linux协议栈之网桥实现之一

linux协议栈之网桥实现之一

网卡驱动最个函数netif_receive_skb.就从说起。简单起见，去掉里面预编译代码int netif_receive_skb(struct

sk_buff *skb) （net/core/dev.c）

{

struct packet_type *ptype, *pt_prev;

int ret = NET_RX_DROP;

unsigned short type;

//打接收时间戳

if (!skb->https://www.360docs.net/doc/a54842636.html,_sec)

net_timestamp(&skb->stamp);

//如果存dev->master。则更新相应指针

skb_bond(skb);

//更新CPU接收统计数据

__get_cpu_var(netdev_rx_stat).total++;

skb->h.raw = skb->nh.raw = skb->data;

skb->mac_len = skb->nh.raw -

skb->mac.raw;

pt_prev = NULL;

rcu_read_lock();

//处理所有协议模块

list_for_each_entry_rcu(ptype, &ptype_all, list) {

if (!ptype->dev || ptype->dev == skb->dev) {

if (pt_prev)

ret =

deliver_skb(skb, pt_prev);

pt_prev = ptype;

}

}

//分片处理

handle_diverter(skb);

//网桥处理

if (handle_bridge(&skb, &pt_prev,

&ret))

goto out;

type = skb->protocol;

//协议调相应模块处理。

list_for_each_entry_rcu(ptype,

&ptype_base[ntohs(type)&15], list) {

if (ptype->type == type

&&

(!ptype->dev || ptype->dev == skb->dev)) {

if (pt_prev)

ret =

deliver_skb(skb, pt_prev);

pt_prev = ptype;

}

}

if (pt_prev) {

ret = pt_prev->func(skb,

skb->dev, pt_prev);

} else {

kfree_skb(skb);

/* Jamal, now you will not able to escape explaining

* me how you were going to use this. :-)

*/

ret = NET_RX_DROP;

}

out:

rcu_read_unlock();

return ret;

}此函数主完成分片重组，网桥处理，根据不同协议调不同传输层处理模块。本节重点概述linux网桥实现处理。传输层协议分层续章节陆续给出。进入网桥处理代码：#if defined(CONFIG_BRIDGE) || defined

(CONFIG_BRIDGE_MODULE)

（net/core/dev.c）

int (*br_handle_frame_hook)(struct net_bridge_port *p, struct sk_buff **pskb);

static __inline__ int handle_bridge(struct sk_buff **pskb,

struct packet_type **pt_prev, int *ret)

{

struct net_bridge_port *port;

//回环接口？非以太网接口？

if ((*pskb)->pkt_type ==

PACKET_LOOPBACK ||

(port =

rcu_dereference((*pskb)->dev->br_port)) == NULL)

return 0;

if (*pt_prev) {

*ret = deliver_skb(*pskb, *pt_prev);

*pt_prev = NULL;

}

// br_handle_frame_hook个全局函数指针

return br_handle_frame_hook(port, pskb);

}

#else

#define handle_bridge(skb, pt_prev, ret) (0)

#endif

从此以看出。如果编译时候选择网桥模式，则进入网桥处理模块，否则，只个空函数，直接返回。

br_handle_frame_hook代表函数什么呢？网桥数据处理框架又什么样呢？

关于网桥：

网桥个二层设备，深入以当成个二层交换机。二层协议转发数据。网桥转发数据，维持个端口MAC应表，通常通CAM表。根据这张表以数据送往相应端口进行发送.网桥转发过程：1：接收个包。判断自己CAM表否含包此包源地址.如果没有，则源地址端口更新至于CAM表.2:判断包否送给本机，如果，则送往本机层协议栈处理。如果不，则查寻CAM表。找相应出口。3：如果找出口，则此包送至出口。

如果不存，各端口发送。4：如果CAM表应表项规定时间没有得更新，则除此项。网桥配置：Brctl个比较好配置网桥工具。源代码配置极其简单。们从网桥配置流程说起，看linux 核怎样步步管理。首先，创建个网桥: brctl addbr br0

(建立个br0网桥)然，接口添加进网桥：brctl addif br0 eth0 (eth0eth1添加进网桥br0)

brctl addif bro eth1OK,网桥现就配置好。这台linux主机以当作交换机使，从eth0包都以转发eth1。现，们看代码进行处理首先brctl addbr 。查看brctl代码调：

ioctl(br_socket_fd, SIOCBRADDBR, brname);然brctl addif brctl代码调：ioctl(br_socket_fd, SIOCBRADDIF, &ifr);呵呵。Brctl代码简单吧，只调户空间配置工具ioctl.Linux网桥分析：好，现就以进入核分析网桥模式：static int __init br_init(void)

（net/brige/br.c）

{

//分配slab缓冲区

br_fdb_init();

//网桥netfiter处理，以章节分析

#ifdef CONFIG_BRIDGE_NETFILTER

if (br_netfilter_init())

return 1;

#endif

//户空间ioctl调函数

brioctl_set(br_ioctl_deviceless_stub);

//接收数据包处理，就们面netif_receive_skb函数看br_handle_frame_hook

br_handle_frame_hook = br_handle_frame;

#if defined(CONFIG_ATM_LANE) ||

defined(CONFIG_ATM_LANE_MODULE)

br_fdb_get_hook = br_fdb_get;

br_fdb_put_hook = br_fdb_put;

#endif

//netdev_chain通知链表注册。关于通知链表，面已经介绍过，这里不再讨论

register_netdevice_notifier(&br_device_notifier);

return 0;

}新建网桥：从面分析以知道，户空间调ioctl(br_socket_fd, SIOCBRADDBR, brname).进入br_ioctl_deviceless_stub,以看相关处理：int br_ioctl_deviceless_stub(unsigned int cmd, void __user *uarg)

{

switch (cmd) {

case SIOCGIFBR:

case SIOCSIFBR:

return old_deviceless(uarg);

//新建网桥

case SIOCBRADDBR:

//除网桥

case SIOCBRDELBR:

{

char buf[IFNAMSIZ];

if (!capable(CAP_NET_ADMIN))

return -EPERM;

//copy_from_user:户空间数据拷入核空间

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;}

这里，们传入cmdSIOCBRADDBR.转入br_add_bridge(buf)进行：int br_add_bridge(const char *name)

{

struct net_device *dev;

int ret;

//虚拟桥新建个net_device

//面“网络设备管理”经讲述此结构

dev = new_bridge_dev(name);

if (!dev)

return -ENOMEM;

//sysfs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;}

网桥注册跟们以看物理网络设备注册样。们关心网桥应net_device结构什么样，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; 网桥私区结构net_bridge br = netdev_priv(dev);

//私区结构dev字段指向本身

br->dev = dev;

br->lock = SPIN_LOCK_UNLOCKED;

//队列始化。port_list保存这个桥端口列表

INIT_LIST_HEAD(&br->port_list);

br->hash_lock = SPIN_LOCK_UNLOCKED;

//面这部份代码跟stp协议相关,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);

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->accept_fastpath =

br_dev_accept_fastpath;

dev->tx_queue_len = 0;

dev->set_mac_address = NULL;

dev->priv_flags = IFF_EBRIDGE;

}这部份，桥设备私区空间进行始化。这里，有必给桥

net_device应私区结构：struct net_bridge

{

//读锁

spinlock_t lock;

//端口列表

struct list_head port_list;

//网桥应虚拟设备

struct net_device *dev;

//网桥应虚拟网卡统计数据

struct net_device_stats

statistics;

//hash表锁

spinlock_t hash_lock;

//MAC PORT应表，即CAM

struct hlist_head

hash[BR_HASH_SIZE];

struct list_head age_list;

/* STP */

u16 root_port;

rtnl_lock();

br->stp_enabled = 0;

br_stp_timer_init(br);

return dev;

//stp 协议应数据

unsigned char stp_enabled; //由核确定接口名字，例如eth0 eth1等

br->designated_root = br->bridge_id;

}

bridge_id designated_root; unsigned char

topology_change;

if (strchr(dev->name, '%')) {

br->root_path_cost = 0;

br_dev_setup还做些另函数指针始化：

bridge_id bridge_id;

ret = dev_alloc_name(dev, dev->name);

br->root_port = 0;

void br_dev_setup(struct net_device *dev)

u32 root_path_cost;

if (ret < 0)

br->bridge_max_age = br->max_age = 20 * HZ;

{

unsigned long max_age;

goto err1;

br->bridge_hello_time = br->hello_time = 2 * HZ;

//桥MAC地址设零

unsigned long hello_time;

}

br->bridge_forward_delay = br->forward_delay = 15 *

HZ;

memset(dev->dev_addr, 0, ETH_ALEN);

unsigned long forward_delay; //向核注册此网络设备

br->topology_change = 0;

//以太网结构始化

unsigned long

bridge_max_age;

ret = register_netdevice(dev);

br->topology_change_detected = 0;

ether_setup(dev);

unsigned long ageing_time; if (ret)

br->ageing_time = 300 * HZ;

//系列函数指针始化

unsigned long

bridge_hello_time;

goto err2;

INIT_LIST_HEAD(&br->age_list);

dev->do_ioctl = br_dev_ioctl;

unsigned long

bridge_forward_delay;

dev_hold(dev);

dev->get_stats = br_dev_get_stats;

rtnl_unlock(); dev->hard_start_xmit = br_dev_xmit;

unsigned char

topology_change_detected;