netfilter之nat

先说一下nat中之前容易忽略的几点总结
a. original方向报文根据规则做了NAT转换,那么reply方向一定是根据连接跟踪conntrack表项保存的tuple做转换

NF_INET_PRE_ROUTING的DNAT和NF_INET_POST_ROUTING 
的SNAT是成对工作的。
比如在NF_INET_PRE_ROUTING设置了DNAT规则,original方向的
数据包会在NF_INET_PRE_ROUTING做DNAT,那么reply的数据
包一定是在NF_INET_POST_ROUTING做SNAT(根据conntrack的
记录,不是根据规则)。
相反的,如果在NF_INET_POST_ROUTING设置了SNAT规则,
original方向的数据包会在NF_INET_POST_ROUTING做SNAT那
么reply的数据包一定是在NF_INET_PRE_ROUTING做DNAT(根据
conntrack的记录,并不是根据规则)。

NF_INET_LOCAL_OUT的DNAT和NF_INET_LOCAL_IN的SNAT也是如此。

b. 不管 SNAT 还是 DNAT,original方向的tuple是不变的。

//SNAT 规则
iptables -A POSTROUTING -t nat -s 10.10.10.8 -j SNAT --to-source 192.168.10.2
//SNAT前的tuple
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple:   192.168.10.4:22 ->10.10.10.8:2222
//SNAT后的tuple,只有reply的变了
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple:   192.168.10.4:22 ->192.168.10.2:2222

//DNAT
iptables -A OUTPUT -t nat -d 192.168.10.4  -j  DNAT --to-destination 10.10.10.12
//DNAT前的tuple
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple:   192.168.10.4:22 ->10.10.10.8:2222
//DNAT后的tuple,只有reply tuple变了
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple:   10.10.10.12:22 ->10.10.10.8:2222

c. 只有数据流的首包才需要查找NAT规则进行匹配,不管有没有匹配到NAT规则,都会将数据流对应的ct->status的IPS_SRC_NAT_DONE和IPS_DST_NAT_DONE位置1,对于后续数据包,只要判断ct->status的这两位设置了,就可以跳过匹配NAT规则流程。当然对于匹配到NAT规则需要做转换的数据包,其ct->status还会将IPS_SRC_NAT和IPS_DST_NAT位置1。

d. netfilter OUTPUT链与路由的位置关系
先查标准路由,再在output链中做DNAT(nf_nat_ipv4_local_fn),再调用ip_route_me_harder重新查找路由,因为目的ip可能已经变了。

route---->output(DNAT)---->reroute

注册nat hook函数

module_init(iptable_nat_init);
static int __init iptable_nat_init(void)
{
    //注册pernet函数,其init函数会初始化nf_nat_ipv4_table
    register_pernet_subsys(&iptable_nat_net_ops);

    //注册nat的hook函数
    nf_register_hooks(nf_nat_ipv4_ops, ARRAY_SIZE(nf_nat_ipv4_ops));

    return 0;
}

在iptable_nat_net_init中,生成默认的四个规则

static struct pernet_operations iptable_nat_net_ops = {
    .init   = iptable_nat_net_init,
    .exit   = iptable_nat_net_exit,
};

static const struct xt_table nf_nat_ipv4_table = {
    .name       = "nat",
    .valid_hooks    = (1 << NF_INET_PRE_ROUTING) |
              (1 << NF_INET_POST_ROUTING) |
              (1 << NF_INET_LOCAL_OUT) |
              (1 << NF_INET_LOCAL_IN),
    .me     = THIS_MODULE,
    .af     = NFPROTO_IPV4,
};
static int __net_init iptable_nat_net_init(struct net *net)
{
    struct ipt_replace *repl;

    repl = ipt_alloc_initial_table(&nf_nat_ipv4_table);

    net->ipv4.nat_table = ipt_register_table(net, &nf_nat_ipv4_table, repl);
    kfree(repl);
    return PTR_ERR_OR_ZERO(net->ipv4.nat_table);
}

NAT在四个hook点生效,分别对应四条链,每条链默认有一条rule,内存布局如下,初始化时没有用户添加的rule,可暂时忽略图中用户添加的rule部分,所以hook_entry和underflow指向同样的位置。


image.png

注册nat的hook函数

static struct nf_hook_ops nf_nat_ipv4_ops[] __read_mostly = {
    /* Before packet filtering, change destination */
    {
        .hook       = iptable_nat_ipv4_in,
        .owner      = THIS_MODULE,
        .pf     = NFPROTO_IPV4,
        .hooknum    = NF_INET_PRE_ROUTING,
        .priority   = NF_IP_PRI_NAT_DST,
    },
    /* After packet filtering, change source */
    {
        .hook       = iptable_nat_ipv4_out,
        .owner      = THIS_MODULE,
        .pf     = NFPROTO_IPV4,
        .hooknum    = NF_INET_POST_ROUTING,
        .priority   = NF_IP_PRI_NAT_SRC,
    },
    /* Before packet filtering, change destination */
    {
        .hook       = iptable_nat_ipv4_local_fn,
        .owner      = THIS_MODULE,
        .pf     = NFPROTO_IPV4,
        .hooknum    = NF_INET_LOCAL_OUT,
        .priority   = NF_IP_PRI_NAT_DST,
    },
    /* After packet filtering, change source */
    {
        .hook       = iptable_nat_ipv4_fn,
        .owner      = THIS_MODULE,
        .pf     = NFPROTO_IPV4,
        .hooknum    = NF_INET_LOCAL_IN,
        .priority   = NF_IP_PRI_NAT_SRC,
    },
};

注册的四个hook函数在netfilter中的位置如下


image.png

四个hook函数都会调用nf_nat_ipv4_fn,此函数实现了nat的核心功能,对于数据包的ctinfo为IP_CT_RELATED,IP_CT_RELATED_REPLY和IP_CT_NEW状态的查找nat表,因为这些状态表示是数据流的首包,如果查到nat规则,则执行target,将ct reply的tuple根据规则进行修改。在nf_nat_packet中根据tuple信息对报文做nat转换。

NF_INET_PRE_ROUTING 的dst nat和 NF_INET_POST_ROUTING 的src nat是成对工作的。
比如在NF_INET_PRE_ROUTING 设置了dst nat规则,reply的数据包一定是在NF_INET_POST_ROUTING 做src nat(根据conntrack的记录,不是根据规则)。
相反的,如果在NF_INET_POST_ROUTING 设置了src nat规则,那么reply的数据包一定是在NF_INET_PRE_ROUTING 做dst nat(根据conntrack的记录,并不是根据规则)。

NF_INET_LOCAL_OUT的dst nat和NF_INET_LOCAL_IN的src nat也是成对工作的。

nat处理流程

以 iptable_nat_ipv4_out 为例

static unsigned int iptable_nat_ipv4_out(const struct nf_hook_ops *ops,
                     struct sk_buff *skb,
                     const struct net_device *in,
                     const struct net_device *out,
                     int (*okfn)(struct sk_buff *))
{
    //iptable_nat_do_chain调用ipt_do_table在net->ipv4.nat_table
    //中匹配规则,并执行SNAT或者DNAT等target
    return nf_nat_ipv4_out(ops, skb, in, out, iptable_nat_do_chain);
}

//做个报文长度检查后执行核心函数nf_nat_ipv4_fn
unsigned int
nf_nat_ipv4_out(const struct nf_hook_ops *ops, struct sk_buff *skb,
        const struct net_device *in, const struct net_device *out,
        unsigned int (*do_chain)(const struct nf_hook_ops *ops,
                      struct sk_buff *skb,
                      const struct net_device *in,
                      const struct net_device *out,
                      struct nf_conn *ct))
{
    unsigned int ret;

    /* root is playing with raw sockets. */
    if (skb->len < sizeof(struct iphdr) ||
        ip_hdrlen(skb) < sizeof(struct iphdr))
        return NF_ACCEPT;
    //nat处理核心函数,看下面的注释
    ret = nf_nat_ipv4_fn(ops, skb, in, out, do_chain);
    ...
    return ret;
}

ct->status的位IPS_SRC_NAT_DONE和IPS_DST_NAT_DONE用来判断属于此ct的数据包是否经过NAT处理了。经过NAT处理指的是经过匹配NAT规则的流程,但是不一定匹配到NAT规则。对于匹配不到NAT规则的数据流,后续数据包也不需要再查找NAT规则,所以根据IPS_SRC_NAT_DONE和IPS_DST_NAT_DONE这两位是否置1来判断。
而ct->status的位IPS_SRC_NAT和IPS_DST_NAT不止说明数据流经过NAT规则处理,还说明属于此ct的数据包需要根据NAT规则修改数据包。

unsigned int
nf_nat_ipv4_fn(const struct nf_hook_ops *ops, struct sk_buff *skb,
           const struct net_device *in, const struct net_device *out,
           unsigned int (*do_chain)(const struct nf_hook_ops *ops,
                    struct sk_buff *skb,
                    const struct net_device *in,
                    const struct net_device *out,
                    struct nf_conn *ct))
{
    struct nf_conn *ct;
    enum ip_conntrack_info ctinfo;
    struct nf_conn_nat *nat;
    /* maniptype == SRC for postrouting. */
    //根据hook点判断出需要做SNAT还是DNAT。
    //在NF_INET_POST_ROUTING和NF_INET_LOCAL_IN做SNAT
    //在NF_INET_PRE_ROUTING和NF_INET_LOCAL_OUT做DNAT
    enum nf_nat_manip_type maniptype = HOOK2MANIP(ops->hooknum);

    /* We never see fragments: conntrack defrags on pre-routing
     * and local-out, and nf_nat_out protects post-routing.
     */
    NF_CT_ASSERT(!ip_is_fragment(ip_hdr(skb)));

    ct = nf_ct_get(skb, &ctinfo);
    /* Can't track?  It's not due to stress, or conntrack would
     * have dropped it.  Hence it's the user's responsibilty to
     * packet filter it out, or implement conntrack/NAT for that
     * protocol. 8) --RR
     */
    if (!ct)
        return NF_ACCEPT;

    /* Don't try to NAT if this packet is not conntracked */
    //如果不是track的包不做处理
    if (nf_ct_is_untracked(ct))
        return NF_ACCEPT;

    //将nat添加到ct扩展空间
    nat = nf_ct_nat_ext_add(ct);
    if (nat == NULL)
        return NF_ACCEPT;

    //根据ctinfo处理
    switch (ctinfo) {
    case IP_CT_RELATED:
    case IP_CT_RELATED_REPLY:
        if (ip_hdr(skb)->protocol == IPPROTO_ICMP) {
            if (!nf_nat_icmp_reply_translation(skb, ct, ctinfo,
                               ops->hooknum))
                return NF_DROP;
            else
                return NF_ACCEPT;
        }
        /* Fall thru... (Only ICMPs can be IP_CT_IS_REPLY) */
    case IP_CT_NEW:
        /* Seen it before?  This can happen for loopback, retrans,
         * or local packets.
         */
        //如果ct->status的IPS_SRC_NAT_DONE或者
        //IPS_DST_NAT_DONE位没有置1,则说明报文还没有经过nat处理
        if (!nf_nat_initialized(ct, maniptype)) {
            unsigned int ret;
            //do_chain为函数iptable_nat_do_chain,其调用
            //ipt_do_table在net->ipv4.nat_table中匹配规则,并
            //执行snat,dnat等target,核心函数是 
            //nf_nat_setup_info用来对匹配nat规则的数据包执行
            //nat处理,并将ct->status的IPS_SRC_NAT_DONE
            //或者IPS_DST_NAT_DONE位置1,如果确实需要修     
            //改数据包,则将ct->status的IPS_DST_NAT或者
            //IPS_SRC_NAT位置1
            ret = do_chain(ops, skb, in, out, ct);
            if (ret != NF_ACCEPT)
                return ret;

            //再次判断是否经过nat处理。
            //如果nat target处理过了,则break
            if (nf_nat_initialized(ct, HOOK2MANIP(ops->hooknum)))
                break;

            //对于那些没有匹配到nat规则的数据包,会使用数据
            //包中的ip构造一个range,也执行 
            //nf_nat_setup_info,但是因为range的ip和数据包ip
            //是一样的,所以不会设置IPS_DST_NAT或者
            //IPS_SRC_NAT。
            ret = nf_nat_alloc_null_binding(ct, ops->hooknum);
            if (ret != NF_ACCEPT)
                return ret;
        } else {
            pr_debug("Already setup manip %s for ct %p\n",
                 maniptype == NF_NAT_MANIP_SRC ? "SRC" : "DST",
                 ct);
            if (nf_nat_oif_changed(ops->hooknum, ctinfo, nat, out))
                goto oif_changed;
        }
        break;

    default:
        /* ESTABLISHED */
        NF_CT_ASSERT(ctinfo == IP_CT_ESTABLISHED ||
                 ctinfo == IP_CT_ESTABLISHED_REPLY);
        if (nf_nat_oif_changed(ops->hooknum, ctinfo, nat, out))
            goto oif_changed;
    }

    //不管有没有匹配到nat规则,都会经过此函数处理。
    //根据ct->status的IPS_DST_NAT或者IPS_SRC_NAT,
    //判断需不需要修改数据包。
    return nf_nat_packet(ct, ctinfo, ops->hooknum, skb);

oif_changed:
    nf_ct_kill_acct(ct, ctinfo, skb);
    return NF_DROP;
}

匹配上nat规则的数据包执行nat target,以snat为例

static unsigned int
xt_snat_target_v0(struct sk_buff *skb, const struct xt_action_param *par)
{
    const struct nf_nat_ipv4_multi_range_compat *mr = par->targinfo;
    struct nf_nat_range range;
    enum ip_conntrack_info ctinfo;
    struct nf_conn *ct;

    ct = nf_ct_get(skb, &ctinfo);
    NF_CT_ASSERT(ct != NULL &&
             (ctinfo == IP_CT_NEW || ctinfo == IP_CT_RELATED ||
              ctinfo == IP_CT_RELATED_REPLY));

    //将nat规则的参数转换到range中
    xt_nat_convert_range(&range, &mr->range[0]);

    //根据range修改tuple
    return nf_nat_setup_info(ct, &range, NF_NAT_MANIP_SRC);
}

//将nat规则的参数转换到range中
static void xt_nat_convert_range(struct nf_nat_range *dst,
                 const struct nf_nat_ipv4_range *src)
{
    memset(&dst->min_addr, 0, sizeof(dst->min_addr));
    memset(&dst->max_addr, 0, sizeof(dst->max_addr));

    dst->flags   = src->flags;
    dst->min_addr.ip = src->min_ip;
    dst->max_addr.ip = src->max_ip;
    dst->min_proto   = src->min;
    dst->max_proto   = src->max;
}

//核心函数,根据target的参数获取ip或者port,并修改ct的reply tuple,original tuple是不变的。
unsigned int
nf_nat_setup_info(struct nf_conn *ct,
          const struct nf_nat_range *range,
          enum nf_nat_manip_type maniptype)
{
    struct net *net = nf_ct_net(ct);
    struct nf_conntrack_tuple curr_tuple, new_tuple;
    struct nf_conn_nat *nat;

    /* nat helper or nfctnetlink also setup binding */
    nat = nf_ct_nat_ext_add(ct);
    if (nat == NULL)
        return NF_ACCEPT;

    NF_CT_ASSERT(maniptype == NF_NAT_MANIP_SRC ||
             maniptype == NF_NAT_MANIP_DST);
    BUG_ON(nf_nat_initialized(ct, maniptype));

    /* What we've got will look like inverse of reply. Normally
     * this is what is in the conntrack, except for prior
     * manipulations (future optimization: if num_manips == 0,
     * orig_tp = ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple)
     */
    nf_ct_invert_tuplepr(&curr_tuple,
                 &ct->tuplehash[IP_CT_DIR_REPLY].tuple);
    //根据range获取唯一的tuple
    get_unique_tuple(&new_tuple, &curr_tuple, range, ct, maniptype);

    //如果新的tuple和当前tuple不一样,说明确实需要做nat
    if (!nf_ct_tuple_equal(&new_tuple, &curr_tuple)) {
        struct nf_conntrack_tuple reply;

        /* Alter conntrack table so will recognize replies. */
        //获取新tuple的invert tuple
        nf_ct_invert_tuplepr(&reply, &new_tuple);
        //修改ct中reply方向的tuple为invert的tuple。original的tuple是不变的
        nf_conntrack_alter_reply(ct, &reply);

        //设置下面的标志位,说明确定需要修改数据包
        /* Non-atomic: we own this at the moment. */
        if (maniptype == NF_NAT_MANIP_SRC)
            ct->status |= IPS_SRC_NAT;
        else
            ct->status |= IPS_DST_NAT;

        if (nfct_help(ct))
            nfct_seqadj_ext_add(ct);
    }

    if (maniptype == NF_NAT_MANIP_SRC) {
        unsigned int srchash;

        srchash = hash_by_src(net, nf_ct_zone(ct),
                      &ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple);
        spin_lock_bh(&nf_nat_lock);
        /* nf_conntrack_alter_reply might re-allocate extension aera */
        nat = nfct_nat(ct);
        nat->ct = ct;
        hlist_add_head_rcu(&nat->bysource,
                   &net->ct.nat_bysource[srchash]);
        spin_unlock_bh(&nf_nat_lock);
    }

    /* It's done. */
    if (maniptype == NF_NAT_MANIP_DST)
        ct->status |= IPS_DST_NAT_DONE;
    else
        ct->status |= IPS_SRC_NAT_DONE;

    return NF_ACCEPT;
}

根据ct->status的IPS_DST_NAT或者IPS_SRC_NAT,
判断需不需要修改数据包。

/* Do packet manipulations according to nf_nat_setup_info. */
unsigned int nf_nat_packet(struct nf_conn *ct,
               enum ip_conntrack_info ctinfo,
               unsigned int hooknum,
               struct sk_buff *skb)
{
    const struct nf_nat_l3proto *l3proto;
    const struct nf_nat_l4proto *l4proto;
    enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
    unsigned long statusbit;
    enum nf_nat_manip_type mtype = HOOK2MANIP(hooknum);

    if (mtype == NF_NAT_MANIP_SRC)
        statusbit = IPS_SRC_NAT;
    else
        statusbit = IPS_DST_NAT;

    //original方向的数据包做SNAT,那reply方向的数据包肯定需要
    //做DNAT。比如在original方向会在POSTROUTING hook点对
    //数据包做SNAT,并且设置ct->status为IPS_SRC_NAT,那么
    //reply方向的DNAT应该是在PREROUTING做的,但是不可能
    //在PREROUTING上也设置一条DNAT规则,那DNAT如何做
    //呢?实现方法就是通过这里的两个if。上面的if中,因为hook点
    //是PREROUTING,所以mtype为NF_NAT_MANIP_DST,
    //statusbit为IPS_DST_NAT,又根据下面这个if,如果是reply方
    //向,则对statusbit做异或操作,statusbit结果是
    //IPS_SRC_NAT,ct->status也为IPS_SRC_NAT,这样才会继 
    //续执行,根据mtype NF_NAT_MANIP_DST修改数据包的目的ip。
    /* Invert if this is reply dir. */
    if (dir == IP_CT_DIR_REPLY)
        statusbit ^= IPS_NAT_MASK;

    /* Non-atomic: these bits don't change. */
    if (ct->status & statusbit) {
        struct nf_conntrack_tuple target;
        //获取当前方向的反方向的tuple的revert tuple
        /* We are aiming to look like inverse of other direction. */
        nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);

        l3proto = __nf_nat_l3proto_find(target.src.l3num);
        l4proto = __nf_nat_l4proto_find(target.src.l3num,
                        target.dst.protonum);
        //根据上面获取的tuple和mtype类型修改数据包
        if (!l3proto->manip_pkt(skb, 0, l4proto, &target, mtype))
            return NF_DROP;
    }
    return NF_ACCEPT;
}

举例说明

  1. snat
    client:10.10.10.8
    server:192.168.10.4
    client访问server时,在client上有一条SNAT规则如下:
    iptables -A POSTROUTING -t nat -s 10.10.10.8 -j SNAT --to-source 192.168.10.2

在client主机上,数据流如下:10.10.10.8:2222 > 192.168.10.4:22
经过conntrack模块后,ct的tuple如下:
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple: 192.168.10.4:22 ->10.10.10.8:2222

经过POSTROUING的nat hook点时,查找到这条SNAT规则,执行其target,设置ct->status为IPS_SRC_NAT,复制一份original的tuple,将此tuple源ip换成--to-source指定的ip,再将此tuple revert后的值赋值给reply的tuple,此时ct的tuple如下,original的tuple是没变化的,reply的目的ip变成 --to-source指定的ip
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple: 192.168.10.4:22 ->192.168.10.2:2222

修改完ct的reply tuple后,在函数nf_nat_packet中,mtype为NF_NAT_MANIP_SRC,所以statusbit = IPS_SRC_NAT,dir为IP_CT_DIR_ORIGINAL,所以ct->status & statusbit 为1,所以需要修改数据包。
首先获取reply方向tuple的revert tuple:192.168.10.2:2222->192.168.10.4:22,
调用 manip_pkt 修改数据包,因为mtype为NF_NAT_MANIP_SRC,所以修改源ip为 192.168.10.2

if (ct->status & statusbit) {
        struct nf_conntrack_tuple target;

        /* We are aiming to look like inverse of other direction. */
        nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);

        l3proto = __nf_nat_l3proto_find(target.src.l3num);
        l4proto = __nf_nat_l4proto_find(target.src.l3num,
                        target.dst.protonum);
        if (!l3proto->manip_pkt(skb, 0, l4proto, &target, mtype))
            return NF_DROP;
    }

所以最后从client发出去的数据流为:192.168.10.2:2222 -> 192.168.10.4:22。

从server返回的数据流为:192.168.10.4:22 -> 192.168.10.2:2222
在client上经过conntrack时,可以查找到ct表项,并且为reply方向,所以设置ctinfo = IP_CT_ESTABLISHED_REPLY,并且设置 ct->status 为 IPS_SEEN_REPLY_BIT。
然后在PREROUTING的nat hook点,此时ctinfo为IP_CT_ESTABLISHED_REPLY,所以不用查找nat表,直接执行nf_nat_packet,
此函数中,mtype为NF_NAT_MANIP_DST,所以statusbit = IPS_DST_NAT,dir为IP_CT_DIR_REPLY,所以执行 statusbit ^= IPS_NAT_MASK 后 statusbit为IPS_SRC_NAT,所以ct->status & statusbit 为1,所以需要修改数据包。
首先获取orginal方向tuple的revert tuple:192.168.10.4:22->10.10.10.8:2222,
调用 manip_pkt 修改数据包,因为mtype为NF_NAT_MANIP_DST,所以修改目的ip为 10.10.10.8.
最后进入client的数据流为:192.168.10.4:22 -> 10.10.10.8:2222.

  1. dnat
    client:10.10.10.8
    server:10.10.10.12
    client访问server时,在client上有一条DNAT规则如下:
    iptables -A OUTPUT -t nat -d 192.168.10.4 -j DNAT --to-destination 10.10.10.12

在client主机上,数据流如下:10.10.10.8:2222 > 192.168.10.4:22
经过conntrack模块后,ct的tuple如下:
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple: 192.168.10.4:22 ->10.10.10.8:2222

经过OUTPUT的nat hook点时,查找到这条DNAT规则,执行其target,设置ct->status为IPS_DST_NAT,复制一份original的tuple,将此tuple目的ip换成--to-destination指定的ip,再将此tuple revert后的值赋值给reply的tuple,此时ct的tuple如下,original的tuple是没变化的,reply的源ip变成 --to-destination指定的ip
original tuple:10.10.10.8:2222 -> 192.168.10.4:22
reply tuple: 10.10.10.12:22 ->10.10.10.8:2222

修改完ct的reply tuple后,在函数nf_nat_packet中,mtype为NF_NAT_MANIP_DST,所以statusbit = IPS_DST_NAT,dir为IP_CT_DIR_ORIGINAL,statusbit 仍然为IPS_DST_NAT,所以ct->status & statusbit 的结果为1,所以需要修改数据包。
首先获取reply方向tuple的revert tuple:10.10.10.8:2222->10.10.10.12:22,
调用 manip_pkt 修改数据包,因为mtype为NF_NAT_MANIP_DST,所以修改目的ip为 10.10.10.12

if (ct->status & statusbit) {
        struct nf_conntrack_tuple target;

        /* We are aiming to look like inverse of other direction. */
        nf_ct_invert_tuplepr(&target, &ct->tuplehash[!dir].tuple);

        l3proto = __nf_nat_l3proto_find(target.src.l3num);
        l4proto = __nf_nat_l4proto_find(target.src.l3num,
                        target.dst.protonum);
        if (!l3proto->manip_pkt(skb, 0, l4proto, &target, mtype))
            return NF_DROP;
    }

所以最后从client发出去的数据流为:10.10.10.8:2222 -> 10.10.10.12:22。

从server返回的数据流为:10.10.10.12:22 -> 10.10.10.8:2222
在client上经过conntrack时,可以查找到ct表项,并且为reply方向,所以设置ctinfo = IP_CT_ESTABLISHED_REPLY,并且设置 ct->status 为 IPS_SEEN_REPLY_BIT。
然后在INPUT的nat hook点,此时ctinfo为IP_CT_ESTABLISHED_REPLY,所以不用查找nat表,直接执行nf_nat_packet,
此函数中,mtype为NF_NAT_MANIP_SRC,所以statusbit = IPS_SRC_NAT,dir为IP_CT_DIR_REPLY,所以执行 statusbit ^= IPS_NAT_MASK 后 statusbit为IPS_DST_NAT,所以ct->status & statusbit 为1,所以需要修改数据包。
首先获取orginal方向tuple的revert tuple:192.168.10.4:22->10.10.10.8:2222,
调用 manip_pkt 修改数据包,因为mtype为NF_NAT_MANIP_SRC,所以修改目的ip为 192.168.10.4
最后进入client的数据流为:192.168.10.4:22 -> 10.10.10.8:2222.

你可能感兴趣的:(netfilter之nat)