Parse prefix, which will be in form of
+ * either default or default/default or default/24 or default/255.255.255.0
+ * or 10.20.30.40 or 10.20.30.40/default or 10.20.30.40/24
+ * or 10.20.30.40/255.255.255.0
+ */
+static void parse_prefix(uint32_t *addr, uint32_t *netmask, uint8_t *len,
+    char *name, int family)
+{
+  uint8_t af = AF_UNSPEC;
+  char *slash = strchr(name, '/');
+
+  if (slash) *slash = 0;
+  if (get_prefix(addr, &af, name, family)) error_exit("Invalid prefix");
+
+  if (slash) { // grab netmask.
+    if (get_nmask_prefix(netmask, af, slash+1, family))
+      error_exit("Invalid prefix");
+    *slash ='/';
+  }
+  else if (af == AF_INET && *addr) *netmask = 32;
+  else if (af == AF_INET6 && (*addr || *(addr+3))) *netmask = 128;
+
+  if (!*addr && !slash && !af) *len = 0;
+  else *len = (af == AF_INET6) ? 16 : 4;
+}
+
+static void add_string_to_rtattr(struct nlmsghdr *n, int maxlen,
+    int type, void *data, int alen)
+{
+  int len = RTA_LENGTH(alen);
+  struct rtattr *rta;
+
+  if ((int)(NLMSG_ALIGN(n->nlmsg_len) + len) > maxlen) return;
+  rta = (struct rtattr*)(((char*)n) + NLMSG_ALIGN(n->nlmsg_len));
+  rta->rta_type = type;
+  rta->rta_len = len;
+  memcpy(RTA_DATA(rta), data, alen);
+  n->nlmsg_len = NLMSG_ALIGN(n->nlmsg_len) + len;
+}
+
+
+
+// ===========================================================================
+// Code for ip link.
+// ===========================================================================
+#ifndef NLMSG_TAIL
+#define NLMSG_TAIL(nmsg) \
+  ((struct rtattr *) (((void *) (nmsg)) + NLMSG_ALIGN((nmsg)->nlmsg_len)))
+#endif
+
+static uint32_t get_ifaceindex(char *name, int ext)
+{
+  struct if_nameindex *if_ni, *i;
+  int index = -1;
+  if_ni = if_nameindex();
+  if (if_ni == NULL) {
+    perror("if_nameindex");
+    exit(EXIT_FAILURE);
+  }
+
+  for (i = if_ni; ! (i->if_index == 0 && i->if_name == NULL); i++)
+    if(!strcmp(name,i->if_name)) { index =i->if_index; break; }
+  if_freenameindex(if_ni);
+  if (index == -1 && ext) perror_exit("can't find device '%s'", name);
+  return index;
+}
+
+static void fill_hwaddr(char *arg, int len, unsigned char *address)
+{
+  int count = 0, val, length;
+
+  while (count < len) {
+    val = length = 0;
+    if (*arg == ':') arg++, count++;
+    sscanf(arg, "%2x%n", &val, &length);
+    if (!length || length > 2)
+      error_exit("bad hw-addr '%s'", arg ? arg : "");
+    arg += length;
+    count += length;
+    *address++ = val;
+  }
+}
+
+// Multimach = 1, single match = 0
+static char *get_flag_string(struct arglist *aflags, int flags, int ismulti)
+{
+  struct arglist *p = aflags;
+  char *out = NULL, *tmp = NULL;
+
+  for (; p->name; p++) {
+    int test = (ismulti ? p->idx & flags : 0) || p->idx == flags;
+    if (test) { // flags can be zero
+      tmp = out ? xmprintf("%s,%s", out, p->name) : xmprintf("%s", p->name);
+      if (out) free(out);
+      out = tmp;
+    }
+  }
+  return out;
+}
+
+static void vlan_parse_opt(char **argv, struct nlmsghdr *n, unsigned int size)
+{
+  struct arglist vlan_optlist[] = {{"id", 0}, {"protocol", 1}, 
+    {"reorder_hdr", 2}, {"gvrp", 3}, {NULL,-1}};
+  struct arglist vlan_protolist[] = {{"802.1q", 0}, {"802.1ad", 1}, {NULL,-1}};
+  struct arglist on_off[] = { {"on", 0}, {"off", 1}, {NULL,-1}};
+  int idx;
+  struct ifla_vlan_flags flags;
+
+  memset(&flags, 0, sizeof(flags));
+  for (; *argv; argv++) {
+    int param, proto;
+
+    if ((idx = substring_to_idx(*argv++, vlan_optlist)) == -1) iphelp();
+    switch (idx) {
+      case 0: // ARG_id
+        if (!*argv) iphelp();
+        param = atolx(*argv);
+        add_string_to_rtattr(n, size, IFLA_VLAN_ID, &param, sizeof(param));
+        break;
+      case 1: // ARG_protocol
+        if (!*argv) error_exit("Invalid vlan id.");
+        if ((idx = substring_to_idx(*argv, vlan_protolist)) == -1) iphelp();
+        if (!idx) proto = ETH_P_8021Q; // PROTO_8021Q - 0
+        else if (idx == 1) proto = 0x88A8; // ETH Protocol - 8021AD
+        // IFLA VLAN PROTOCOL - 5
+        add_string_to_rtattr(n, size, 5, &proto, sizeof(proto));
+        break;
+      case 2: // ARG_reorder_hdr
+      case 3: // ARG_gvrp
+        if ((param = substring_to_idx(*argv, on_off)) == -1) iphelp();
+
+        flags.mask |= (idx -1); // VLAN FLAG REORDER Header              
+        flags.flags &= ~(idx -1); // VLAN FLAG REORDER Header            
+        if (!param) flags.flags |= (idx -1); // VLAN FLAG REORDER Header
+        break;
+    }
+  }
+  if (flags.mask) 
+    add_string_to_rtattr(n, size, IFLA_VLAN_FLAGS, &flags, sizeof(flags));
+}
+
+static int linkupdate(char **argv)
+{
+  struct {
+    struct nlmsghdr mhdr;
+    struct ifinfomsg info;
+    char buf[1024];
+  } request;  
+  char *name, *dev, *type, *link, *addr;
+  struct rtattr *attr = NULL;
+  int len = 0, add = (*argv[-1] == 'a') ? 1 : 0;
+
+  name = dev = type = link = addr = NULL;
+  for (; *argv; argv++) {
+    struct arglist objectlist[] = { {"type", 0}, {"name", 1}, {"link", 2}, 
+      {"address", 3}, {NULL,-1}};
+    uint8_t idx = substring_to_idx(*argv, objectlist);
+
+    if (!idx) {
+      type = *++argv;
+      break;
+    }
+    else if (idx == 1) dev = name = *++argv;
+    else if (idx == 2) link = *++argv;
+    else if (idx == 3) addr = *++argv;
+    else if (!dev) name = dev = *argv;
+  }
+
+  if (!name && !add)
+    error_exit("Not enough information: \"dev\" argument is required.\n");
+  else if (!type  && add)
+    error_exit("Not enough information: \"type\" argument is required.\n");
+
+  memset(&request, 0, sizeof(request));
+  request.mhdr.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifinfomsg));
+  request.mhdr.nlmsg_flags = NLM_F_REQUEST|NLM_F_ACK;
+  if (add) {
+    request.mhdr.nlmsg_flags |= NLM_F_CREATE|NLM_F_EXCL;
+    request.mhdr.nlmsg_type = RTM_NEWLINK;
+  } else {
+    request.mhdr.nlmsg_type = RTM_DELLINK;
+    request.info.ifi_index = get_ifaceindex(name, 1);
+  }
+  request.info.ifi_family = AF_UNSPEC;
+  attr = NLMSG_TAIL(&request.mhdr);
+  if (type) {
+    add_string_to_rtattr(&request.mhdr, sizeof(request),
+        IFLA_LINKINFO, NULL, 0);
+    add_string_to_rtattr(&request.mhdr, sizeof(request),
+        IFLA_INFO_KIND, type, strlen(type));
+    if (!strcmp(type, "vlan")) {
+      struct rtattr *data = NLMSG_TAIL(&request.mhdr);
+      add_string_to_rtattr(&request.mhdr, sizeof(request), 
+          IFLA_INFO_DATA, NULL, 0);
+      vlan_parse_opt(++argv, &request.mhdr, sizeof(request));
+      data->rta_len = (void *)NLMSG_TAIL(&request.mhdr) - (void *)data;
+    }
+    attr->rta_len = (void *)NLMSG_TAIL(&request.mhdr) - (void *)attr;
+  }
+
+  if (link) {
+    uint32_t idx = get_ifaceindex(link, 1);
+    add_string_to_rtattr(&request.mhdr, sizeof(request), 
+        IFLA_LINK, &idx, sizeof(uint32_t));
+  }
+  if (addr) {
+    char abuf[IF_NAMESIZE] = {0,};
+
+    fill_hwaddr(addr, IF_NAMESIZE, (unsigned char *)abuf);
+    add_string_to_rtattr(&request.mhdr, sizeof(request), 
+        IFLA_ADDRESS, abuf, strlen(abuf));
+  }
+  if (!name) {
+    snprintf(toybuf, IFNAMSIZ, "%s%d", type, 0);
+    for (len = 1; ; len++) {
+      if (!get_ifaceindex(toybuf, 0)) break;
+      snprintf(toybuf, IFNAMSIZ, "%s%d", type, len);
+    }
+    name = toybuf;
+  }
+  len = strlen(name) + 1;
+  if (len < 2 || len > IFNAMSIZ) error_exit("Invalid device name.");
+  add_string_to_rtattr(&request.mhdr, sizeof(request), IFLA_IFNAME, name, len);
+
+  send_nlmesg(0, 0, 0, (void *)&request, request.mhdr.nlmsg_len);
+  return (filter_nlmesg(NULL,NULL));
+}
+
+static int link_set(char **argv)
+{
+  struct arglist cmd_objectlist[] = {{"up", 0}, {"down", 1}, {"arp", 2}, 
+    {"multicast", 3}, {"dynamic", 4}, {"name", 5}, {"txqueuelen", 6}, 
+    {"mtu", 7},{"address", 8}, {"broadcast", 9}, {NULL,-1}};
+  int case_flags[] = {IFF_NOARP,IFF_MULTICAST,IFF_DYNAMIC};
+  struct ifreq req;
+  int idx, flags = 0, masks = 0xffff, fd;
+
+  memset(&req, 0, sizeof(req));
+  if (!*argv) error_exit("\"dev\" missing");
+  strncpy(req.ifr_name, *argv, IF_NAMESIZE);
+  fd = xsocket(AF_INET, SOCK_DGRAM, 0);
+  xioctl(fd, SIOCGIFINDEX, &req);
+  for (++argv; *argv;) {
+    if ((idx = substring_to_idx(*argv++, cmd_objectlist)) == -1) iphelp();
+    switch(idx) {
+      case 0:
+        flags |= IFF_UP; break;
+      case 1:
+        masks &= ~IFF_UP; break;
+      case 2:
+      case 3:
+      case 4:
+        if (!*argv) iphelp();
+        else if (!strcmp(*argv, "on")) {
+          if (idx == 2) {
+            masks &= ~case_flags[idx-2];
+            flags &= ~case_flags[idx-2];
+          } else flags |= case_flags[idx-2];
+        } else if (!strcmp(*argv,"off")) {
+          if (idx == 2) {
+            masks |= case_flags[idx-2];
+            flags |= case_flags[idx-2];
+          } else masks &= ~case_flags[idx-2];
+        } else iphelp();
+        ++argv;
+        break;
+      case 5:
+        strncpy(req.ifr_ifru.ifru_newname, *argv, IF_NAMESIZE);
+        xioctl(fd, SIOCSIFNAME, &req);
+        strncpy(req.ifr_name, *argv++, IF_NAMESIZE);
+        xioctl(fd, SIOCGIFINDEX, &req);
+        break;
+      case 6:
+        req.ifr_ifru.ifru_ivalue = atolx(*argv++);
+        xioctl(fd, SIOCSIFTXQLEN, &req);
+        break;
+      case 7:
+        req.ifr_ifru.ifru_mtu = atolx(*argv++);
+        xioctl(fd, SIOCSIFMTU, &req);
+        break;
+      case 8:
+        xioctl(fd, SIOCGIFHWADDR, &req);
+        fill_hwaddr(*argv++, IF_NAMESIZE, 
+            (unsigned char *)(req.ifr_hwaddr.sa_data));
+        xioctl(fd, SIOCSIFHWADDR, &req);
+        break;
+      case 9:
+        xioctl(fd, SIOCGIFHWADDR, &req);
+        fill_hwaddr(*argv++, IF_NAMESIZE,
+            (unsigned char *)(req.ifr_hwaddr.sa_data));
+        xioctl(fd, SIOCSIFHWBROADCAST, &req);
+        break;
+    }
+  }
+  xioctl(fd, SIOCGIFFLAGS, &req);
+  req.ifr_ifru.ifru_flags |= flags;
+  req.ifr_ifru.ifru_flags &= masks;
+  xioctl(fd, SIOCSIFFLAGS, &req);
+  xclose(fd);
+  return 0;
+}
+
+static void print_stats(struct  rtnl_link_stats *rtstat)
+{
+  char *line_feed = (!TT.singleline ? "\n    " : " ");
+
+  if (TT.stats > 0) {
+    xprintf("    RX: bytes  packets  errors  "
+        "dropped  overrun  mcast%s%-10u %-8u %-7u %-8u %-8u %-8u\n",
+        line_feed, rtstat->rx_bytes, rtstat->rx_packets, rtstat->rx_errors,
+        rtstat->rx_dropped, rtstat->rx_over_errors, rtstat->multicast);
+    if (TT.stats > 1) {
+      xprintf("    RX: errors  length  crc  "
+          "frame  fifo  missed%s%-10u %-8u %-7u %-8u %-8u %-8u\n",
+          line_feed, rtstat->rx_errors, rtstat->rx_length_errors,
+          rtstat->rx_crc_errors, rtstat->rx_frame_errors,
+          rtstat->rx_fifo_errors, rtstat->rx_missed_errors);
+    }
+    xprintf("    TX: bytes  packets  errors  "
+        "dropped  carrier  collsns%s%-10u %-8u %-7u %-8u %-8u %-8u\n",
+        line_feed, rtstat->tx_bytes, rtstat->tx_packets, rtstat->tx_errors,
+        rtstat->tx_dropped, rtstat->tx_carrier_errors, rtstat->collisions);
+    if (TT.stats > 1) {
+      xprintf("    TX: errors  aborted  fifo  window  "
+          "heartbeat%s%-10u %-8u %-7u %-8u %-8u\n",
+          line_feed, rtstat->tx_errors, rtstat->tx_aborted_errors,
+          rtstat->tx_fifo_errors, rtstat->tx_window_errors, 
+          rtstat->tx_heartbeat_errors);
+    }
+  }
+}
+
+static int print_link_output(struct linkdata *link)
+{
+  char *line_feed = " ", *flags,*peer = "brd";
+  struct arglist iface_flags[] = {{"",0},{"UP", IFF_UP}, 
+    {"BROADCAST", IFF_BROADCAST}, {"DEBUG", IFF_DEBUG},
+    {"LOOPBACK", IFF_LOOPBACK}, {"POINTOPOINT", IFF_POINTOPOINT},
+    {"NOTRAILERS", IFF_NOTRAILERS}, {"RUNNING", IFF_RUNNING},
+    {"NOARP", IFF_NOARP}, {"PROMISC",IFF_PROMISC},
+    {"ALLMULTI", IFF_ALLMULTI}, {"MASTER", IFF_MASTER}, {"SLAVE", IFF_SLAVE},
+    {"MULTICAST", IFF_MULTICAST}, {"PORTSEL", IFF_PORTSEL},
+    {"AUTOMEDIA", IFF_AUTOMEDIA}, {"DYNAMIC", IFF_DYNAMIC}, {NULL,-1}};
+
+  if (link->parent != -1) {
+    int fd = 0;
+    struct ifreq req;
+
+    memset(&req, 0, sizeof(req));
+    if_indextoname( link->parent,req.ifr_ifrn.ifrn_name);
+    fd = xsocket(AF_INET, SOCK_DGRAM, 0);
+    if (ioctl(fd, SIOCGIFTXQLEN, &req)) perror("");
+    else link->txqueuelen = req.ifr_ifru.ifru_ivalue;
+    xclose(fd);
+  }
+
+  if(TT.is_addr && addrinfo.label && fnmatch(addrinfo.label, link->iface, 0))
+    return 0;
+
+
+  if (!(flags = get_flag_string(iface_flags, link->flags, 1)))
+    error_exit("Invalid data.");    
+  if (!TT.singleline) line_feed="\n    ";
+  if (link->parent != -1) {
+    char iface[IF_NAMESIZE];
+
+    if (!if_indextoname(link->parent, iface)) perror_exit(NULL);
+    sprintf(toybuf,"%s@%s", link->iface, iface);
+  }
+  if(link->flags & IFF_POINTOPOINT) peer = "peer";
+  if (TT.is_addr && TT.singleline && TT.addressfamily)
+    xprintf("%d: %s", link->iface_idx,
+        ((link->parent == -1) ? link->iface : toybuf));
+  else xprintf("%d: %s: <%s> mtu %d qdisc %s state %s qlen %d",
+      link->iface_idx, ((link->parent == -1) ? link->iface : toybuf), flags,
+      link->mtu, link->qdiscpline, link->state, link->txqueuelen);
+
+  if (!TT.addressfamily || TT.addressfamily == AF_PACKET)
+    xprintf("%slink/%s %s %s %s",
+        line_feed, link->type, link->laddr, peer ,link->bcast);
+
+  xputc('\n');
+
+  //user can specify stats flag two times
+  //one for stats and other for erros e.g. -s and -s -s
+  print_stats(&link->rt_stat);
+  free(flags);
+
+  return 0;
+}
+
+static void fill_address(void *p, char *ip)
+{
+  unsigned char *ptr = (unsigned char*)p;
+  snprintf(ip, 64, " %02x:%02x:%02x:%02x:%02x:%02x",
+      ptr[0], ptr[1], ptr[2], ptr[3], ptr[4], ptr[5]);
+}
+
+static int get_link_info(struct nlmsghdr* h,struct linkdata* link,char **argv)
+{
+  struct ifinfomsg *iface = NLMSG_DATA(h);
+  struct rtattr *attr = IFLA_RTA(iface);
+  int len = h->nlmsg_len - NLMSG_LENGTH(sizeof(*iface));
+  struct arglist hwtypes[]={{"generic",0},{"ether",ARPHRD_ETHER},
+    {"loopback", ARPHRD_LOOPBACK},{"sit",ARPHRD_SIT},
+#ifdef ARPHRD_INFINIBAND
+    {"infiniband",ARPHRD_INFINIBAND},
+#endif
+#ifdef ARPHRD_IEEE802_TR
+    {"ieee802",ARPHRD_IEEE802}, {"tr",ARPHRD_IEEE802_TR},
+#else
+    {"tr",ARPHRD_IEEE802},
+#endif
+#ifdef ARPHRD_IEEE80211
+    {"ieee802.11",ARPHRD_IEEE80211},
+#endif
+#ifdef ARPHRD_IEEE1394
+    {"ieee1394",ARPHRD_IEEE1394},
+#endif
+    {"irda",ARPHRD_IRDA},{"slip",ARPHRD_SLIP},{"cslip",ARPHRD_CSLIP},
+    {"slip6",ARPHRD_SLIP6}, {"cslip6",ARPHRD_CSLIP6}, {"ppp",ARPHRD_PPP},
+    {"ipip",ARPHRD_TUNNEL}, {"tunnel6",ARPHRD_TUNNEL6},
+    {"gre",ARPHRD_IPGRE},
+#ifdef ARPHRD_VOID
+    {"void",ARPHRD_VOID},
+#endif
+    {NULL,-1}};
+  char *lname = get_flag_string(hwtypes, iface->ifi_type, 0);
+
+  link->next = link->prev = 0;
+  link->iface_type = iface->ifi_type;
+  if (!lname) error_exit("Invalid link.");
+  strncpy(link->type, lname, IFNAMSIZ);
+  free(lname);
+  link->iface_idx = iface->ifi_index;
+  link->flags = iface->ifi_flags;
+  if (*argv && !strcasecmp("up",*argv) && !(link->flags & IFF_UP)) return 1;
+  link->parent =  -1;
+  for (; RTA_OK(attr, len); attr = RTA_NEXT(attr, len)) {
+    switch(attr->rta_type) {
+      case IFLA_IFNAME:
+        snprintf(link->iface, IFNAMSIZ, "%s",(char *) RTA_DATA(attr));
+        break;
+      case IFLA_ADDRESS:
+        if ( iface->ifi_type== ARPHRD_TUNNEL ||
+            iface->ifi_type == ARPHRD_SIT ||
+            iface->ifi_type == ARPHRD_IPGRE)
+          inet_ntop(AF_INET, RTA_DATA(attr), link->laddr, 64);
+        else fill_address(RTA_DATA(attr), link->laddr);
+        break;
+      case IFLA_BROADCAST:
+        if (iface->ifi_type== ARPHRD_TUNNEL ||
+            iface->ifi_type == ARPHRD_SIT ||
+            iface->ifi_type == ARPHRD_IPGRE)
+          inet_ntop(AF_INET, RTA_DATA(attr), link->bcast, 64);
+        else  fill_address(RTA_DATA(attr), link->bcast);
+        break;
+      case IFLA_MTU:
+        link->mtu = *((int*)(RTA_DATA(attr)));
+        break;
+      case IFLA_QDISC:
+        snprintf(link->qdiscpline, IFNAMSIZ, "%s", (char *) RTA_DATA(attr));
+        break;
+      case IFLA_STATS  :
+        link->rt_stat = *((struct rtnl_link_stats*) RTA_DATA(attr));
+        break;
+      case IFLA_LINK:
+        link->parent = *((int*)(RTA_DATA(attr)));
+        break;
+      case IFLA_TXQLEN:
+        link->txqueuelen = *((int*)(RTA_DATA(attr)));
+        break;
+      case IFLA_OPERSTATE: 
+        {
+          struct arglist flags[]={{"UNKNOWN", 0}, {"NOTPRESENT", 1}, 
+            {"DOWN", 2}, {"LOWERLAYERDOWN", 3}, {"TESTING", 4}, 
+            {"DORMANT", 5}, {"UP", 6}, {NULL, -1}};
+          if (!(lname = get_flag_string(flags, *((int*)(RTA_DATA(attr))), 0)))
+            error_exit("Invalid state.");
+          strncpy(link->state, lname,IFNAMSIZ);
+          free(lname);
+        }
+        break;
+      default: break;
+    }
+  }
+  return 0;
+}
+
+static int display_link_info(struct nlmsghdr *mhdr, char **argv)
+{
+  struct linkdata link;
+
+  if(!get_link_info(mhdr, &link, argv)){
+    if(TT.is_addr) {
+      struct linkdata *lnk = xzalloc(sizeof(struct linkdata));
+      memcpy(lnk, &link, sizeof(struct linkdata));
+      dlist_add_nomalloc((struct double_list **)&linfo,
+          (struct double_list *)lnk);
+    }
+    else print_link_output(&link);
+  }
+  return 0;
+}
+
+static int link_show(char **argv)
+{
+  struct {
+    struct nlmsghdr mhdr;
+    struct ifinfomsg info;
+  } request;
+  uint32_t index = 0;
+
+  if (*argv && strcasecmp("up",*argv)) index = get_ifaceindex(*argv, 1);
+  memset(&request, 0, sizeof(request));
+  request.mhdr.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifinfomsg));
+  request.mhdr.nlmsg_flags = NLM_F_REQUEST|NLM_F_ACK;
+  if (!index) request.mhdr.nlmsg_flags |= NLM_F_ROOT|NLM_F_MATCH;
+  else request.info.ifi_change =  0xffffffff; // used in single operation
+  request.mhdr.nlmsg_type = RTM_GETLINK;
+  request.info.ifi_index = index;
+  request.info.ifi_family = AF_UNSPEC;
+  send_nlmesg(0, 0, 0, (void*)&request, sizeof(request));
+  return (filter_nlmesg(display_link_info, argv));
+}
+
+static int iplink(char **argv)
+{
+  int idx;
+  cmdobj ipcmd, cmdobjlist[] = {linkupdate, link_set, link_show};
+  struct arglist cmd_objectlist[] = {{"add", 0}, {"delete", 0},
+    {"set", 1}, {"show", 2}, {"list", 2}, {"lst", 2}, {NULL,-1}};
+
+  if (!*argv) idx = 2;
+  else if ((idx = substring_to_idx(*argv++, cmd_objectlist)) == -1) iphelp();
+  ipcmd = cmdobjlist[idx];
+  return ipcmd(argv);
+}
+
+// ===========================================================================
+// Code for ip addr.
+// ===========================================================================
+
+static int print_addrinfo(struct nlmsghdr *h, int flag_l)
+{
+  struct rtattr *rta, *rta_tb[IFA_MAX+1] = {0,};
+  char *family = toybuf, *scope = toybuf+256, *label = toybuf+512,
+       *brd = toybuf+768, *peer = toybuf+1024, *any = toybuf+1280,
+       lbuf[INET6_ADDRSTRLEN] = {0,}, lbuf_ifa[INET6_ADDRSTRLEN] = {0,};
+  struct ifaddrmsg *ifa = NLMSG_DATA(h);
+  int len;
+
+  if ((len = h->nlmsg_len - NLMSG_LENGTH(sizeof(*ifa))) < 0) {
+    error_msg("wrong nlmsg len %d", len);
+    return 0;
+  }
+
+  for (rta = IFA_RTA(ifa); RTA_OK(rta, len); rta=RTA_NEXT(rta, len))
+    if (rta->rta_type <= IFA_MAX) rta_tb[rta->rta_type] = rta;
+
+  if (!rta_tb[IFA_LOCAL]) rta_tb[IFA_LOCAL] = rta_tb[IFA_ADDRESS];
+  if (!rta_tb[IFA_ADDRESS]) rta_tb[IFA_ADDRESS] = rta_tb[IFA_LOCAL];
+  if ((addrinfo.scope ^ ifa->ifa_scope)&addrinfo.scopemask) return 0;
+  if (addrinfo.ifindex && addrinfo.ifindex != ifa->ifa_index) return 0;
+
+  if (flag_l && addrinfo.label && ifa->ifa_family == AF_INET6) return 0;
+  if ((rta_tb[IFA_LABEL])) {
+    strcpy(label, RTA_DATA(rta_tb[IFA_LABEL]));
+    if(addrinfo.label && fnmatch(addrinfo.label, label, 0))
+      return 0;
+  }
+
+  if (TT.flush) {
+    if (ifa->ifa_index == addrinfo.ifindex) {
+      h->nlmsg_type = RTM_DELADDR;
+      h->nlmsg_flags = NLM_F_REQUEST;
+      send_nlmesg(RTM_DELADDR, 0, 0, h, h->nlmsg_len);
+      return 0;
+    }
+  }
+
+  if (h->nlmsg_type == RTM_DELADDR) printf("Deleted ");
+
+  if (TT.singleline) {
+    if (!if_indextoname(ifa->ifa_index, lbuf)) perror_exit(NULL);
+    printf("%u: %s",ifa->ifa_index, lbuf);
+  }
+
+  sprintf(scope, " scope %s ", namefromRPDB(ifa->ifa_scope, RPDB_rtscopes));
+
+  if (ifa->ifa_family == AF_INET) strcpy(family, "    inet ");
+  else if (ifa->ifa_family == AF_INET6) strcpy(family, "    inet6 ");
+  else sprintf(family, "    family %d", ifa->ifa_family);
+
+  if (rta_tb[IFA_LOCAL]) {
+    if (!inet_ntop(ifa->ifa_family, RTA_DATA(rta_tb[IFA_LOCAL]),
+          lbuf, sizeof(lbuf))) perror_exit("inet");
+
+    sprintf(family+strlen(family), lbuf, strlen(lbuf));
+    if (!rta_tb[IFA_ADDRESS] || !memcmp(RTA_DATA(rta_tb[IFA_ADDRESS]),
+          RTA_DATA(rta_tb[IFA_LOCAL]), 4))
+      sprintf(family+strlen(family), "/%d ", ifa->ifa_prefixlen);
+    else {
+      if (!inet_ntop(ifa->ifa_family, RTA_DATA(rta_tb[IFA_ADDRESS]),
+            lbuf_ifa, sizeof(lbuf_ifa))) perror_exit("inet");
+      sprintf(peer, " peer %s/%d ", lbuf_ifa, ifa->ifa_prefixlen);
+    }
+  }
+
+  if (addrinfo.to && strcmp(addrinfo.addr, lbuf))
+    return 0;
+
+  if (rta_tb[IFA_BROADCAST]) {
+    if (!inet_ntop(ifa->ifa_family, RTA_DATA(rta_tb[IFA_BROADCAST]),
+          lbuf, sizeof(lbuf))) perror_exit("inet");
+    sprintf(brd, " brd %s", lbuf);
+  }else brd = "";
+
+  if (rta_tb[IFA_ANYCAST]) {
+    if (!inet_ntop(ifa->ifa_family, RTA_DATA(rta_tb[IFA_ANYCAST]),
+          lbuf, sizeof(lbuf))) perror_exit("inet");
+    sprintf(any, " any %s", lbuf);
+  }
+
+  if (ifa->ifa_family == AF_INET)
+    printf("%s%s%s%s%s %c", family, brd, peer, scope, label,
+        (TT.singleline? '\0' : '\n'));
+  else printf("%s%s %c", family, scope, (TT.singleline? '\0' : '\n'));
+  if (TT.singleline && (ifa->ifa_family == AF_INET)) xputc('\n');
+
+  if (rta_tb[IFA_CACHEINFO]) {
+    struct ifa_cacheinfo *ci = RTA_DATA(rta_tb[IFA_CACHEINFO]);
+
+    printf("%c      valid_lft ", (TT.singleline? '\\' : '\0'));
+    if (ci->ifa_valid ==  0xFFFFFFFFU) printf("forever");
+    else printf("%usec", ci->ifa_valid);
+    printf(" preferred_lft ");
+    if (ci->ifa_prefered ==  0xFFFFFFFFU) printf("forever");
+    else printf("%dsec", ci->ifa_prefered);
+    xputc('\n');
+  }
+  return 0;
+}
+
+static int ipaddrupdate(char **argv)
+{
+  int length, cmd = !memcmp("add", argv[-1], strlen(argv[-1]))
+    ? RTM_NEWADDR: RTM_DELADDR;
+  int idx = 0,length_brd = 0, length_peer = 0,length_any = 0,length_local = 0,
+      scoped = 0;
+  char *dev = NULL,*label = NULL, reply[8192];
+
+  unsigned    scope;
+  struct nlmsghdr *addr_ptr = NULL;
+  struct nlmsgerr *err = NULL;
+  struct arglist cmd_objectlist[] = {{"dev",0}, {"peer", 1},
+    {"remote", 2}, {"broadcast", 3}, {"brd", 4}, {"label", 5},
+    {"anycast", 6},{"scope", 7}, {"local", 8}, {NULL, -1}};
+  struct{
+    struct nlmsghdr nlm;
+    struct ifaddrmsg ifadd;
+    char buf[256];
+  } req;
+  typedef struct{
+    int family, bytelen, bitlen;
+    __u32  data[8];
+  }option_data;
+  option_data local;
+
+  memset(&req, 0, sizeof(req));
+  req.nlm.nlmsg_len = NLMSG_LENGTH(sizeof(struct ifaddrmsg));
+  req.nlm.nlmsg_flags = NLM_F_REQUEST|NLM_F_ACK;
+  req.nlm.nlmsg_type = cmd;
+  req.ifadd.ifa_family = TT.addressfamily;
+
+  while(*argv) {
+    idx = substring_to_idx(*argv, cmd_objectlist);
+    if (idx >= 0)
+      if (!*++argv)
+        error_exit("Incomplete Command line");
+    switch(idx) {
+      case 0:
+        dev = *argv;
+        break;
+      case 1:
+      case 2:
+        {
+          uint32_t addr[4] = {0,}, netmask = 0;
+          uint8_t len = 0;
+          parse_prefix(addr, &netmask, &len, *argv,
+              req.ifadd.ifa_family);
+          if (len)
+            req.ifadd.ifa_family = ((len == 4) ? AF_INET : AF_INET6);
+          length_peer = len;
+          add_string_to_rtattr(&req.nlm, sizeof(req),
+              IFA_ADDRESS, addr, len);
+          req.ifadd.ifa_prefixlen = netmask;
+        }
+        break;
+      case 3:
+      case 4:
+        if (*argv[0] == '+') {
+          length_brd = -1;
+        } else if (*argv[0] == '-') {
+          length_brd = -2;
+        } else {
+          uint32_t addr[4] = {0,};
+          uint8_t af = AF_UNSPEC;
+
+          if (get_prefix(addr, &af, *argv, req.ifadd.ifa_family))
+            error_exit("Invalid prefix");
+
+          length_brd = ((af == AF_INET6) ? 16 : 4);
+          if (req.ifadd.ifa_family == AF_UNSPEC)
+            req.ifadd.ifa_family = af;
+          add_string_to_rtattr(&req.nlm, sizeof(req),
+              IFA_BROADCAST, &addr, length_brd);
+        }
+        break;
+      case 5:
+        label = *argv;
+        add_string_to_rtattr(&req.nlm, sizeof(req),
+            IFA_LABEL, label, strlen(label) + 1);
+        break;
+      case 6:
+        {
+          uint32_t addr[4] = {0,};
+          uint8_t af = AF_UNSPEC;
+
+          if (get_prefix(addr, &af, *argv, req.ifadd.ifa_family))
+            error_exit("Invalid prefix");
+          length_any = ((af == AF_INET6) ? 16 : 4);
+          if (req.ifadd.ifa_family == AF_UNSPEC)
+            req.ifadd.ifa_family = af;
+          add_string_to_rtattr(&req.nlm, sizeof(req),
+              IFA_ANYCAST, &addr, length_brd);
+        }
+        break;
+      case 7:
+        scope = idxfromRPDB(*argv, RPDB_rtscopes);
+        if (scope){
+          if (!strcmp(*argv, "scope"))
+            error_exit("wrong scope '%s'", *argv);
+        }
+        req.ifadd.ifa_scope = scope;
+        scoped =1;
+        break;
+      default:
+        {
+          //local is by default
+          uint32_t addr[8] = {0,}, netmask = 0;
+          uint8_t len = 0;
+
+          parse_prefix(addr, &netmask, &len, *argv,
+              req.ifadd.ifa_family);
+          if (len)
+            req.ifadd.ifa_family = ((len == 4) ? AF_INET : AF_INET6);
+          length_local = len;
+          local.bitlen = netmask;
+          local.bytelen = len;
+          memcpy(local.data, addr, sizeof(local.data));
+          local.family = req.ifadd.ifa_family;
+          add_string_to_rtattr(&req.nlm, sizeof(req),
+              IFA_LOCAL, &local.data, local.bytelen);
+        }
+        break;
+    }
+    argv++;
+  }
+  if (!dev) error_exit("need \"dev \" argument");
+  if (label && strncmp(dev, label, strlen(dev)) != 0)
+    error_exit("\"dev\" (%s) must match \"label\" (%s)", dev, label);
+
+  if (length_peer == 0 && length_local && cmd != RTM_DELADDR){
+    add_string_to_rtattr(&req.nlm, sizeof(req),
+        IFA_ADDRESS, &local.data, local.bytelen);
+  }
+
+  if (length_brd < 0 && cmd != RTM_DELADDR){
+    int i;
+
+    if (req.ifadd.ifa_family != AF_INET)
+      error_exit("broadcast can be set only for IPv4 addresses");
+
+    if (local.bitlen <= 30) {
+      for (i = 31; i >= local.bitlen; i--) {
+        if (length_brd == -1)
+          local.data[0] |= htonl(1<<(31-i));
+        else
+          local.data[0] &= ~htonl(1<<(31-i));
+      }
+      add_string_to_rtattr(&req.nlm, sizeof(req),
+          IFA_BROADCAST, &local.data, local.bytelen);
+      length_brd = local.bytelen;
+    }
+  }
+  if (req.ifadd.ifa_prefixlen == 0)
+    req.ifadd.ifa_prefixlen = local.bitlen;
+  if (!scoped && (cmd != RTM_DELADDR) && (local.family == AF_INET)
+      && (local.bytelen >= 1 && *(uint8_t*)&local.data == 127))
+    req.ifadd.ifa_scope = RT_SCOPE_HOST;
+  req.ifadd.ifa_index = get_ifaceindex(dev, 1);
+
+  send_nlmesg(RTM_NEWADDR, 0, AF_UNSPEC, (void *)&req, req.nlm.nlmsg_len);
+  length = recv(TT.sockfd, reply, sizeof(reply), 0);
+  addr_ptr = (struct nlmsghdr *) reply;
+  for (; NLMSG_OK(addr_ptr, length); addr_ptr = NLMSG_NEXT(addr_ptr, length)) {
+    if (addr_ptr->nlmsg_type == NLMSG_DONE)
+      return 1;
+    if (addr_ptr->nlmsg_type == NLMSG_ERROR)
+      err = (struct nlmsgerr*) NLMSG_DATA(addr_ptr);
+    if (err->error != 0){
+      errno = -err->error;
+      perror_exit("RTNETLINK answers:");
+    }
+  }
+  return 0;
+}
+
+static int ipaddr_listflush(char **argv)
+{
+  int idx; uint32_t netmask = 0, found = 0;
+  char *tmp = NULL, *name = NULL;
+  struct double_list *dlist;
+  struct arglist cmd_objectlist[] = {{"to", 0}, {"scope", 1}, {"up", 2},
+    {"label", 3}, {"dev", 4}, {NULL, -1}};
+
+  TT.flush = *argv[-1] == 'f' ? 1 : 0;
+  memset(&addrinfo, 0, sizeof(addrinfo));
+
+  if (TT.flush) {
+    if (!*argv)
+      error_exit("Incomplete command for \"flush\"");
+    if (TT.addressfamily == AF_PACKET)
+      error_exit("Can't flush link Addressess");
+  }
+  addrinfo.scope = -1;
+  while (*argv) {
+    unsigned scope = 0;
+
+    switch (idx = substring_to_idx(*argv, cmd_objectlist)) {
+      case 0: 
+        {// ADDR_TO
+          if (!*++argv) error_exit("Incomplete Command line");
+          else if(!strcmp(*argv, "0")) return 0;
+          uint32_t addr[4] = {0,};
+          uint8_t len = 0;
+
+          addrinfo.to = 1;
+          parse_prefix(addr, &netmask, &len, *argv, TT.addressfamily);
+          if (len)
+            TT.addressfamily = ((len == 4) ? AF_INET : AF_INET6);
+          addrinfo.addr  = strtok(*argv, "/");
+        }
+        break;
+      case 1: // ADDR_SCOPE
+        if (!*++argv) error_exit("Incomplete Command line");
+        name = *argv;
+
+        addrinfo.scopemask = -1;
+        if (isdigit(**argv)) {
+          int idx = atolx(*argv);
+
+          name = xstrdup(namefromRPDB(idx, RPDB_rtscopes));
+        }
+        scope = idxfromRPDB(name, RPDB_rtscopes);
+
+        if (!scope && strcmp(name, "global")) {
+          if (strcmp(name, "all"))
+            error_exit("wrong scope '%s'", name);
+          scope = RT_SCOPE_NOWHERE;
+          addrinfo.scopemask = 0;
+        }
+        if (isdigit(**argv))
+          free(name);
+        addrinfo.scope = scope;
+        break;       
+      case 2: // ADDR_UP
+        addrinfo.up = 1;
+        break;            
+      case 3: // ADDR_LABEL
+        if (!*++argv) error_exit("Incomplete Command line");
+        addrinfo.label = *argv;
+        break;
+      case 4: // ADDR_DEV
+        if (!*++argv) error_exit("Incomplete Command line");
+
+      default:                               
+        if (TT.filter_dev)
+          error_exit("Either \"dev\" is duplicate or %s is garbage",
+              *argv);
+        TT.filter_dev = *argv;
+        break;
+    }
+    argv++;
+  }
+
+  link_show(&tmp);
+  while ( linfo && (dlist = dlist_pop(&linfo))){    
+    struct linkdata *tmp  = (struct linkdata*) dlist;
+    char *temp = &tmp->iface[0];
+
+    if (TT.filter_dev && strcmp(TT.filter_dev, temp))
+      continue;
+    found = 1;
+    if (TT.flush && addrinfo.label) ipaddr_print( tmp, 0);
+    if (addrinfo.up && !(tmp->flags & IFF_UP)){
+      ipaddr_print(tmp, 0);
+      continue;
+    }
+    if (addrinfo.label){
+      if ( fnmatch(addrinfo.label, temp, 0)) {
+        ipaddr_print(tmp, 1);
+        continue;
+      }      
+    }
+    if (!TT.addressfamily && ! TT.flush ) print_link_output(tmp);
+
+    ipaddr_print(tmp, 0);
+    free(tmp);
+  }
+  if (TT.filter_dev && !found)
+    error_exit("Device \"%s\" doesn't exist. \n", TT.filter_dev);
+  return 0;
+}
+
+static int ipaddr_print( struct linkdata *link, int flag_l)
+{
+  struct nlmsghdr *addr_ptr;
+  int ip_match = 0;
+
+  addrinfo.ifindex = link->iface_idx;
+  send_nlmesg(RTM_GETADDR, NLM_F_ROOT|NLM_F_MATCH|NLM_F_REQUEST,
+      AF_UNSPEC, NULL, 0);
+  if (TT.addressfamily == AF_PACKET) print_link_output(link);
+
+  if(addrinfo.label){
+    char *col = strchr(addrinfo.label, ':');
+    if (!col && (fnmatch(addrinfo.label, &link->iface[0], 0)))
+      return 0;
+  }
+
+  while (1){
+    int len = recv(TT.sockfd, TT.gbuf, sizeof(TT.gbuf), 0);
+    addr_ptr = (struct nlmsghdr *)TT.gbuf;
+    struct ifaddrmsg *addressInfo = NLMSG_DATA(addr_ptr);
+    char lbuf[INET6_ADDRSTRLEN];
+    struct rtattr *rta, *rta_tb[IFA_MAX+1] = {0,};
+
+    int len1 = addr_ptr->nlmsg_len - NLMSG_LENGTH(sizeof(*addressInfo));
+    if (len1 > 0) {
+      for (; NLMSG_OK(addr_ptr, len); addr_ptr = NLMSG_NEXT(addr_ptr, len)) {
+        addressInfo = NLMSG_DATA(addr_ptr);     
+        if (TT.addressfamily && TT.addressfamily != addressInfo->ifa_family)
+          continue;
+        if (addrinfo.ifindex && addrinfo.ifindex != addressInfo->ifa_index)
+          continue;
+
+        if(addrinfo.to){        
+          memset(rta_tb, 0, sizeof(rta_tb));
+          int rt_len = IFA_PAYLOAD(addr_ptr);
+          for (rta = IFA_RTA(addressInfo); RTA_OK(rta, rt_len); rta=RTA_NEXT(rta, rt_len)) {
+            if (rta->rta_type <= IFA_MAX) rta_tb[rta->rta_type] = rta;
+          }
+          if (!rta_tb[IFA_LOCAL]) rta_tb[IFA_LOCAL] = rta_tb[IFA_ADDRESS];              
+          if (rta_tb[IFA_LOCAL]) {
+            if (!inet_ntop(TT.addressfamily, RTA_DATA(rta_tb[IFA_LOCAL]),
+                  lbuf, sizeof(lbuf))) perror_exit("inet");
+            if (strcmp(addrinfo.addr, lbuf))
+              continue;
+            ip_match=1;
+          }
+          if(!ip_match)
+            continue;
+        }
+
+        if(!TT.flush){
+          if (addrinfo.scope != -1 && TT.addressfamily && TT.addressfamily ==
+              addressInfo->ifa_family &&
+              (addrinfo.ifindex == addressInfo->ifa_index)) {
+            if(addrinfo.scope != addressInfo->ifa_scope)
+              continue;
+            else if (addrinfo.up && (link->flags & IFF_UP))
+              print_link_output(link);
+            else if (!addrinfo.up) print_link_output(link);
+          }
+          if (TT.addressfamily &&
+              (addrinfo.ifindex == addressInfo->ifa_index) &&
+              (addrinfo.scope == -1)){
+            if (addrinfo.up && (link->flags & IFF_UP))
+              print_link_output(link);
+            else if (!addrinfo.up) print_link_output(link);
+          }
+        }
+        for (; NLMSG_OK(addr_ptr, len); addr_ptr = NLMSG_NEXT(addr_ptr, len)) {
+          if ((addr_ptr->nlmsg_type == RTM_NEWADDR))
+            print_addrinfo(addr_ptr, flag_l);
+          if ((addr_ptr->nlmsg_type == NLMSG_DONE) ||
+              (addr_ptr->nlmsg_type == NLMSG_ERROR) ||
+              (TT.flush && addrinfo.to))
+            goto ret_stop;          
+        }
+      }
+    }
+    else
+      return 0;
+  }
+
+ret_stop:
+  return 0;
+}
+
+static int ipaddr(char **argv)
+{
+  int    idx;
+  cmdobj ipcmd, cmdobjlist[] = {ipaddrupdate, ipaddr_listflush};
+  struct arglist cmd_objectlist[] = { {"add", 0}, {"delete", 0},
+    {"list", 1},{"show", 1},{"lst", 1}, {"flush", 1}, {NULL,-1}};
+
+  TT.is_addr++;
+  if (!*argv) idx = 1;
+  else if ((idx = substring_to_idx(*argv++, cmd_objectlist)) == -1) iphelp();
+
+  ipcmd = cmdobjlist[idx];
+  return ipcmd(argv);
+}
+
+// ===========================================================================
+// TODO: code for ip route.
+// ===========================================================================
+static int iproute(char **argv)
+{
+  printf("__FUNCTION__ = %s\n", __FUNCTION__);
+  return 0;
+}
+
+// ===========================================================================
+// code for ip rule.
+// ===========================================================================
+static void show_iprule_help(void)
+{
+  char *errmsg = "Usage: ip rule [ list | add | del ] SELECTOR ACTION\n"
+    "SELECTOR := [ from PREFIX ] [ to PREFIX ] [pref NUMBER] [ tos TOS ]\n"
+    "            [ fwmark FWMARK] [ dev/iif STRING ] [type TYPE]\n"
+    "ACTION := [ table TABLE_ID ] [ realms [SRCREALM/]DSTREALM ]";
+
+  error_exit(errmsg);
+}
+
+static int ruleupdate(char **argv)
+{
+  int8_t idx, tflag = 0, opt = (*argv[-1] == 'a') ? RTM_NEWRULE : RTM_DELRULE;
+  struct arglist options[] = {{"from", 0}, {"to", 1}, {"preference", 2},
+    {"order", 2}, {"priority", 2}, {"tos", 3}, {"dsfield", 3}, {"fwmark", 4},
+    {"realms", 5}, {"table", 6}, {"lookup", 6}, {"dev", 7}, {"iif", 7},
+    {"nat", 8}, {"map-to", 8}, {"type", 9}, {"help", 10}, {NULL, -1}};
+  struct {
+    struct nlmsghdr mhdr;
+    struct rtmsg    msg;
+    char buf[1024];
+  } request;
+
+  memset(&request, 0, sizeof(request));
+  request.mhdr.nlmsg_type = opt;
+  request.mhdr.nlmsg_len = NLMSG_LENGTH(sizeof(struct rtmsg));
+  request.mhdr.nlmsg_flags = NLM_F_REQUEST | NLM_F_ACK |
+    ((opt == RTM_DELRULE) ? 0 : NLM_F_CREATE | NLM_F_EXCL);
+  request.msg.rtm_family = TT.addressfamily;
+  request.msg.rtm_protocol = RTPROT_BOOT;
+  request.msg.rtm_scope = RT_SCOPE_UNIVERSE;
+  request.msg.rtm_table = 0;
+  request.msg.rtm_type = ((opt == RTM_DELRULE) ? RTN_UNSPEC : RTN_UNICAST);
+
+  for (; *argv; argv++) {
+    switch ((idx = substring_to_idx(*argv, options))) {
+      case 0:
+      case 1: 
+        { // e.g. from IP/Netmask and to IP/Netmask.
+          uint32_t addr[4] = {0,}, netmask = 0;
+          uint8_t len = 0, *tmp;
+
+          if (!*++argv) error_exit("'%s': Missing Prefix", argv[-1]);
+          parse_prefix(addr, &netmask, &len, *argv, request.msg.rtm_family);
+
+          tmp = idx ? &request.msg.rtm_dst_len : &request.msg.rtm_src_len;
+          if (!netmask) *tmp = 0;
+          else *tmp = netmask;
+
+          add_string_to_rtattr(&request.mhdr, sizeof(request),
+              (idx ? RTA_DST : RTA_SRC), addr, len);
+        }
+        break;
+      case 2:
+      case 4: 
+        { // e.g. Preference p# and fwmark MARK
+          uint32_t pref;
+          char *ptr;
+
+          if (!*++argv)
+            error_exit("Missing %s", (idx == 2) ? "Preference" : "fwmark");
+          pref = strtoul(*argv, &ptr, 0);
+          if (!ptr || (ptr == *argv) || *ptr  || pref > 0xFFFFFFFFUL)
+            error_exit("Invalid %s",  (idx == 2) ? "Preference" : "fwmark");
+          add_string_to_rtattr(&request.mhdr, sizeof(request),
+              ((idx == 2) ? RTA_PRIORITY : RTA_PROTOINFO),
+              (void *)&pref, sizeof(uint32_t));
+        }
+        break;
+      case 3:
+        {
+          if (!*++argv) error_exit("Missing TOS key");
+          request.msg.rtm_tos = idxfromRPDB(*argv, RPDB_rtdsfield);
+        }
+        break;
+      case 5:
+        { // e.g. realms FROM_realm/TO_realm
+          uint32_t realms = 0;
+          char *ptr;
+
+          if (!*++argv) error_exit("Missing REALMSID");
+          if ((ptr = strchr(*argv, '/'))) {
+            *ptr = 0;
+            realms = idxfromRPDB(*argv, RPDB_rtrealms);
+            realms <<= 16;
+            *ptr++ = '/';
+          } else ptr = *argv;
+          realms |= idxfromRPDB(ptr, RPDB_rtrealms);
+          add_string_to_rtattr(&request.mhdr, sizeof(request),
+              RTA_FLOW, (void *)&realms, sizeof(uint32_t));
+        }
+        break;
+      case 6:
+        { // e.g. table tid/tableName
+          if (!*++argv) error_exit("Missing TableID");
+          request.msg.rtm_table = idxfromRPDB(*argv, RPDB_rttables);
+          tflag = 1;
+        }
+        break;
+      case 7:
+        {
+          if (!*++argv) error_exit("Missing dev/iif NAME");
+          add_string_to_rtattr(&request.mhdr, sizeof(request),
+              RTA_IIF, *argv, strlen(*argv)+1);
+        }
+        break;
+      case 8:
+        {
+          uint32_t addr[4] = {0,};
+          uint8_t af = AF_UNSPEC;
+
+          if (!*++argv) error_exit("Missing nat/map-to ADDRESS");
+          if (get_prefix(addr, &af /* Un-used variable */, *argv, AF_INET))
+            error_exit("Invalid mapping Address");
+
+          add_string_to_rtattr(&request.mhdr, sizeof(request),
+              RTA_GATEWAY, addr, sizeof(uint32_t));
+          request.msg.rtm_type = RTN_NAT;
+        }
+        break;
+      case 9:
+        {
+          if (!*++argv) error_exit("TYPE Missing");
+          request.msg.rtm_type = rtmtype_str2idx(*argv);
+        }
+        break;
+      case 10: 
+        show_iprule_help();
+        break; // Unreachable code.
+      default: 
+        error_exit("Invalid argument '%s'", *argv);
+        break; // Unreachable code.
+    }
+  }
+
+  if (!request.msg.rtm_family) request.msg.rtm_family = AF_INET;
+  if (!tflag && opt == RTM_NEWRULE) request.msg.rtm_table = RT_TABLE_MAIN;
+
+  send_nlmesg(0, 0, 0, &request, sizeof(request));
+  return (filter_nlmesg(NULL, NULL));
+}
+
+static int show_rules(struct nlmsghdr *mhdr,
+    char **argv __attribute__ ((__unused__)))
+{
+  struct rtmsg *msg = NLMSG_DATA(mhdr);
+  struct rtattr *rta, *attr[RTA_MAX+1] = {0,};
+  int32_t tvar, msglen = mhdr->nlmsg_len - NLMSG_LENGTH(sizeof(struct rtmsg));
+  int hlen = ((msg->rtm_family == AF_INET) ? 32
+      : ((msg->rtm_family == AF_INET6) ? 128 : -1));
+
+  if (mhdr->nlmsg_type != RTM_NEWRULE) return 0;
+  if (msglen < 0) return 1;
+
+  tvar = msglen;
+  for (rta = RTM_RTA(msg); RTA_OK(rta, tvar); rta=RTA_NEXT(rta, tvar))
+    if (rta->rta_type <= RTA_MAX) attr[rta->rta_type] = rta;
+
+  if (tvar) error_msg("deficit %d, rtalen = %d!", tvar, rta->rta_len);
+
+  printf("%u:\tfrom ", attr[RTA_PRIORITY] ?
+      *(unsigned *)RTA_DATA(attr[RTA_PRIORITY]) : 0);
+
+  if (attr[RTA_SRC]) {
+    printf("%s", (msg->rtm_family == AF_INET || msg->rtm_family == AF_INET6)
+        ? inet_ntop(msg->rtm_family, RTA_DATA(attr[RTA_SRC]),
+          toybuf, sizeof(toybuf))
+        : "???");
+    (msg->rtm_src_len != hlen) ? printf("/%u", msg->rtm_src_len) : 0;
+  } else msg->rtm_src_len ? printf("0/%d", msg->rtm_src_len) : printf("all");
+
+  xputc(' ');
+  if (attr[RTA_DST]) {
+    printf("to %s", (msg->rtm_family == AF_INET || msg->rtm_family == AF_INET6)
+        ? inet_ntop(msg->rtm_family, RTA_DATA(attr[RTA_DST]),
+          toybuf, sizeof(toybuf))  : "???");
+    (msg->rtm_dst_len != hlen) ? printf("/%u", msg->rtm_dst_len) : xputc(' ');
+  } else if (msg->rtm_dst_len)
+    printf("to 0/%d ", msg->rtm_dst_len);
+
+  if (msg->rtm_tos)
+    printf("tos %s ", namefromRPDB(msg->rtm_tos, RPDB_rtdsfield));
+
+  if (attr[RTA_PROTOINFO])
+    printf("fwmark %#x ", *(uint32_t*)RTA_DATA(attr[RTA_PROTOINFO]));
+
+  if (attr[RTA_IIF]) printf("iif %s ", (char*)RTA_DATA(attr[RTA_IIF]));
+
+  if (msg->rtm_table)
+    printf("lookup %s ", namefromRPDB(msg->rtm_table, RPDB_rttables));
+
+  if (attr[RTA_FLOW]) {
+    u_int32_t from, to = *(u_int32_t *)RTA_DATA(attr[RTA_FLOW]);
+    char *format = "realms %s/";
+
+    to = (from = (to >> 16)) & 0xFFFF;
+    format = (from ? format: "%s");
+    printf(format, namefromRPDB((from ? from : to), RPDB_rtrealms));
+  }
+
+  if (msg->rtm_type == RTN_NAT) {
+    if (!attr[RTA_GATEWAY]) printf("masquerade");
+    else printf("map-to %s ", inet_ntop(msg->rtm_family,
+          RTA_DATA(attr[RTA_GATEWAY]), toybuf, sizeof(toybuf)));
+  } else if (msg->rtm_type != RTN_UNICAST)
+    printf("%s", rtmtype_idx2str(msg->rtm_type));
+
+  xputc('\n');
+  return 0;
+}
+
+static int rulelist(char **argv)
+{
+  if (*argv) {
+    error_msg("'ip rule show' does not take any arguments.");
+    return 1;
+  }
+  send_nlmesg(RTM_GETRULE, NLM_F_ROOT | NLM_F_MATCH | NLM_F_REQUEST,
+      ((TT.addressfamily != AF_UNSPEC) ? TT.addressfamily : AF_INET), NULL, 0);
+  return filter_nlmesg(show_rules, argv);
+}
+
+static int iprule(char **argv)
+{
+  int idx;
+  struct arglist options[] = {{"add", 0}, {"delete", 0}, {"list", 1},
+    {"show", 1}, {NULL, -1}};
+  cmdobj ipcmd, cmdobjlist[] = {ruleupdate, rulelist};
+
+  if (!*argv) idx = 1;
+  else if ((idx = substring_to_idx(*argv++, options)) == -1)
+    show_iprule_help();
+  ipcmd = cmdobjlist[idx];
+  return ipcmd(argv);
+}
+//============================================================================
+// TODO: code for ip tunnel.
+//============================================================================
+static int iptunnel(char **argv)
+{
+  printf("__FUNCTION__ = %s\n", __FUNCTION__);
+  return 0;
+}
+
+// ===========================================================================
+// Common code, which is used for all ip options.
+// ===========================================================================
+
+// Parse netlink messages and call input callback handler for action
+static int filter_nlmesg(int (*fun)(struct nlmsghdr *mhdr, char **argv),
+    char **argv)
+{
+  while (1) {
+    struct nlmsghdr *mhdr;
+    int msglen = recv(TT.sockfd, TT.gbuf, MESG_LEN, 0);
+
+    if ((msglen < 0) && (errno == EINTR || errno == EAGAIN)) continue;
+    else if (msglen < 0) {
+      error_msg("netlink receive error %s", strerror(errno));
+      return 1;
+    } else if (!msglen) {
+      error_msg("EOF on netlink");
+      return 1;
+    }
+
+    for (mhdr = (struct nlmsghdr*)TT.gbuf; NLMSG_OK(mhdr, msglen);
+        mhdr = NLMSG_NEXT(mhdr, msglen)) {
+      int err;
+      if (mhdr->nlmsg_pid != getpid())
+        continue;
+      switch (mhdr->nlmsg_type) {
+        case NLMSG_DONE:
+          return 0;
+        case NLMSG_ERROR:
+          {
+            struct nlmsgerr *merr = (struct nlmsgerr*)NLMSG_DATA(mhdr);
+
+            if (merr->error == 0) return 0;
+            if (mhdr->nlmsg_len < NLMSG_LENGTH(sizeof(struct nlmsgerr)))
+              error_msg("ERROR truncated");
+            else {
+              errno = -merr->error;
+              perror_msg("RTNETLINK answers");
+            }
+            return 1;
+          }
+        default:
+          if (fun && (err = fun(mhdr, argv))) return err;
+          break;
+      }
+    } // End of for loop.
+  } // End of while loop.
+  return 0;
+}
+
+void ip_main(void)
+{
+  char **optargv = toys.argv;
+  //TT.addressfamily = AF_UNSPEC;
+  int idx, isip = !(toys.which->name[2]); //1 -> if only ip
+  cmdobj ipcmd, cmdobjlist[] = {ipaddr, iplink, iproute, iprule, iptunnel};
+
+  for (++optargv; *optargv; ++optargv) {
+    char *ptr = *optargv;
+    struct arglist ip_options[] = {{"oneline", 0}, {"family",  1},
+      {"4", 1}, {"6", 1}, {"0", 1}, {"stats", 2}, {NULL, -1}};
+
+    if (*ptr != '-') break;
+    else if ((*(ptr+1) == '-') && (*(ptr+2))) ptr +=2;
+    //escape "--" and stop ip arg parsing.
+    else if ((*(ptr+1) == '-') && (!*(ptr+2))) {
+      *ptr +=1;
+      break;
+    } else ptr +=1;
+    switch (substring_to_idx(ptr, ip_options)) {
+      case 0: TT.singleline = 1;
+              break;
+      case 1: {
+                if (isdigit(*ptr)) {
+                  long num = atolx(ptr);
+                  if (num == 4) TT.addressfamily  = AF_INET;
+                  else if (num == 6) TT.addressfamily  = AF_INET6;
+                  else TT.addressfamily = AF_PACKET;
+                } else {
+                  struct arglist ip_aflist[] = {{"inet", AF_INET},
+                    {"inet6", AF_INET6}, {"link", AF_PACKET}, {NULL, -1}};
+
+                  if (!*++optargv) iphelp();
+                  if ((TT.addressfamily = string_to_idx(*optargv, ip_aflist)) == -1)
+                    error_exit("wrong family '%s'", *optargv);
+                }
+              }
+              break;
+      case 2:
+              TT.stats++;
+              break;
+      default: iphelp();
+               break; // unreachable code.
+    }
+  }
+
+  TT.sockfd = xsocket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
+
+  if (isip) {// only for ip
+    if (*optargv) {
+      struct arglist ip_objectlist[] = { {"address", 0}, {"link", 1},
+        {"route", 2}, {"rule", 3}, {"tunnel", 4}, {"tunl", 4}, {NULL, -1}};
+
+      if ((idx = substring_to_idx(*optargv, ip_objectlist)) == -1) iphelp();
+      ipcmd = cmdobjlist[idx];
+      toys.exitval = ipcmd(++optargv);
+    } else iphelp();
+  } else {
+    struct arglist ip_objectlist[] = { {"ipaddr", 0}, {"iplink", 1},
+      {"iproute", 2}, {"iprule", 3}, {"iptunnel", 4}, {NULL, -1}};
+    if ((idx = string_to_idx(toys.which->name, ip_objectlist)) == -1)
+      iphelp();
+    ipcmd = cmdobjlist[idx];
+    toys.exitval = ipcmd(optargv);
+  }
+  xclose(TT.sockfd);
+}