linux 获取网络状态信息（Rtnetlink）

一、Rtnetlink

　　Rtnetlink 允许对内核路由表进行读和更改，它用于内核与各个子系统之间（路由子系统、IP地址、链接参数等）的通信，

用户空间可以通过NET_LINK_ROUTER socket 与内核进行通信，该过程基于标准的netlink消息进行。

　　注：netlink用法在上一篇博文中有提到 http://www.cnblogs.com/wenqiang/p/6306727.html

一些rtnetlink消息在初始头后有一些可选属性，下面是该属性的结构：

 struct rtattr {
     unsigned short rta_len;    /* Length of option */
     unsigned short rta_type;   /* Type of option */
     /* Data follows */
 };

操作这些属性只可以用RTA_*这些宏来造作

 /* Macros to handle rtattributes */
 
 /* 对齐 */
 #define RTA_ALIGNTO 4
 #define RTA_ALIGN(len) ( ((len)+RTA_ALIGNTO-1) & ~(RTA_ALIGNTO-1) )
 
 /* 判断是否为合法的路由属性 */
 #define RTA_OK(rta,len) ((len) >= (int)sizeof(struct rtattr) && \
              (rta)->rta_len >= sizeof(struct rtattr) && \
              (rta)->rta_len <= (len))
 
 /* 获取下一个rtattr的首地址*/
 #define RTA_NEXT(rta,attrlen)   ((attrlen) -= RTA_ALIGN((rta)->rta_len), \
                  (struct rtattr*)(((char*)(rta)) + RTA_ALIGN((rta)->rta_len)))
 
 /* 返回加上 rtattr header的总长度 */
 #define RTA_LENGTH(len) (RTA_ALIGN(sizeof(struct rtattr)) + (len))
 
 /* 返回数据对齐的最小值 */
 #define RTA_SPACE(len)  RTA_ALIGN(RTA_LENGTH(len))
 
 /* 返回属性数据部分首地址 */
 #define RTA_DATA(rta)   ((void*)(((char*)(rta)) + RTA_LENGTH(0)))
 
 /*返回属性数据部分的长度 */
 #define RTA_PAYLOAD(rta) ((int)((rta)->rta_len) - RTA_LENGTH(0))
 
 /*****************************************************************
 ******************************************************************/
 rtnetlink_socket = socket(AF_NETLINK, int socket_type, NETLINK_ROUTE);
 
 int RTA_OK(struct rtattr *rta, int rtabuflen);
 
 void *RTA_DATA(struct rtattr *rta);
 
 unsigned int RTA_PAYLOAD(struct rtattr *rta);
 
 struct rtattr *RTA_NEXT(struct rtattr *rta, unsigned int rtabuflen);
 
 unsigned int RTA_LENGTH(unsigned int length);
 
 unsigned int RTA_SPACE(unsigned int length);

Rtnetlink 由下面这些消息类型构成（新加在标准的netlink消息上）
（1）#RTM_NEWLINK, RTM_DELLINK, RTM_GETLINK
创建或者删除一个特定的网络接口，或者从一个特定的网络接口上获得信息。
这些消息含有一个ifinfomsg类型的结构，紧跟在后面的是一系列的rtattr结构。

 /*****************************************************************
  *      Link layer specific messages.
  ****/
 
 /* struct ifinfomsg
  * passes link level specific information, not dependent
  * on network protocol.
  */
 
  struct ifinfomsg {
     unsigned char  ifi_family; /* AF_UNSPEC */
     unsigned short ifi_type;   /* Device type */
     int            ifi_index;  /* Interface index */
     unsigned int   ifi_flags;  /* Device flags  */
     unsigned int   ifi_change; /* change mask */
 };
 /*
 * ifi_family: 接口地址类型
 * ifi_type: 设备类型
 * ifi_index: 是结构唯一的索引
 * ifi_flags:  设备标志，可以看netdevice 结构
 * ifi_change: 保留值，通常设置为0xFFFFFFFF
 */
 
 /*
 ifi_type代表硬件设备的类型：
     ARPHRD_ETHER                   10M以太网
     ARPHRD_PPP                     PPP拨号
     ARPHRDLOOPBACK                 环路设备
 
 ifi_flags包含设备的一些标志：
     IFF_UP                            接口正在运行
     IFF_BROADCAST                     有效的广播地址集
     IFF_DEBUG                         内部调试标志
     IFF_LOOPBACK                      这是自环接口
     IFF_POINTOPOINT                   这是点到点的链路设备
     IFF_RUNNING                       资源已分配
     IFF_NOARP                         无arp协议，没有设置第二层目的地址
     IFF_PROMISC                       接口为杂凑(promiscuous)模式
     IFF_NOTRAILERS                    避免使用trailer
     IFF_ALLMULTI                      接收所有组播(multicast)报文
     IFF_MASTER                        主负载平衡群(bundle)
     IFF_SLAVE                         从负载平衡群(bundle)
     IFF_MULTICAST                     支持组播(multicast)
     IFF_PORTSEL                       可以通过ifmap选择介质(media)类型
     IFF_AUTOMEDIA                     自动选择介质
     IFF_DYNAMIC                       接口关闭时丢弃地址    
 
             Routing attributes（rtattr部分属性，rta_type）
 
 rta_type         value type         description
  ──────────────────────────────────────────────────────────
 IFLA_UNSPEC      -                  未说明，未指定的数据
 IFLA_ADDRESS     hardware address   L2硬件地址
 IFLA_BROADCAST   hardware address   L2广播地址.
 IFLA_IFNAME      asciiz string      char型设备名.
 IFLA_MTU         unsigned int       MTU of the device.
 IFLA_LINK        int                Link type.
 IFLA_QDISC       asciiz string      Queueing discipline.
 IFLA_STATS       see below          struct rtnl_link_stats的设备信息
 
 //用来获取ifinfomsg后面的rtattr结构
 #define IFLA_RTA(r)  ((struct rtattr*)(((char*)(r)) + NLMSG_ALIGN(sizeof(struct ifinfomsg))))
 */

（2）# RTM_NEWADDR, RTM_DELADDR, RTM_GETADDR

添加,删除或者接收一个和接口相关的IP地址的信息。
在linux2.2中，一个网口是可以有多个IP地址信息的。这些消息含有一个ifaddrmsg类型的结构，紧跟在后面的是一系列的rtattr结构。

  struct ifaddrmsg {
     unsigned char ifa_family;    /* Address type */
     unsigned char ifa_prefixlen; /* Prefixlength of address */
     unsigned char ifa_flags;     /* Address flags */
     unsigned char ifa_scope;     /* Address scope */
     int           ifa_index;     /* Interface index */
   };
 /*
 * ifa_family: 地址类型（通常为AF_INET or AF_INET6)）
 * ifa_prefixlen: 地址的地址掩码长度，如果改地址定义在这个family
 * ifa_flags:
 * ifa_scope: 地址的作用域
 * ifa_index:  接口索引与接口地址关联
 */
 
  /*
             Attributes （rtattr部分属性，rta_type）
 rta_type        value type             description
 ─────────────────────────────────────────────────────────────
 IFA_UNSPEC      -                      unspecified.
 IFA_ADDRESS     raw protocol address   接口地址 interface address
 IFA_LOCAL       raw protocol address   本地地址 local address
 IFA_LABEL       asciiz string          接口名称 name of the interface
 IFA_BROADCAST   raw protocol address   广播 broadcast address.
 IFA_ANYCAST     raw protocol address   anycast address
 IFA_CACHEINFO   struct ifa_cacheinfo   Address information.
 
 */

（3）#RTM_NEWROUTE, RTM_DELROUTE, RTM_GETROUTE

创建，删除或者获取网络设备的路由信息；这些消息包含一个rtmsg结构，其后跟数目可选的rtattr结构。
对于RTM_GETROUTE，设置rtm_dst_len以及rtm_src_len为0表示获取指定路由表的所有条目（entries）。
其它的成员，除了rtm_table、rtm_protocol，0是通配符

 struct rtmsg {
     unsigned char       rtm_family;
     unsigned char       rtm_dst_len;
     unsigned char       rtm_src_len;
     unsigned char       rtm_tos;
 
     unsigned char       rtm_table;  /* Routing table id */
     unsigned char       rtm_protocol;   /* Routing protocol; see below  */
     unsigned char       rtm_scope;  /* See below */
     unsigned char       rtm_type;   /* See below    */
 
     unsigned        rtm_flags;
 };
 
 rtm_type          Route type
 ───────────────────────────────────────────────────────────
 RTN_UNSPEC        unknown route                /*位置路由*/
 RTN_UNICAST       a gateway or direct route  /* 网关或直连路由 */
 RTN_LOCAL         a local interface route    /* 本地接口路由 */
 RTN_BROADCAST     a local broadcast route (sent as a broadcast)  /* 本地广播式接收，发送 */
 RTN_ANYCAST       a local broadcast route (sent as a unicast)    /* 本地单播路由 */
 RTN_MULTICAST     a multicast route            /* 多播路由 */
 RTN_BLACKHOLE     a packet dropping route    /* 丢弃 */
 RTN_UNREACHABLE   an unreachable destination /* 目标不可达 */
 RTN_PROHIBIT      a packet rejection route    /* 拒绝 */
 RTN_THROW         continue routing lookup in another table /* 不在本表 */
 RTN_NAT           a network address translation rule        /* nat */
 RTN_XRESOLVE      refer to an external resolver (not implemented)
 
 rtm_protocol      Route origin.
 ───────────────────────────────────────
 RTPROT_UNSPEC     unknown
 RTPROT_REDIRECT   by an ICMP redirect (currently unused)   /* 通过icmp转发建立路由 （目前没用）*/
 RTPROT_KERNEL     by the kernel     /* 通过内核建立路由 */
 RTPROT_BOOT       during boot       /* 启动时建立路由 */
 RTPROT_STATIC     by the administrator  /* 管理员建立 */
 
 rtm_scope is the distance to the destination:
 
 RT_SCOPE_UNIVERSE   global route
 RT_SCOPE_SITE       interior route in the local autonomous system
 RT_SCOPE_LINK       route on this link
 RT_SCOPE_HOST       route on the local host
 RT_SCOPE_NOWHERE    destination doesn't exist 
 
 /* 用户可用范围 */
 RT_SCOPE_UNIVERSE ~ RT_SCOPE_SITE are available to the user.
 
 The rtm_flags have the following meanings:
 
 RTM_F_NOTIFY     if the route changes, notify the user via rtnetlink
 RTM_F_CLONED     route is cloned from another route
 RTM_F_EQUALIZE   a multipath equalizer (not yet implemented)
 
 rtm_table specifies the routing table
 
 RT_TABLE_UNSPEC    an unspecified routing table /* 0 未指定的表 */
 RT_TABLE_DEFAULT   the default table            /* 253 默认表 */
 RT_TABLE_MAIN      the main table                /* 254 main 表 */
 RT_TABLE_LOCAL     the local table                /* 255 local 表 */
 
 //用户可以使用 RT_TABLE_UNSPEC 到 RT_TABLE_DEFAULT 之间的任意值
 
             Attributes
 
 rta_type        value type         description
 ──────────────────────────────────────────────────────────────
 RTA_UNSPEC      -                  ignored.
 RTA_DST         protocol address   Route destination address.   /* 目的 */
 RTA_SRC         protocol address   Route source address.        /* 源地址 */
 RTA_IIF         int                Input interface index.        /* 输入设备 index */
 RTA_OIF         int                Output interface index.
 RTA_GATEWAY     protocol address   The gateway of the route        /* 网关 */
 RTA_PRIORITY    int                Priority of route.            /* 优先级 */
 RTA_PREFSRC
 RTA_METRICS     int                Route metric                    /* 路由metric 值*/
 RTA_MULTIPATH
 RTA_PROTOINFO
 RTA_FLOW
 RTA_CACHEINFO

一面是一个具体实例：

 /*********************************************************
 * Filename: nl_netinfo.c
 * Author: zhangwj
 * Date:
 * Descripte:
 * Email:
 * Warnning:
 **********************************************************/
 #include <stdio.h>
 #include <string.h>
 #include <stdlib.h>
 #include <sys/types.h>          /* See NOTES */
 #include <sys/socket.h>
 #include <arpa/inet.h>
 #include <sys/epoll.h>
 #include <linux/netlink.h>
 #include <linux/rtnetlink.h>
 #include <linux/route.h>
 #include <errno.h>
 
 #define EPOLL_LISTEN_MAX_CNT    256
 #define EPOLL_LISTEN_TIMEOUT    500
 
 int g_nlfd = -;
 int g_epollfd = -;
 
 void parse_rtattr(struct rtattr **tb, int max, struct rtattr *attr, int len)
 {
     for (; RTA_OK(attr, len); attr = RTA_NEXT(attr, len)) {
         if (attr->rta_type <= max) {
             tb[attr->rta_type] = attr;
         }
     }
 }
 
 void nl_netroute_handle(struct nlmsghdr *nlh)
 {
     int len;
     struct rtattr *tb[RTA_MAX + ];
     struct rtmsg *rt;
     char tmp[];  
 
     bzero(tb, sizeof(tb));
     rt = NLMSG_DATA(nlh);
     len = nlh->nlmsg_len - NLMSG_SPACE(sizeof(*rt));
     parse_rtattr(tb, RTA_MAX, RTM_RTA(rt), len);
     printf("%s: ", (nlh->nlmsg_type==RTM_NEWROUTE)?"NEWROUT":"DELROUT");
     if (tb[RTA_DST] != NULL) {
         inet_ntop(rt->rtm_family, RTA_DATA(tb[RTA_DST]), tmp, sizeof(tmp));
         printf("DST: %s ", tmp);
     }
     if (tb[RTA_SRC] != NULL) {
         inet_ntop(rt->rtm_family, RTA_DATA(tb[RTA_SRC]), tmp, sizeof(tmp));
         printf("SRC: %s ", tmp);
     }
     if (tb[RTA_GATEWAY] != NULL) {
         inet_ntop(rt->rtm_family, RTA_DATA(tb[RTA_GATEWAY]), tmp, sizeof(tmp));
         printf("GATEWAY: %s ", tmp);
     }
     printf("\n");
 }
 
 void nl_netifinfo_handle(struct nlmsghdr *nlh)
 {
     int len;
     struct rtattr *tb[IFLA_MAX + ];
     struct ifinfomsg *ifinfo; 
 
     bzero(tb, sizeof(tb));
     ifinfo = NLMSG_DATA(nlh);
     len = nlh->nlmsg_len - NLMSG_SPACE(sizeof(*ifinfo));
     parse_rtattr(tb, IFLA_MAX, IFLA_RTA (ifinfo), len);  
 
     printf("%s: %s ", (nlh->nlmsg_type==RTM_NEWLINK) ? "NEWLINK" : "DELLINK", (ifinfo->ifi_flags & IFF_UP) ? "up" : "down");
     if(tb[IFLA_IFNAME]) {
         printf("%s", RTA_DATA(tb[IFLA_IFNAME]));
     }
     printf("\n");
 }
 
 void nl_netifaddr_handle(struct nlmsghdr *nlh)
 {
     int len;
     struct rtattr *tb[IFA_MAX + ];
     struct ifaddrmsg *ifaddr;
     char tmp[];  
 
     bzero(tb, sizeof(tb));
     ifaddr = NLMSG_DATA(nlh);
     len =nlh->nlmsg_len - NLMSG_SPACE(sizeof(*ifaddr));
     parse_rtattr(tb, IFA_MAX, IFA_RTA (ifaddr), len);
 
     printf("%s ", (nlh->nlmsg_type == RTM_NEWADDR)? "NEWADDR":"DELADDR");
     if (tb[IFA_LABEL] != NULL) {
         printf("%s ", RTA_DATA(tb[IFA_LABEL]));
     }
     if (tb[IFA_ADDRESS] != NULL) {
         inet_ntop(ifaddr->ifa_family, RTA_DATA(tb[IFA_ADDRESS]), tmp, sizeof(tmp));
         printf("%s ", tmp);
     }
     printf("\n");
 }
 
 void nl_netlink_handle(int fd)
 {
     int r_size;
     socklen_t len = ;
     char buff[] = {};
     struct sockaddr_nl addr;
     struct nlmsghdr *nlh;
 
     while()
     {
         len = sizeof(addr);
         r_size = recvfrom(fd, (void *)buff, sizeof(buff), , (struct sockaddr *)&addr, &len);
         nlh = (struct nlmsghdr *)buff;
         for(; NLMSG_OK(nlh, r_size); nlh = NLMSG_NEXT(nlh, r_size))
         {
             switch(nlh->nlmsg_type) {
             case NLMSG_DONE:
             case NLMSG_ERROR:
                 break;
             case RTM_NEWLINK:     /* */
             case RTM_DELLINK:
                 nl_netifinfo_handle(nlh);
                 break;
             case RTM_NEWADDR:
             case RTM_DELADDR:    /* */
                 nl_netifaddr_handle(nlh);
                 break;
             case RTM_NEWROUTE:
             case RTM_DELROUTE:   /* */
                 nl_netroute_handle(nlh);
                 break;
             default:
                 break;
             }
         }
     }
 }
 
 void epoll_event_handle(void)
 {
     int i = ;
     int fd_cnt = ;
     int sfd;
     struct epoll_event events[EPOLL_LISTEN_MAX_CNT];    
 
     memset(events, , sizeof(events));
     while()
     {
         /* wait epoll event */
         fd_cnt = epoll_wait(g_epollfd, events, EPOLL_LISTEN_MAX_CNT, EPOLL_LISTEN_TIMEOUT);
         for(i = ; i < fd_cnt; i++)
         {
             sfd = events[i].data.fd;
             if(events[i].events & EPOLLIN)
             {
                 if (sfd == g_nlfd)
                 {
                     nl_netlink_handle(sfd);
                 }
             }
         }
     }
 }
 
 int epoll_add_fd(int fd)
 {
     struct epoll_event ev;
 
     ev.data.fd = fd;
     ev.events = EPOLLIN | EPOLLET;
 
     if (epoll_ctl(g_epollfd, EPOLL_CTL_ADD, fd, &ev) < ) {
         perror("epoll add fd error");
         return -;
     }
 
     printf("epoll add fd[%d] success\n", fd);
     return ;
 }
 
 int init_epoll_fd()
 {
     int epollfd = -;
 
     epollfd = epoll_create(EPOLL_LISTEN_MAX_CNT);
     if (epollfd < ) {
         perror("epoll create failure!...");
         return -;
     }
     g_epollfd = epollfd;
 
     printf("epool create fd [%d] success\n", epollfd);
     return g_epollfd;
 }
 
 int init_nl_sockfd()
 {
     int ret = ;
     int nlfd = -;
     struct sockaddr_nl sa;
 
     /* open a netlink fd */
     nlfd = socket(AF_NETLINK, SOCK_RAW, NETLINK_ROUTE);
     if (nlfd < ) {
         perror("create netlink socket failure");
         return -;
     }
 
     memset(&sa, , sizeof(sa));
     sa.nl_family = AF_NETLINK;
     sa.nl_groups = RTMGRP_LINK | RTMGRP_IPV4_IFADDR | RTMGRP_IPV4_ROUTE | RTMGRP_IPV6_IFADDR | RTMGRP_IPV6_ROUTE;
 
     /* bind netlink */
     ret = bind(nlfd, (struct sockaddr *)&sa, sizeof(sa));
     if (ret < ) {
         perror("bind nlfd error");
         close(nlfd);
         return -;
     }
 
     if (epoll_add_fd(nlfd)) {
         close(nlfd);
         return -;
     }
     g_nlfd = nlfd;
 
     printf("netlink create fd [%d] success\n", nlfd);
     return nlfd;
 }
 
 int main(int argc, char **argv)
 {
     if (init_epoll_fd() < ) { /* 创建epoll 监听fd */
         return -;
     }
 
     if (init_nl_sockfd() < ) { /* 创建netlink */
         return -;
     }
 
     /* 循环接收处理 */
     epoll_event_handle();
 
     return ;
 }

参考资料：

http://www.man7.org/linux/man-pages/man7/rtnetlink.7.html

http://www.man7.org/linux/man-pages/man3/rtnetlink.3.html