转载请注明 作者: Alanx Email:zhangsuozhu@tom.com QQ:8540426
http://hi.baidu.com/alanx/ 当内核准备好要发包时,将调用以下函数,其中最重要的结构是struct sk_buff请查相关资料 int ip_output(struct sk_buff *skb) { struct net_device *dev = skb_dst(skb)->dev; 取发包设备 IP_UPD_PO_STATS(dev_net(dev), IPSTATS_MIB_OUT, skb->len); 此处应为snmp网网协议相关的操作 skb->dev = dev; 设发发包的设备 skb->protocol = htons(ETH_P_IP); 指定IP协议 return NF_HOOK_COND(PF_INET, NF_INET_POST_ROUTING, skb, NULL, dev, ip_finish_output, !(IPCB(skb)->flags & IPSKB_REROUTED)); } 最后一句是调用勾子,就是netfilter机制最著名的勾子函数。别急,我们像下看。 NF_HOOK_COND()是一个宏,定义如下: #define NF_HOOK_COND(pf, hook, skb, indev, outdev, okfn, cond) / ({int __ret; / if ((__ret=nf_hook_thresh(pf, hook, (skb), indev, outdev, okfn, INT_MIN, cond)) == 1)/ __ret = (okfn)(skb); / __ret;}) 再像下看: static inline int nf_hook_thresh(u_int8_t pf, unsigned int hook, struct sk_buff *skb, struct net_device *indev, struct net_device *outdev, int (*okfn)(struct sk_buff *), int thresh, int cond) { if (!cond) return 1; #ifndef CONFIG_NETFILTER_DEBUG if (list_empty(&nf_hooks[pf][hook])) return 1; #endif return nf_hook_slow(pf, hook, skb, indev, outdev, okfn, thresh); } 好像明白点什么了,嗯。cond非0时或nf_hooks[pf][hook]链表为空时,int nf_hook_thresh()返回1。接着NF_HOOK_COND()会调用(okfn)(skb)的指针函数。 根据传过来的参数,我们得知,实际上就是调用:ip_finish_output(sk).最后我们再分析ip_finish_output(sk),先让我们看看cond和nf_hooks[pf][hook]链表吧。 cond: 根据我们传来的参数得知就是!(IPCB(skb)->flags & IPSKB_REROUTED),具体没细看,猜想可能是和路由重发有关吧。感兴趣的朋友可以了解一下。 nf_hooks[pf][hook]: 根据上层传过来的参数据应为:nf_hooks[PF_INET][NF_INET_POST_ROUTING] 。 让我们看看它的定义struct list_head nf_hooks[NFPROTO_NUMPROTO][NF_MAX_HOOKS] __read_mostly; 明白了吧。就是的一链表数组。我们这里用的是其中的nf_hooks[PF_INET][NF_INET_POST_ROUTING]。解释一下的话,应 为PF_INET协议时,路由发包的勾子函数链表。当我们用iptables或tc时。用netlink和内核通信,在这个链表上挂上很多结点。 有什么用呢?我们接着往下看:nf_hook_slow(pf, hook, skb, indev, outdev, okfn, thresh) 我们看一下它的构成: nt nf_hook_slow(u_int8_t pf, unsigned int hook, struct sk_buff *skb, struct net_device *indev, struct net_device *outdev, int (*okfn)(struct sk_buff *), int hook_thresh) { struct list_head *elem; unsigned int verdict; int ret = 0; /* We may already have this, but read-locks nest anyway */ rcu_read_lock(); elem = &nf_hooks[pf][hook]; next_hook: verdict = nf_iterate(&nf_hooks[pf][hook], skb, hook, indev, outdev, &elem, okfn, hook_thresh); if (verdict == NF_ACCEPT || verdict == NF_STOP) { ret = 1; } else if (verdict == NF_DROP) { kfree_skb(skb); ret = -EPERM; } else if ((verdict & NF_VERDICT_MASK) == NF_QUEUE) { if (!nf_queue(skb, elem, pf, hook, indev, outdev, okfn, verdict >> NF_VERDICT_BITS)) goto next_hook; } rcu_read_unlock(); return ret; } 关键是这句 verdict = nf_iterate(&nf_hooks[pf][hook], skb, hook, indev, outdev, &elem, okfn, hook_thresh); 根据上面返回的值,决定这个包是否通过NF_ACCEPT 或 丢弃 NF_DROP 还是进行流控入排队 NF_QUEUE.注意。这是个循环goto next_hook 接下来,我们看看关键的verdict = nf_iterate(&nf_hooks[pf][hook], skb, hook, indev, outdev, &elem, okfn, hook_thresh); unsigned int nf_iterate(struct list_head *head, struct sk_buff *skb, unsigned int hook, const struct net_device *indev, const struct net_device *outdev, struct list_head **i, int (*okfn)(struct sk_buff *), int hook_thresh) { unsigned int verdict; /* * The caller must not block between calls to this * function because of risk of continuing from deleted element. */ list_for_each_continue_rcu(*i, head) { struct nf_hook_ops *elem = (struct nf_hook_ops *)*i; if (hook_thresh > elem->priority) continue; /* Optimization: we don't need to hold module reference here, since function can't sleep. --RR */ verdict = elem->hook(hook, skb, indev, outdev, okfn); if (verdict != NF_ACCEPT) { #ifdef CONFIG_NETFILTER_DEBUG if (unlikely((verdict & NF_VERDICT_MASK) > NF_MAX_VERDICT)) { NFDEBUG("Evil return from %p(%u)./n", elem->hook, hook); continue; } #endif if (verdict != NF_REPEAT) return verdict; *i = (*i)->prev; } } return NF_ACCEPT; } 其中 1、ist_for_each_continue_rcu(*i, head) 是循环遍历head 既刚才所说的nf_hooks[PF_INET][NF_INET_POST_ROUTING]。 2、struct nf_hook_ops *elem = (struct nf_hook_ops *)*i; 取出上面链表的结点。 3、if (hook_thresh > elem->priority) 如果优先级小。则下一个结点这里 hook_thresh为INT_MIN continue; 4、verdict = elem->hook(hook, skb, indev, outdev, okfn);结于调到勾子了。万岁。最激动人心的事。上面说了勾子是我们用iptables或tc通过与内核通讯的netlink,挂到nf_hooks上的。 让我们看看。 nf_hooks上的结点的结构吧。 typedef unsigned int nf_hookfn(unsigned int hooknum, struct sk_buff *skb, const struct net_device *in, const struct net_device *out, int (*okfn)(struct sk_buff *)); struct nf_hook_ops { struct list_head list; 链表 /* User fills in from here down. */ nf_hookfn *hook; 勾子函数,就是上面调用的 struct module *owner; 模块 u_int8_t pf; 如:PF_INT unsigned int hooknum; 如:NF_INET_POST_ROUTING /* Hooks are ordered in ascending priority. */ int priority; 优先级:如INT_MIN等 }; 等以上的都处理完,最后,要么把包丢弃了。要最一开始的ip_finish_output处理发包啦!先写到此了!