本文主要讲解了Linux内核IP层的整体架构和对从网卡接受的报文处理流程,使用的内核的版本是2.6.32.27
为了方便理解,本文采用整体流程图加伪代码的方式对Linxu内核中IP整体实现架构和对网卡报文的处理流程进行了讲解,希望可以对大家有所帮助。阅读本文章假设大家对C语言有了一定的了解
IP层的整体实现架构
IP层接受底层数据报文的处理流程
/*
* 在NET_RX_SOFTIRQ软中后,由ETH_P_IP触发的ipv4协议入口函数
*/
int ip_rcv(struct sk_buff *skb, struct net_device *dev, struct packet_type *pt, struct net_device *orig_dev)
{
/*
* 过滤掉送往其他主机的数据包(这时网卡正在处于混杂模式)
*/
if (skb->pkt_type == PACKET_OTHERHOST)
goto drop;
iph = ip_hdr(skb);
/*头的长度是否至少是IP头长度(5); 是否是IPV4报文*/
if (iph->ihl < 5 || iph->version != 4)
goto inhdr_error;
/*IP头长度是否正确,不是伪造的长度*/
if (!pskb_may_pull(skb, iph->ihl*4))
goto inhdr_error;
iph = ip_hdr(skb);
/*检查校验和*/
if (unlikely(ip_fast_csum((u8 *)iph, iph->ihl)))
goto inhdr_error;
len = ntohs(iph->tot_len);
if (skb->len < len) {
IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INTRUNCATEDPKTS);
goto drop;
} else if (len < (iph->ihl*4))
goto inhdr_error;
/*实际尺寸不匹配套接字缓冲(skb->len)中维护的信息,则调用skb_trim调整数据包的长度*/
if (pskb_trim_rcsum(skb, len)) {
IP_INC_STATS_BH(dev_net(dev), IPSTATS_MIB_INDISCARDS);
goto drop;
}
/*调用IP_PRE_ROUTING(NF_INET_PRE_ROUTING)上注册的钩子,
*在调用钩子处理完之后,调用钩子处理完成之后,调用ip_rcv_finish
* 后面讲防火墙的时候,我们会仔细梳理*/
return NF_HOOK(PF_INET, NF_INET_PRE_ROUTING, skb, dev, NULL, ip_rcv_finish);
}
/* NF_HOOK(PF_INET, NF_INET_PRE_ROUTING, skb, dev, NULL, ip_rcv_finish)*/
#define NF_HOOK(pf, hook, skb, indev, outdev, okfn) \
NF_HOOK_THRESH(pf, hook, skb, indev, outdev, okfn, INT_MIN)
{
int __ret; \
if ((__ret=nf_hook_thresh(pf, hook, (skb), indev, outdev, okfn, thresh, 1)) == 1)\
__ret = (okfn)(skb); \
__ret;
}
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)
{
/*逐个调用注册的防火墙钩子*/
return nf_hook_slow(pf, hook, skb, indev, outdev, okfn, thresh);
}
/*
* 接收完数据包后的后续处理函数
*/
static int ip_rcv_finish(struct sk_buff *skb)
{
const struct iphdr *iph = ip_hdr(skb);
struct rtable *rt;
/*
* 激活ip_route_input,确定报文的路由,如果ip_route_input无法从FIB中找到路由
* 则丢弃数据报文,ip_route_input将在IP路由中的专题中进行讲解
*/
if (skb_dst(skb) == NULL) {
int err = ip_route_input(skb, iph->daddr, iph->saddr, iph->tos, skb->dev);
if (unlikely(err)) {
goto drop;
}
}
/*检查IP报头里面是否含有选项,如果含有建立ip_options*/
if (iph->ihl > 5 && ip_rcv_options(skb))
goto drop;
/*根据dst_entry的结果,使用skb_dst(skb)->input(skb)进行IP的路由选择
*传递给本地计算机的单播或多播,进入 ip_local_deliver();
*单播转发的报文进入ip_forward()
*多播转发进入ip_mr_input()
*/
return dst_input(skb);
{
skb_dst(skb)->input(skb)
}
drop:
kfree_skb(skb);
return NET_RX_DROP;
}
/*目的地分发策略的注册*/
static int __mkroute_input(struct sk_buff *skb,
struct fib_result *res,
struct in_device *in_dev,
__be32 daddr, __be32 saddr, u32 tos,
struct rtable **result)
{
//.......
rth->u.dst.input = ip_forward;
rth->u.dst.output = ip_output;
//......
}
static int __mkroute_output(struct rtable **result,
struct fib_result *res,
const struct flowi *fl,
const struct flowi *oldflp,
struct net_device *dev_out,
unsigned flags)
{
//......
if (flags & RTCF_LOCAL) {
rth->u.dst.input = ip_local_deliver;
rth->rt_spec_dst = fl->fl4_dst;
}
if (flags & (RTCF_BROADCAST | RTCF_MULTICAST)) {
rth->rt_spec_dst = fl->fl4_src;
if (flags & RTCF_LOCAL &&
!(dev_out->flags & IFF_LOOPBACK)) {
rth->u.dst.output = ip_mc_output;
RT_CACHE_STAT_INC(out_slow_mc);
}
#ifdef CONFIG_IP_MROUTE
if (res->type == RTN_MULTICAST) {
if (IN_DEV_MFORWARD(in_dev) &&
!ipv4_is_local_multicast(oldflp->fl4_dst)) {
rth->u.dst.input = ip_mr_input;
rth->u.dst.output = ip_mc_output;
}
}
#endif
}
//......
}
//-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
/*单播转发处理,负责处理转发相关的所有动作*/
int ip_forward(struct sk_buff *skb)
{
/*删除不是PACKET_HOST的数据包*/
if (skb->pkt_type != PACKET_HOST)
goto drop;
/*TTL递减为1之间,丢弃该报,并返回ICMP_TIME_EXCEEDED*/
if (ip_hdr(skb)->ttl <= 1)
goto too_many_hops;
/*如果skb->len大于MTU值,且Dont-Fragment被职位,则丢弃此报文,
*并返回ICMP_FRAG_NEEDED*/
if (unlikely(skb->len > dst_mtu(&rt->u.dst) && (ip_hdr(skb)->frag_off & htons(IP_DF)))) {
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_FRAG_NEEDED, htonl(dst_mtu(&rt->u.dst)));
goto drop;
}
/*检查是否有足够的空间用于输出网络设备中的MAC报头dst.header_len()
*调用skb_cow来创建一个新的足够长的skb,并且拷贝原来的所有数据
*/
if (skb_cow(skb, LL_RESERVED_SPACE(rt->u.dst.dev)+rt->u.dst.header_len))
goto drop;
iph = ip_hdr(skb);
/*TTL减少1*/
ip_decrease_ttl(iph);
/*使用IP_FORWARD中注册的钩子函数,当防火墙中的钩子都与运行完成后,
*进入ip_forward_finish*/
return NF_HOOK(PF_INET, NF_INET_FORWARD, skb, skb->dev, rt->u.dst.dev, ip_forward_finish);
sr_failed:
/*
* Strict routing permits no gatewaying
*/
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_SR_FAILED, 0);
goto drop;
too_many_hops:
/* Tell the sender its packet died... */
IP_INC_STATS_BH(dev_net(skb_dst(skb)->dev), IPSTATS_MIB_INHDRERRORS);
icmp_send(skb, ICMP_TIME_EXCEEDED, ICMP_EXC_TTL, 0);
drop:
kfree_skb(skb);
return NET_RX_DROP;
}
/*
* 该函数没有什么用途,除非启用了FASTROUTE,
* 将处理后的函数报文送入output阶段
*/
static int ip_forward_finish(struct sk_buff *skb)
{
struct ip_options * opt = &(IPCB(skb)->opt);
/*使用ip_forward_options处理IP选项*/
if (unlikely(opt->optlen))
ip_forward_options(skb);
/*送入到输出阶段*/
return dst_output(skb);
{
skb_dst(skb)->output(skb);
}
}
/*目的地分发策略的注册*/
static int __mkroute_input(struct sk_buff *skb,
struct fib_result *res,
struct in_device *in_dev,
__be32 daddr, __be32 saddr, u32 tos,
struct rtable **result)
{
//.......
rth->u.dst.input = ip_forward;
rth->u.dst.output = ip_output;
//......
}
int ip_output(struct sk_buff *skb)
{
struct net_device *dev = skb_dst(skb)->dev;
/*将skb->dev指向输出设备的dev*/
skb->dev = dev;
/*设置2层包类型为ETH_P_IP*/
skb->protocol = htons(ETH_P_IP);
/*使用防火墙中的NF_IP_POST_ROUTING中注册的钩子函数进行处理,
*处理完成之后进入ip_finish_output处理*/
return NF_HOOK_COND(PF_INET, NF_INET_POST_ROUTING, skb, NULL, dev,
ip_finish_output,
!(IPCB(skb)->flags & IPSKB_REROUTED));
}
/*判定是否进行IP分片*/
static int ip_finish_output(struct sk_buff *skb)
{
/*如果报文尺寸大于MTU,则进行IP分片后送入ip_finish_output2
*否则直接送入ip_finish_output2
*/
if (skb->len > ip_skb_dst_mtu(skb) && !skb_is_gso(skb))
return ip_fragment(skb, ip_finish_output2);
else
return ip_finish_output2(skb);
}
static const struct neigh_ops arp_generic_ops = {
.family = AF_INET,
.output = neigh_resolve_output,
.hh_output = dev_queue_xmit,
};
static const struct neigh_ops arp_hh_ops = {
.family = AF_INET,
.output = neigh_resolve_output,
.hh_output = dev_queue_xmit,
};
static void neigh_hh_init(struct neighbour *n, struct dst_entry *dst, __be16 protocol)
{
struct hh_cache *hh;
//......
if (n->nud_state & NUD_CONNECTED)
hh->hh_output = n->ops->hh_output; /*也就是dev_queue_xmit*/
else
hh->hh_output = n->ops->output;
//......
}
static void neigh_suspect(struct neighbour *neigh)
{
//.....
neigh->output = neigh->ops->output; /*也就是neigh_resolve_output*/
}
static inline int ip_finish_output2(struct sk_buff *skb)
{
/*如果2层头数据空间不够,则重新分配足够长度的SKB,并将数据复制到新的SKB后释放原来SKB*/
if (unlikely(skb_headroom(skb) < hh_len && dev->header_ops)) {
struct sk_buff *skb2;
skb2 = skb_realloc_headroom(skb, LL_RESERVED_SPACE(dev));
if (skb->sk)
skb_set_owner_w(skb2, skb->sk);
kfree_skb(skb);
skb = skb2;
}
/*如果路由出口项中已经含有2层包头缓存的引用(dst->hh),进入neigh_hh_output*/
if (dst->hh)
return neigh_hh_output(dst->hh, skb);
/*如果没有dst->hh,有dst->neighbour,则启动地址解析协议,也就是neigh_resolve_output*/
else if (dst->neighbour)
return dst->neighbour->output(skb);
}
static inline int neigh_hh_output(struct hh_cache *hh, struct sk_buff *skb)
{
/*直接复制2层包头到套接字的包数据空间中*/
memcpy(skb->data - hh_alen, hh->hh_data, hh_alen);
skb_push(skb, hh_len);
/*调用hh->hh_output(skb),也就是dev_queue_xmit进行硬件发送*/
return hh->hh_output(skb);
}
//-----------------------------------------------------------------------------------------------------------------------------------------------------------------------
/*包的本地投递*/
int ip_local_deliver(struct sk_buff *skb)
{
/*收集并组装IP分片,如果还没有收集完成,那么就等待IP分片组装完成*/
if (ip_hdr(skb)->frag_off & htons(IP_MF | IP_OFFSET)) {
if (ip_defrag(skb, IP_DEFRAG_LOCAL_DELIVER))
return 0;
}
/*进入NF_IP_LOCAL_IN的过滤器处理,处理完成后进入ip_local_deliver_finish*/
return NF_HOOK(PF_INET, NF_INET_LOCAL_IN, skb, skb->dev, NULL, ip_local_deliver_finish);
}
/*IP层处理完成后的协议分发函数*/
static int ip_local_deliver_finish(struct sk_buff *skb)
{
resubmit:
/*如果是RAW-IP报文,送往RAW-IP对应的处理???*/
raw = raw_local_deliver(skb, protocol);
/*MAX_INET_PROTOS-1 为IP报头中协议的模,
*这里计算对应协议在ipprot中被散列的位置*/
hash = protocol & (MAX_INET_PROTOS - 1);
/*IP层上的ipprot负责管理所有的传输协议*/
ipprot = rcu_dereference(inet_protos[hash]);
/*如果找到相应的协议,那么调用对应的处理例程*/
if (ipprot != NULL) {
ret = ipprot->handler(skb);
if (ret < 0) {
protocol = -ret;
goto resubmit;
}
}
/*找不到相应的处理例程*/
else {
/*又是RAW-IP报文,会在RAW-IP处理例程???
* 就丢弃,并想对端发送ICMP_DEST_UNREACH,ICMP_PROT_UNREACH*/
if (!raw)
{
icmp_send(skb, ICMP_DEST_UNREACH, ICMP_PROT_UNREACH, 0);
}
kfree_skb(skb);
}
return 0;
}
static const struct net_protocol tcp_protocol = {
.handler = tcp_v4_rcv, /*TCP*/
};
static const struct net_protocol udp_protocol = {
.handler = udp_rcv, /*UDP*/
};
static const struct net_protocol icmp_protocol = {
.handler = icmp_rcv, /*ICMP*/
};
static const struct net_protocol igmp_protocol = {
.handler = igmp_rcv, /*IGMP*/
};
关于二层数据报文的处理流程和是如何送到三层进行处理,请参考我前面的博客
《Linux内核二层数据包接收流程》
希望大家批评指正