在内核,网桥是以模块的方式存在,注册源码路径:\net\brige\br.c:
4.1 初始化
static int __init br_init(void)
{
br_fdb_init(); //网桥数据库初始化,分配slab缓冲区
#ifdef CONFIG_BRIDGE_NETFILTER
if (br_netfilter_init()) //netfilter钩子初始化
return 1;
#endif
brioctl_set(br_ioctl_deviceless_stub); //设置ioctl钩子函数:br_ioctl_hook
br_handle_frame_hook = br_handle_frame;//设置报文处理钩子:br_ioctl_hook
//网桥数据库处理钩子
br_fdb_get_hook = br_fdb_get;
br_fdb_put_hook = br_fdb_put;
//在netdev_chain通知链表上注册
register_netdevice_notifier(&br_device_notifier);
return 0;
}
|
4.2 新建网桥
前面说到通过brctl addbr br0命令建立网桥,此处用户控件调用的brctl命令最终对应到内核中的br_ioctl_deviceless_stub处理函数:
int br_ioctl_deviceless_stub(unsigned int cmd, void __user *uarg)
{
switch (cmd) {
case SIOCGIFBR:
case SIOCSIFBR:
return old_deviceless(uarg);
case SIOCBRADDBR: //新建网桥
case SIOCBRDELBR: //删除网桥
{
char buf[IFNAMSIZ];
if (!capable(CAP_NET_ADMIN))
return -EPERM;
//copy_from_user:把用户空间的数据拷入内核空间
if (copy_from_user(buf, uarg, IFNAMSIZ))
return -EFAULT;
buf[IFNAMSIZ-1] = 0;
if (cmd == SIOCBRADDBR)
return br_add_bridge(buf);
return br_del_bridge(buf);
}
}
return -EOPNOTSUPP;
}
|
在这里,我们传入的cmd为SIOCBRADDBR.转入br_add_bridge(buf)中进行:
int br_add_bridge(const char *name)
{
struct net_device *dev;
int ret;
//为虚拟桥新建一个net_device
dev = new_bridge_dev(name);
if (!dev)
return -ENOMEM;
rtnl_lock();
//由内核确定接口名字,例如eth0 eth1等
if (strchr(dev->name, '%')) {
ret = dev_alloc_name(dev, dev->name);
if (ret < 0)
goto err1;
}
//向内核注册此网络设备
ret = register_netdevice(dev);
if (ret)
goto err2;
/* network device kobject is not setup until
* after rtnl_unlock does it's hotplug magic.
* so hold reference to avoid race.
*/
dev_hold(dev);
rtnl_unlock();
//在sysfs中建立相关信息
ret = br_sysfs_addbr(dev);
dev_put(dev);
if (ret)
unregister_netdev(dev);
out:
return ret;
err2:
free_netdev(dev);
err1:
rtnl_unlock();
goto out;
}
|
网桥是一个虚拟的设备,它的注册跟实际的物理网络设备注册是一样的。我们关心的是网桥对应的net_device结构是什么样的,继续跟踪进new_bridge_dev:
static struct net_device *new_bridge_dev(const char *name)
{
struct net_bridge *br;
struct net_device *dev;
//分配net_device
dev = alloc_netdev(sizeof(struct net_bridge), name,
br_dev_setup);
if (!dev)
return NULL;
//网桥的私区结构为net_bridge
br = netdev_priv(dev);
//私区结构中的dev字段指向设备本身
br->dev = dev;
spin_lock_init(&br->lock);
//队列初始化。在port_list中保存了这个桥上的端口列表
INIT_LIST_HEAD(&br->port_list);
spin_lock_init(&br->hash_lock);
//下面这部份代码跟stp协议相关,我们暂不关心
br->bridge_id.prio[0] = 0x80;
br->bridge_id.prio[1] = 0x00;
memset(br->bridge_id.addr, 0, ETH_ALEN);
br->stp_enabled = 0;
br->designated_root = br->bridge_id;
br->root_path_cost = 0;
br->root_port = 0;
br->bridge_max_age = br->max_age = 20 * HZ;
br->bridge_hello_time = br->hello_time = 2 * HZ;
br->bridge_forward_delay = br->forward_delay = 15 * HZ;
br->topology_change = 0;
br->topology_change_detected = 0;
br->ageing_time = 300 * HZ;
INIT_LIST_HEAD(&br->age_list);
br_stp_timer_init(br);
return dev;
}
|
在br_dev_setup中还做了一些另外在函数指针初始化:
void br_dev_setup(struct net_device *dev)
{
//将桥的MAC地址设为零
memset(dev->dev_addr, 0, ETH_ALEN);
//初始化dev的部分函数指针,因为目前网桥设备主适用于以及网,
//以太网的部分功能对它也适用
ether_setup(dev);
//设置设备的ioctl函数为br_dev_ioctl
dev->do_ioctl = br_dev_ioctl;
//网桥与一般网卡不同,网桥统一统计它的数据包和字节数等信息
dev->get_stats = br_dev_get_stats;
// 网桥接口的数据包发送函数,真实设备要向外发送数据时,是通过网卡向外发送数据
// 而该网桥设备要向外发送数据时,它的处理逻辑与网桥其它接口的基本一致。
dev->hard_start_xmit = br_dev_xmit;
dev->open = br_dev_open;
dev->set_multicast_list = br_dev_set_multicast_list;
dev->change_mtu = br_change_mtu;
dev->destructor = free_netdev;
SET_MODULE_OWNER(dev);
dev->stop = br_dev_stop;
dev->tx_queue_len = 0;
dev->set_mac_address = NULL;
dev->priv_flags = IFF_EBRIDGE;
}
|
仅仅创建网桥,还是不够的。实际应用中的网桥需要添加实际的端口(即物理接口),如例子中的eth1, eth2等。应用程序在使用ioctl来为网桥增加物理接口,对应内核函数br_dev_ioctl的代码和分析如下:
int br_dev_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct net_bridge *br = netdev_priv(dev);
switch(cmd) {
case SIOCDEVPRIVATE:
return old_dev_ioctl(dev, rq, cmd);
case SIOCBRADDIF: //添加
case SIOCBRDELIF: //删除
//同一处理函数,默认为添加
return add_del_if(br, rq->ifr_ifindex, cmd == SIOCBRADDIF);
}
pr_debug("Bridge does not support ioctl 0x%x\n", cmd);
return -EOPNOTSUPP;
}
|
下面分析具体的添加删除函数add_del_if:
static int add_del_if(struct net_bridge *br, int ifindex, int isadd)
{
struct net_device *dev;
int ret;
if (!capable(CAP_NET_ADMIN))
return -EPERM;
dev = dev_get_by_index(ifindex);
if (dev == NULL)
return -EINVAL;
if (isadd)
ret = br_add_if(br, dev);
else
ret = br_del_if(br, dev);
dev_put(dev);
return ret;
}
|
对应的添加删除函数分别为:br_add_if, br_del_if;
br_add_if:int br_add_if(struct net_bridge *br, struct net_device *dev)
{
struct net_bridge_port *p;
int err = 0;
/*--Kernel仅支持以太网网桥--*/
if (dev->flags & IFF_LOOPBACK || dev->type != ARPHRD_ETHER)
return -EINVAL;
/*--把网桥接口当作物理接口加入到另一个网桥中,是不行的,
逻辑和代码上都会出现 loop--*/
if (dev->hard_start_xmit == br_dev_xmit)
return -ELOOP;
/*--该物理接口已经绑定到另一个网桥了--*/
if (dev->br_port != NULL)
return -EBUSY;
/*--为该接口创建一个网桥端口数据,并初始化好该端口的相关数据--*/
if (IS_ERR(p = new_nbp(br, dev, br_initial_port_cost(dev))))
return PTR_ERR(p);
/*--将该接口的物理地址写入到 MAC-端口映射表中,
该MAC是属于网桥内部端口的固定MAC地址,
它在fdb中的记录是固定的,不会失效(agged)--*/
if ((err = br_fdb_insert(br, p, dev->dev_addr)))
destroy_nbp(p);
/*--添加相应的系统文件信息--*/
else if ((err = br_sysfs_addif(p)))
del_nbp(p);
else {
/*--打开该接口的混杂模式,网桥中的各个端口必须处于混杂模式,
网桥才能正确工作--*/
dev_set_promiscuity(dev, 1);
/*--加到端口列表--*/
list_add_rcu(&p->list, &br->port_list);
/*--STP相关设置-*/
spin_lock_bh(&br->lock);
br_stp_recalculate_bridge_id(br);
br_features_recompute(br);
if ((br->dev->flags & IFF_UP)
&& (dev->flags & IFF_UP) && netif_carrier_ok(dev))
br_stp_enable_port(p);
spin_unlock_bh(&br->lock);
/*--设置设备的mtu--*/
dev_set_mtu(br->dev, br_min_mtu(br));
}
return err;
}
|
br_del_if:int br_del_if(struct net_bridge *br, struct net_device *dev)
{
struct net_bridge_port *p = dev->br_port;
if (!p || p->br != br)
return -EINVAL;
br_sysfs_removeif(p);
del_nbp(p);
spin_lock_bh(&br->lock);
br_stp_recalculate_bridge_id(br);
br_features_recompute(br);
spin_unlock_bh(&br->lock);
return 0;
}
|
在内核,网桥是以模块的方式存在,注册源码路径:\net\brige\br.c:
4.1 初始化
static int __init br_init(void)
{
br_fdb_init(); //网桥数据库初始化,分配slab缓冲区
#ifdef CONFIG_BRIDGE_NETFILTER
if (br_netfilter_init()) //netfilter钩子初始化
return 1;
#endif
brioctl_set(br_ioctl_deviceless_stub); //设置ioctl钩子函数:br_ioctl_hook
br_handle_frame_hook = br_handle_frame;//设置报文处理钩子:br_ioctl_hook
//网桥数据库处理钩子
br_fdb_get_hook = br_fdb_get;
br_fdb_put_hook = br_fdb_put;
//在netdev_chain通知链表上注册
register_netdevice_notifier(&br_device_notifier);
return 0;
}
|
4.2 新建网桥
前面说到通过brctl addbr br0命令建立网桥,此处用户控件调用的brctl命令最终对应到内核中的br_ioctl_deviceless_stub处理函数:
int br_ioctl_deviceless_stub(unsigned int cmd, void __user *uarg)
{
switch (cmd) {
case SIOCGIFBR:
case SIOCSIFBR:
return old_deviceless(uarg);
case SIOCBRADDBR: //新建网桥
case SIOCBRDELBR: //删除网桥
{
char buf[IFNAMSIZ];
if (!capable(CAP_NET_ADMIN))
return -EPERM;
//copy_from_user:把用户空间的数据拷入内核空间
if (copy_from_user(buf, uarg, IFNAMSIZ))
return -EFAULT;
buf[IFNAMSIZ-1] = 0;
if (cmd == SIOCBRADDBR)
return br_add_bridge(buf);
return br_del_bridge(buf);
}
}
return -EOPNOTSUPP;
}
|
在这里,我们传入的cmd为SIOCBRADDBR.转入br_add_bridge(buf)中进行:
int br_add_bridge(const char *name)
{
struct net_device *dev;
int ret;
//为虚拟桥新建一个net_device
dev = new_bridge_dev(name);
if (!dev)
return -ENOMEM;
rtnl_lock();
//由内核确定接口名字,例如eth0 eth1等
if (strchr(dev->name, '%')) {
ret = dev_alloc_name(dev, dev->name);
if (ret < 0)
goto err1;
}
//向内核注册此网络设备
ret = register_netdevice(dev);
if (ret)
goto err2;
/* network device kobject is not setup until
* after rtnl_unlock does it's hotplug magic.
* so hold reference to avoid race.
*/
dev_hold(dev);
rtnl_unlock();
//在sysfs中建立相关信息
ret = br_sysfs_addbr(dev);
dev_put(dev);
if (ret)
unregister_netdev(dev);
out:
return ret;
err2:
free_netdev(dev);
err1:
rtnl_unlock();
goto out;
}
|
网桥是一个虚拟的设备,它的注册跟实际的物理网络设备注册是一样的。我们关心的是网桥对应的net_device结构是什么样的,继续跟踪进new_bridge_dev:
static struct net_device *new_bridge_dev(const char *name)
{
struct net_bridge *br;
struct net_device *dev;
//分配net_device
dev = alloc_netdev(sizeof(struct net_bridge), name,
br_dev_setup);
if (!dev)
return NULL;
//网桥的私区结构为net_bridge
br = netdev_priv(dev);
//私区结构中的dev字段指向设备本身
br->dev = dev;
spin_lock_init(&br->lock);
//队列初始化。在port_list中保存了这个桥上的端口列表
INIT_LIST_HEAD(&br->port_list);
spin_lock_init(&br->hash_lock);
//下面这部份代码跟stp协议相关,我们暂不关心
br->bridge_id.prio[0] = 0x80;
br->bridge_id.prio[1] = 0x00;
memset(br->bridge_id.addr, 0, ETH_ALEN);
br->stp_enabled = 0;
br->designated_root = br->bridge_id;
br->root_path_cost = 0;
br->root_port = 0;
br->bridge_max_age = br->max_age = 20 * HZ;
br->bridge_hello_time = br->hello_time = 2 * HZ;
br->bridge_forward_delay = br->forward_delay = 15 * HZ;
br->topology_change = 0;
br->topology_change_detected = 0;
br->ageing_time = 300 * HZ;
INIT_LIST_HEAD(&br->age_list);
br_stp_timer_init(br);
return dev;
}
|
在br_dev_setup中还做了一些另外在函数指针初始化:
void br_dev_setup(struct net_device *dev)
{
//将桥的MAC地址设为零
memset(dev->dev_addr, 0, ETH_ALEN);
//初始化dev的部分函数指针,因为目前网桥设备主适用于以及网,
//以太网的部分功能对它也适用
ether_setup(dev);
//设置设备的ioctl函数为br_dev_ioctl
dev->do_ioctl = br_dev_ioctl;
//网桥与一般网卡不同,网桥统一统计它的数据包和字节数等信息
dev->get_stats = br_dev_get_stats;
// 网桥接口的数据包发送函数,真实设备要向外发送数据时,是通过网卡向外发送数据
// 而该网桥设备要向外发送数据时,它的处理逻辑与网桥其它接口的基本一致。
dev->hard_start_xmit = br_dev_xmit;
dev->open = br_dev_open;
dev->set_multicast_list = br_dev_set_multicast_list;
dev->change_mtu = br_change_mtu;
dev->destructor = free_netdev;
SET_MODULE_OWNER(dev);
dev->stop = br_dev_stop;
dev->tx_queue_len = 0;
dev->set_mac_address = NULL;
dev->priv_flags = IFF_EBRIDGE;
}
|
仅仅创建网桥,还是不够的。实际应用中的网桥需要添加实际的端口(即物理接口),如例子中的eth1, eth2等。应用程序在使用ioctl来为网桥增加物理接口,对应内核函数br_dev_ioctl的代码和分析如下:
int br_dev_ioctl(struct net_device *dev, struct ifreq *rq, int cmd)
{
struct net_bridge *br = netdev_priv(dev);
switch(cmd) {
case SIOCDEVPRIVATE:
return old_dev_ioctl(dev, rq, cmd);
case SIOCBRADDIF: //添加
case SIOCBRDELIF: //删除
//同一处理函数,默认为添加
return add_del_if(br, rq->ifr_ifindex, cmd == SIOCBRADDIF);
}
pr_debug("Bridge does not support ioctl 0x%x\n", cmd);
return -EOPNOTSUPP;
}
|
下面分析具体的添加删除函数add_del_if:
static int add_del_if(struct net_bridge *br, int ifindex, int isadd)
{
struct net_device *dev;
int ret;
if (!capable(CAP_NET_ADMIN))
return -EPERM;
dev = dev_get_by_index(ifindex);
if (dev == NULL)
return -EINVAL;
if (isadd)
ret = br_add_if(br, dev);
else
ret = br_del_if(br, dev);
dev_put(dev);
return ret;
}
|
对应的添加删除函数分别为:br_add_if, br_del_if;
br_add_if:int br_add_if(struct net_bridge *br, struct net_device *dev)
{
struct net_bridge_port *p;
int err = 0;
/*--Kernel仅支持以太网网桥--*/
if (dev->flags & IFF_LOOPBACK || dev->type != ARPHRD_ETHER)
return -EINVAL;
/*--把网桥接口当作物理接口加入到另一个网桥中,是不行的,
逻辑和代码上都会出现 loop--*/
if (dev->hard_start_xmit == br_dev_xmit)
return -ELOOP;
/*--该物理接口已经绑定到另一个网桥了--*/
if (dev->br_port != NULL)
return -EBUSY;
/*--为该接口创建一个网桥端口数据,并初始化好该端口的相关数据--*/
if (IS_ERR(p = new_nbp(br, dev, br_initial_port_cost(dev))))
return PTR_ERR(p);
/*--将该接口的物理地址写入到 MAC-端口映射表中,
该MAC是属于网桥内部端口的固定MAC地址,
它在fdb中的记录是固定的,不会失效(agged)--*/
if ((err = br_fdb_insert(br, p, dev->dev_addr)))
destroy_nbp(p);
/*--添加相应的系统文件信息--*/
else if ((err = br_sysfs_addif(p)))
del_nbp(p);
else {
/*--打开该接口的混杂模式,网桥中的各个端口必须处于混杂模式,
网桥才能正确工作--*/
dev_set_promiscuity(dev, 1);
/*--加到端口列表--*/
list_add_rcu(&p->list, &br->port_list);
/*--STP相关设置-*/
spin_lock_bh(&br->lock);
br_stp_recalculate_bridge_id(br);
br_features_recompute(br);
if ((br->dev->flags & IFF_UP)
&& (dev->flags & IFF_UP) && netif_carrier_ok(dev))
br_stp_enable_port(p);
spin_unlock_bh(&br->lock);
/*--设置设备的mtu--*/
dev_set_mtu(br->dev, br_min_mtu(br));
}
return err;
}
|
br_del_if:int br_del_if(struct net_bridge *br, struct net_device *dev)
{
struct net_bridge_port *p = dev->br_port;
if (!p || p->br != br)
return -EINVAL;
br_sysfs_removeif(p);
del_nbp(p);
spin_lock_bh(&br->lock);
br_stp_recalculate_bridge_id(br);
br_features_recompute(br);
spin_unlock_bh(&br->lock);
return 0;
}
|