在內核,網橋是以模塊的方式存在,註冊源碼路徑:\net\brige\br.c:
4.1 初始化
static int __init br_init(void) {
br_fdb_init(); //網橋數據庫初始化,分配slab緩衝區
#ifdef CONFIG_BRIDGE_NETFILTER
if (br_netfilter_init()) //netfilter鉤子初始化
return 1; #endif
brioctl_set(br_ioctl_deviceless_stub); //設置ioctl鉤子函數:br_ioctl_hook
br_handle_frame_hook = br_handle_frame;//設置報文處理鉤子:br_ioctl_hook
//網橋數據庫處理鉤子
br_fdb_get_hook = br_fdb_get;
br_fdb_put_hook = br_fdb_put;
//在netdev_chain通知鏈表上註冊
register_netdevice_notifier(&br_device_notifier);
return 0; }
|
4.2 新建網橋
前面說到通過brctl addbr br0命令建立網橋,此處用戶控件調用的brctl命令最終對應到內核中的br_ioctl_deviceless_stub處理函數:
int br_ioctl_deviceless_stub(unsigned int cmd, void __user *uarg) {
switch (cmd) {
case SIOCGIFBR:
case SIOCSIFBR:
return old_deviceless(uarg);
case SIOCBRADDBR: //新建網橋
case SIOCBRDELBR: //刪除網橋
{
char buf[IFNAMSIZ];
if (!capable(CAP_NET_ADMIN))
return -EPERM;
//copy_from_user:把用戶空間的數據拷入內核空間
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; }
|
在這裏,我們傳入的cmd爲SIOCBRADDBR.轉入br_add_bridge(buf)中進行:
int br_add_bridge(const char *name) {
struct net_device *dev;
int ret;
//爲虛擬橋新建一個net_device
dev = new_bridge_dev(name);
if (!dev)
return -ENOMEM;
rtnl_lock();
//由內核確定接口名字,例如eth0 eth1等
if (strchr(dev->name, '%')) {
ret = dev_alloc_name(dev, dev->name);
if (ret < 0)
goto err1;
}
//向內核註冊此網絡設備
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中建立相關信息
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; }
|
網橋是一個虛擬的設備,它的註冊跟實際的物理網絡設備註冊是一樣的。我們關心的是網橋對應的net_device結構是什麼樣的,繼續跟蹤進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;
//網橋的私區結構爲net_bridge
br = netdev_priv(dev);
//私區結構中的dev字段指向設備本身
br->dev = dev;
spin_lock_init(&br->lock);
//隊列初始化。在port_list中保存了這個橋上的端口列表
INIT_LIST_HEAD(&br->port_list);
spin_lock_init(&br->hash_lock);
//下面這部份代碼跟stp協議相關,我們暫不關心
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中還做了一些另外在函數指針初始化:
void br_dev_setup(struct net_device *dev) {
//將橋的MAC地址設爲零
memset(dev->dev_addr, 0, ETH_ALEN);
//初始化dev的部分函數指針,因爲目前網橋設備主適用於以及網,
//以太網的部分功能對它也適用
ether_setup(dev);
//設置設備的ioctl函數爲br_dev_ioctl
dev->do_ioctl = br_dev_ioctl;
//網橋與一般網卡不同,網橋統一統計它的數據包和字節數等信息
dev->get_stats = br_dev_get_stats;
// 網橋接口的數據包發送函數,真實設備要向外發送數據時,是通過網卡向外發送數據
// 而該網橋設備要向外發送數據時,它的處理邏輯與網橋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; }
|
僅僅創建網橋,還是不夠的。實際應用中的網橋需要添加實際的端口(即物理接口),如例子中的eth1, eth2等。應用程序在使用ioctl來爲網橋增加物理接口,對應內核函數br_dev_ioctl的代碼和分析如下:
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: //刪除
//同一處理函數,默認爲添加
return add_del_if(br, rq->ifr_ifindex, cmd == SIOCBRADDIF);
}
pr_debug("Bridge does not support ioctl 0x%x\n", cmd);
return -EOPNOTSUPP; }
|
下面分析具體的添加刪除函數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; }
|
對應的添加刪除函數分別爲: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僅支持以太網網橋--*/
if (dev->flags & IFF_LOOPBACK || dev->type != ARPHRD_ETHER)
return -EINVAL;
/*--把網橋接口當作物理接口加入到另一個網橋中,是不行的,
邏輯和代碼上都會出現 loop--*/
if (dev->hard_start_xmit == br_dev_xmit)
return -ELOOP;
/*--該物理接口已經綁定到另一個網橋了--*/
if (dev->br_port != NULL)
return -EBUSY;
/*--爲該接口創建一個網橋端口數據,並初始化好該端口的相關數據--*/
if (IS_ERR(p = new_nbp(br, dev, br_initial_port_cost(dev))))
return PTR_ERR(p);
/*--將該接口的物理地址寫入到 MAC-端口映射表中,
該MAC是屬於網橋內部端口的固定MAC地址,
它在fdb中的記錄是固定的,不會失效(agged)--*/
if ((err = br_fdb_insert(br, p, dev->dev_addr)))
destroy_nbp(p);
/*--添加相應的系統文件信息--*/
else if ((err = br_sysfs_addif(p)))
del_nbp(p);
else {
/*--打開該接口的混雜模式,網橋中的各個端口必須處於混雜模式,
網橋才能正確工作--*/
dev_set_promiscuity(dev, 1);
/*--加到端口列表--*/
list_add_rcu(&p->list, &br->port_list);
/*--STP相關設置-*/
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);
/*--設置設備的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; }
|
在內核,網橋是以模塊的方式存在,註冊源碼路徑:\net\brige\br.c:
4.1 初始化
static int __init br_init(void) {
br_fdb_init(); //網橋數據庫初始化,分配slab緩衝區
#ifdef CONFIG_BRIDGE_NETFILTER
if (br_netfilter_init()) //netfilter鉤子初始化
return 1; #endif
brioctl_set(br_ioctl_deviceless_stub); //設置ioctl鉤子函數:br_ioctl_hook
br_handle_frame_hook = br_handle_frame;//設置報文處理鉤子:br_ioctl_hook
//網橋數據庫處理鉤子
br_fdb_get_hook = br_fdb_get;
br_fdb_put_hook = br_fdb_put;
//在netdev_chain通知鏈表上註冊
register_netdevice_notifier(&br_device_notifier);
return 0; }
|
4.2 新建網橋
前面說到通過brctl addbr br0命令建立網橋,此處用戶控件調用的brctl命令最終對應到內核中的br_ioctl_deviceless_stub處理函數:
int br_ioctl_deviceless_stub(unsigned int cmd, void __user *uarg) {
switch (cmd) {
case SIOCGIFBR:
case SIOCSIFBR:
return old_deviceless(uarg);
case SIOCBRADDBR: //新建網橋
case SIOCBRDELBR: //刪除網橋
{
char buf[IFNAMSIZ];
if (!capable(CAP_NET_ADMIN))
return -EPERM;
//copy_from_user:把用戶空間的數據拷入內核空間
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; }
|
在這裏,我們傳入的cmd爲SIOCBRADDBR.轉入br_add_bridge(buf)中進行:
int br_add_bridge(const char *name) {
struct net_device *dev;
int ret;
//爲虛擬橋新建一個net_device
dev = new_bridge_dev(name);
if (!dev)
return -ENOMEM;
rtnl_lock();
//由內核確定接口名字,例如eth0 eth1等
if (strchr(dev->name, '%')) {
ret = dev_alloc_name(dev, dev->name);
if (ret < 0)
goto err1;
}
//向內核註冊此網絡設備
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中建立相關信息
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; }
|
網橋是一個虛擬的設備,它的註冊跟實際的物理網絡設備註冊是一樣的。我們關心的是網橋對應的net_device結構是什麼樣的,繼續跟蹤進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;
//網橋的私區結構爲net_bridge
br = netdev_priv(dev);
//私區結構中的dev字段指向設備本身
br->dev = dev;
spin_lock_init(&br->lock);
//隊列初始化。在port_list中保存了這個橋上的端口列表
INIT_LIST_HEAD(&br->port_list);
spin_lock_init(&br->hash_lock);
//下面這部份代碼跟stp協議相關,我們暫不關心
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中還做了一些另外在函數指針初始化:
void br_dev_setup(struct net_device *dev) {
//將橋的MAC地址設爲零
memset(dev->dev_addr, 0, ETH_ALEN);
//初始化dev的部分函數指針,因爲目前網橋設備主適用於以及網,
//以太網的部分功能對它也適用
ether_setup(dev);
//設置設備的ioctl函數爲br_dev_ioctl
dev->do_ioctl = br_dev_ioctl;
//網橋與一般網卡不同,網橋統一統計它的數據包和字節數等信息
dev->get_stats = br_dev_get_stats;
// 網橋接口的數據包發送函數,真實設備要向外發送數據時,是通過網卡向外發送數據
// 而該網橋設備要向外發送數據時,它的處理邏輯與網橋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; }
|
僅僅創建網橋,還是不夠的。實際應用中的網橋需要添加實際的端口(即物理接口),如例子中的eth1, eth2等。應用程序在使用ioctl來爲網橋增加物理接口,對應內核函數br_dev_ioctl的代碼和分析如下:
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: //刪除
//同一處理函數,默認爲添加
return add_del_if(br, rq->ifr_ifindex, cmd == SIOCBRADDIF);
}
pr_debug("Bridge does not support ioctl 0x%x\n", cmd);
return -EOPNOTSUPP; }
|
下面分析具體的添加刪除函數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; }
|
對應的添加刪除函數分別爲: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僅支持以太網網橋--*/
if (dev->flags & IFF_LOOPBACK || dev->type != ARPHRD_ETHER)
return -EINVAL;
/*--把網橋接口當作物理接口加入到另一個網橋中,是不行的,
邏輯和代碼上都會出現 loop--*/
if (dev->hard_start_xmit == br_dev_xmit)
return -ELOOP;
/*--該物理接口已經綁定到另一個網橋了--*/
if (dev->br_port != NULL)
return -EBUSY;
/*--爲該接口創建一個網橋端口數據,並初始化好該端口的相關數據--*/
if (IS_ERR(p = new_nbp(br, dev, br_initial_port_cost(dev))))
return PTR_ERR(p);
/*--將該接口的物理地址寫入到 MAC-端口映射表中,
該MAC是屬於網橋內部端口的固定MAC地址,
它在fdb中的記錄是固定的,不會失效(agged)--*/
if ((err = br_fdb_insert(br, p, dev->dev_addr)))
destroy_nbp(p);
/*--添加相應的系統文件信息--*/
else if ((err = br_sysfs_addif(p)))
del_nbp(p);
else {
/*--打開該接口的混雜模式,網橋中的各個端口必須處於混雜模式,
網橋才能正確工作--*/
dev_set_promiscuity(dev, 1);
/*--加到端口列表--*/
list_add_rcu(&p->list, &br->port_list);
/*--STP相關設置-*/
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);
/*--設置設備的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; }
|