IP 层收发报文简要剖析3--ip输入报文分片重组

在ip_local_deliver中，如果检测到是分片包，则需要将报文进行重组。其所有的分片被重新组合后才能提交到上层协议，每一个被重新组合的数据包文用ipq结构实例来表示

struct ipq {
    struct inet_frag_queue q;

    u32        user;//分片来源
    __be32        saddr;//原地址
    __be32        daddr;//目的地址
    __be16        id;//ip报文序列号
    u8        protocol;//上层协议号
//这四个字段来自ip首部是为了确定来自哪个ip数据报文
    u8        ecn; /* RFC3168 support */
    u16        max_df_size; /* largest frag with DF set seen */
    int             iif;
    int             vif;   /* L3 master device index */
    unsigned int    rid;//已收到的分片计数器
    struct inet_peer *peer;//记录发送方信息
   //通过rid peer 可以防止Dos攻击
};

网络空间分段管理结构

struct inet_frags {
    struct inet_frag_bucket    hash[INETFRAGS_HASHSZ];//哈希队列

    struct work_struct    frags_work;//工作队列
    unsigned int next_bucket;
    unsigned long last_rebuild_jiffies;
    bool rebuild;

    /* The first call to hashfn is responsible to initialize
     * rnd. This is best done with net_get_random_once.
     *
     * rnd_seqlock is used to let hash insertion detect
     * when it needs to re-lookup the hash chain to use.
     */
    u32            rnd;//随机数
    seqlock_t        rnd_seqlock;//
    int            qsize;//队列长度

    unsigned int        (*hashfn)(const struct inet_frag_queue *);
    bool            (*match)(const struct inet_frag_queue *q,
                     const void *arg);//分段队列匹配函数
    void            (*constructor)(struct inet_frag_queue *q,
                           const void *arg);
    void            (*destructor)(struct inet_frag_queue *);
    void            (*frag_expire)(unsigned long data);//队列过期处理函数
    struct kmem_cache    *frags_cachep;
    const char        *frags_cache_name;
};

struct netns_frags {
    /* The percpu_counter "mem" need to be cacheline aligned.
     *  mem.count must not share cacheline with other writers
     */
    struct percpu_counter   mem ____cacheline_aligned_in_smp;

    /* sysctls */
    int            timeout;超时时间
    int            high_thresh;内存使用上限
    int            low_thresh;内存使用下限
    int            max_dist;
};

/**
 * struct inet_frag_queue - fragment queue
 *
 * @lock: spinlock protecting the queue
 * @timer: queue expiration timer
 * @list: hash bucket list
 * @refcnt: reference count of the queue
 * @fragments: received fragments head
 * @fragments_tail: received fragments tail
 * @stamp: timestamp of the last received fragment
 * @len: total length of the original datagram
 * @meat: length of received fragments so far
 * @flags: fragment queue flags
 * @max_size: maximum received fragment size
 * @net: namespace that this frag belongs to
 * @list_evictor: list of queues to forcefully evict (e.g. due to low memory)
 */
struct inet_frag_queue {//inet分段队列头
    spinlock_t        lock;smp环境下 需要
    struct timer_list    timer;队列定时器，组装非常耗时，不能无休止的等待分片的到达
    struct hlist_node    list;哈希节点，链入inet分段管理结构的哈希队列
    atomic_t        refcnt;计数器
    struct sk_buff        *fragments;分段数据包队列
    struct sk_buff        *fragments_tail;
    ktime_t            stamp;时间戳
    int            len;数据包结束位置offset+len
    int            meat;与原数据长度的差距，如果和原数据包长度一样代表接收完成
    __u8            flags;
    u16            max_size;
    struct netns_frags    *net;指向网络空寂分段管理结构
    struct hlist_node    list_evictor;
};

1.1、 IP分组的初始化

void __init ipfrag_init(void)
{
    ip4_frags_ctl_register();
    register_pernet_subsys(&ip4_frags_ops);//向内核注册ipv4分段管理函数
    ip4_frags.hashfn = ip4_hashfn;//设置计算hash的函数
    //设置初始化ip 分段队列的构造函数
    ip4_frags.constructor = ip4_frag_init;
    //析构函数
    ip4_frags.destructor = ip4_frag_free;
    //队列机构长度
    ip4_frags.qsize = sizeof(struct ipq);
    //对比ip分段队列hook
    ip4_frags.match = ip4_frag_match;
    //设置分段队列过期处理函数
    ip4_frags.frag_expire = ip_expire;
    ip4_frags.frags_cache_name = ip_frag_cache_name;
    if (inet_frags_init(&ip4_frags))
        panic("IP: failed to allocate ip4_frags cache\n");
}

int inet_frags_init(struct inet_frags *f)
{
    int i;
//初始化工作队列
    INIT_WORK(&f->frags_work, inet_frag_worker);

    for (i = 0; i < INETFRAGS_HASHSZ; i++) {
        struct inet_frag_bucket *hb = &f->hash[i];//初始化hash 队列头

        spin_lock_init(&hb->chain_lock);
        INIT_HLIST_HEAD(&hb->chain);
    }

    seqlock_init(&f->rnd_seqlock);
    f->last_rebuild_jiffies = 0;
    f->frags_cachep = kmem_cache_create(f->frags_cache_name, f->qsize, 0, 0,
                        NULL);
    if (!f->frags_cachep)
        return -ENOMEM;

    return 0;
}
EXPORT_SYMBOL(inet_frags_init);

int ip_local_deliver(struct sk_buff *skb)
{
    /*
     *    Reassemble IP fragments.
     */
    struct net *net = dev_net(skb->dev);

    /* 分片重组 */
    if (ip_is_fragment(ip_hdr(skb))) {
        if (ip_defrag(net, skb, IP_DEFRAG_LOCAL_DELIVER))
            return 0;
    }

    /* 经过LOCAL_IN钩子点 */
    return NF_HOOK(NFPROTO_IPV4, NF_INET_LOCAL_IN,
               net, NULL, skb, skb->dev, NULL,
               ip_local_deliver_finish);
}

1.2、 ip分片报文重组的处理

/* Process an incoming IP datagram fragment. */
int ip_defrag(struct net *net, struct sk_buff *skb, u32 user)
{
    struct net_device *dev = skb->dev ? : skb_dst(skb)->dev;
    int vif = l3mdev_master_ifindex_rcu(dev);
    struct ipq *qp;
//递增计数
    __IP_INC_STATS(net, IPSTATS_MIB_REASMREQDS);
    skb_orphan(skb);

    /* Lookup (or create) queue header* 查找或创建IP分片队列  */
    qp = ip_find(net, ip_hdr(skb), user, vif);
    if (qp) {/* 分片队列存在 */
        int ret;

        spin_lock(&qp->q.lock);

        ret = ip_frag_queue(qp, skb);//分片数据包入队重组数据包

        spin_unlock(&qp->q.lock);
        ipq_put(qp);
        return ret;
    }
    /* 创建新的ip分片队列失败，内存不足递增失败计数*/
    __IP_INC_STATS(net, IPSTATS_MIB_REASMFAILS);
    kfree_skb(skb);
    return -ENOMEM;
}
EXPORT_SYMBOL(ip_defrag);

1.2.2 ip_find 根据ip首部以及user标志 在ipq散列表中查找对应的ipq。

/* Find the correct entry in the "incomplete datagrams" queue for
 * this IP datagram, and create new one, if nothing is found.
enum ip_defrag_users {
    IP_DEFRAG_LOCAL_DELIVER,
    IP_DEFRAG_CALL_RA_CHAIN,
    IP_DEFRAG_CONNTRACK_IN,
    __IP_DEFRAG_CONNTRACK_IN_END    = IP_DEFRAG_CONNTRACK_IN + USHRT_MAX,
    IP_DEFRAG_CONNTRACK_OUT,
    __IP_DEFRAG_CONNTRACK_OUT_END    = IP_DEFRAG_CONNTRACK_OUT + USHRT_MAX,
    IP_DEFRAG_CONNTRACK_BRIDGE_IN,
    __IP_DEFRAG_CONNTRACK_BRIDGE_IN = IP_DEFRAG_CONNTRACK_BRIDGE_IN + USHRT_MAX,
    IP_DEFRAG_VS_IN,
    IP_DEFRAG_VS_OUT,
    IP_DEFRAG_VS_FWD,
    IP_DEFRAG_AF_PACKET,
    IP_DEFRAG_MACVLAN,
};
 */
static struct ipq *ip_find(struct net *net, struct iphdr *iph,
               u32 user, int vif)
{
    struct inet_frag_queue *q;
    struct ip4_create_arg arg;
    unsigned int hash;
 /* 记录ip头和输入信息 */
    arg.iph = iph;
    arg.user = user;
    arg.vif = vif;
 /* 通过id，源地址，目的地址，协议计算hash */
    hash = ipqhashfn(iph->id, iph->saddr, iph->daddr, iph->protocol);
/* 根据hash值查找或创建队列 */
    q = inet_frag_find(&net->ipv4.frags, &ip4_frags, &arg, hash);
    if (IS_ERR_OR_NULL(q)) {
        inet_frag_maybe_warn_overflow(q, pr_fmt());
        return NULL;
    }
    return container_of(q, struct ipq, q);
}

struct inet_frag_queue *inet_frag_find(struct netns_frags *nf,
                       struct inet_frags *f, void *key,
                       unsigned int hash)
{
    struct inet_frag_bucket *hb;
    struct inet_frag_queue *q;
    int depth = 0;
 /* 分片内存已经超过了低限 */
    if (frag_mem_limit(nf) > nf->low_thresh)
        /* 进行节点回收 */
        inet_frag_schedule_worker(f); //工作队列回调函数为inet_frag_worker

    hash &= (INETFRAGS_HASHSZ - 1);
    hb = &f->hash[hash]; /* 找到hash桶 */

    spin_lock(&hb->chain_lock);
    hlist_for_each_entry(q, &hb->chain, list) { /* 遍历链表 */
        if (q->net == nf && f->match(q, key)) {
            atomic_inc(&q->refcnt); /* 增加引用计数 */
            spin_unlock(&hb->chain_lock);
            return q;
        }
        depth++;/* 记录查找深度 */
    }
    spin_unlock(&hb->chain_lock);
 /* 未找到 */
    /* 桶节点的链表深度不超过限定 */
    if (depth <= INETFRAGS_MAXDEPTH)
        return inet_frag_create(nf, f, key);/* 创建节点返回 */

    if (inet_frag_may_rebuild(f)) {
        /* 如果已经超过了重建间隔时间，则重建 */
        if (!f->rebuild)
            f->rebuild = true;
        inet_frag_schedule_worker(f);
    }

    return ERR_PTR(-ENOBUFS);
}
EXPORT_SYMBOL(inet_frag_find);

如果查找不到则会创建一个ipq 并将其插入链表中

static struct inet_frag_queue *inet_frag_create(struct netns_frags *nf,
                        struct inet_frags *f,
                        void *arg)
{
    struct inet_frag_queue *q;

    q = inet_frag_alloc(nf, f, arg);//分配队列头结构空间
    if (!q)
        return NULL;

    return inet_frag_intern(nf, q, f, arg);
}
static struct inet_frag_queue *inet_frag_alloc(struct netns_frags *nf,
                           struct inet_frags *f,
                           void *arg)
{
    struct inet_frag_queue *q;

    if (frag_mem_limit(nf) > nf->high_thresh) {//内存超过警戒线 回收内存
        inet_frag_schedule_worker(f);
        return NULL;
    }

    q = kmem_cache_zalloc(f->frags_cachep, GFP_ATOMIC);
    if (!q)
        return NULL;

    q->net = nf;//记录下网络空间的分段管理结构指针
    f->constructor(q, arg);//之前初始化时，构造函数来初始化-ip4_frag_init
    add_frag_mem_limit(nf, f->qsize);//sum 网络空间的分段内存

    setup_timer(&q->timer, f->frag_expire, (unsigned long)q);//定时器initand run
    spin_lock_init(&q->lock);
    atomic_set(&q->refcnt, 1);

    return q;
}
static void ip4_frag_init(struct inet_frag_queue *q, const void *a)
{
    struct ipq *qp = container_of(q, struct ipq, q);//获取分段队列指针
    struct netns_ipv4 *ipv4 = container_of(q->net, struct netns_ipv4,
                           frags);

    struct net *net = container_of(ipv4, struct net, ipv4);

    const struct ip4_create_arg *arg = a;//ipv4的分段信息指针

    qp->protocol = arg->iph->protocol;//IP层头部协议
    qp->id = arg->iph->id;//ip层id
    qp->ecn = ip4_frag_ecn(arg->iph->tos);
    qp->saddr = arg->iph->saddr;
    qp->daddr = arg->iph->daddr;
    qp->vif = arg->vif;
    qp->user = arg->user;
    //记录对方信息
    qp->peer = q->net->max_dist ?
        inet_getpeer_v4(net->ipv4.peers, arg->iph->saddr, arg->vif, 1) :
        NULL;
}

static struct inet_frag_queue *inet_frag_intern(struct netns_frags *nf,
                        struct inet_frag_queue *qp_in,
                        struct inet_frags *f,
                        void *arg)
{
    struct inet_frag_bucket *hb = get_frag_bucket_locked(qp_in, f);
    struct inet_frag_queue *qp;

#ifdef CONFIG_SMP
    /* With SMP race we have to recheck hash table, because
     * such entry could have been created on other cpu before
     * we acquired hash bucket lock.
     */
    hlist_for_each_entry(qp, &hb->chain, list) {
        if (qp->net == nf && f->match(qp, arg)) {
            atomic_inc(&qp->refcnt);
            spin_unlock(&hb->chain_lock);
            qp_in->flags |= INET_FRAG_COMPLETE;
            inet_frag_put(qp_in, f);
            return qp;
        }
    }
#endif
    qp = qp_in;
    if (!mod_timer(&qp->timer, jiffies + nf->timeout))
        atomic_inc(&qp->refcnt);

    atomic_inc(&qp->refcnt);//链入inet分段管理结构的hash队列
    hlist_add_head(&qp->list, &hb->chain);

    spin_unlock(&hb->chain_lock);

    return qp;
}

1/2/3 分片数据包加入重组数据包

/* Add new segment to existing queue. */
static int ip_frag_queue(struct ipq *qp, struct sk_buff *skb)
{
    struct sk_buff *prev, *next;
    struct net_device *dev;
    unsigned int fragsize;
    int flags, offset;
    int ihl, end;
    int err = -ENOENT;
    u8 ecn;

    if (qp->q.flags & INET_FRAG_COMPLETE) //分段队列接收完成 则释放此分片返回
        goto err;
/*数据包没有分段标志or  分段队列间隔过大
//重现调整分段队列是否出错
如果不是本地生成的分片，则调用ip_frag_too_far 检测
是否存在 dos攻击，存在攻击则调用邋ip_frag_reinit释放
所用分片
*/
    if (!(IPCB(skb)->flags & IPSKB_FRAG_COMPLETE) &&
        unlikely(ip_frag_too_far(qp)) &&
        unlikely(err = ip_frag_reinit(qp))) {
        ipq_kill(qp);//将ipq从散列表中移除停止定时器 计数器减一
        // 调用ipq_unlink 设置ipq为complete状态，只有complete状态才能释放
        goto err;
    }

    ecn = ip4_frag_ecn(ip_hdr(skb)->tos);
    offset = ntohs(ip_hdr(skb)->frag_off);
    flags = offset & ~IP_OFFSET;
    offset &= IP_OFFSET;
    offset <<= 3;        /* offset is in 8-byte chunks */
    ihl = ip_hdrlen(skb);
/* 获取ip首部中的数据标志位 片的偏移 首部长度 */
    /* Determine the position of this fragment. */
    end = offset + skb->len - skb_network_offset(skb) - ihl;
    err = -EINVAL;
     /**/
    /* Is this the final fragment?
如果是最后一个片则先对分片进行检测
    */
    if ((flags & IP_MF) == 0) {
        /* If we already have some bits beyond end
         * or have different end, the segment is corrupted.
         结束位置小于前一个位置，ipq已经有
         last_in 标志且分片末尾不等于原始数据长度
         */
        if (end < qp->q.len ||
            ((qp->q.flags & INET_FRAG_LAST_IN) && end != qp->q.len))
            goto err;
        qp->q.flags |= INET_FRAG_LAST_IN;
        qp->q.len = end;
        /*通过校验并设置为last_in标志，存储完整的数据长度*/
    } else {
        if (end&7) {//按8字节对其
            end &= ~7;
            if (skb->ip_summed != CHECKSUM_UNNECESSARY)
                skb->ip_summed = CHECKSUM_NONE;
        }
        if (end > qp->q.len) {
            /* 结束地址大于前一个分段数据地址
            Some bits beyond end -> corruption.
            如果设置了最后一个分段数据标志
            表示最后一个包，则错误*/
            if (qp->q.flags & INET_FRAG_LAST_IN)
                goto err;
            qp->q.len = end;//记录当前分段数据块的结束位置
        }
    }
    if (end == offset)//等于起始位置 即分片区数据长度为0
        goto err;

    err = -ENOMEM;//去掉ip首部
    if (!pskb_pull(skb, skb_network_offset(skb) + ihl))
        goto err;
//skb 数据长度为end-offset ip 有效载荷长度
    err = pskb_trim_rcsum(skb, end - offset);
    if (err)
        goto err;

    /* Find out which fragments are in front and at the back of us
     * in the chain of fragments so far.  We must know where to put
     * this fragment, right?
     */
    prev = qp->q.fragments_tail;
    if (!prev || FRAG_CB(prev)->offset < offset) {
        next = NULL;
        goto found;
    }
    prev = NULL;
    for (next = qp->q.fragments; next != NULL; next = next->next) {
        if (FRAG_CB(next)->offset >= offset)
            break;    /* bingo! */
        prev = next;
    }/*确定分片在链表中的位置，分片到达的时间顺序不同
    ipq 上的分片按照分片偏移值大小排序
    */

found:
    /* We found where to put this one.  Check for overlap with
     * preceding fragment, and, if needed, align things so that
     * any overlaps are eliminated.
     检验和和上一个分片数据是否有重叠
     */
    if (prev) {
        int i = (FRAG_CB(prev)->offset + prev->len) - offset;

        if (i > 0) {//有重叠 调用pskb_pull 消除重叠
            offset += i;
            err = -EINVAL;
            if (end <= offset)
                goto err;
            err = -ENOMEM;
            if (!pskb_pull(skb, i))
                goto err;
            if (skb->ip_summed != CHECKSUM_UNNECESSARY)
                skb->ip_summed = CHECKSUM_NONE;
        }
    }

    err = -ENOMEM;
/*如果和后面一个分片的数据有重叠，
部分重叠还是完全重叠；
重叠部分数据超过下一个分片的数据长度，咋释放
下发一个分片并在检查与后面第二个分片的数据是否
有重叠，如果没有超过下一个则调整下一个分片。
如此反复直到对所有分片都检测完。
调整片的偏移以及分片总长度
*/
    while (next && FRAG_CB(next)->offset < end) {
        int i = end - FRAG_CB(next)->offset; /* overlap is ‘i‘ bytes */

        if (i < next->len) {
            /* Eat head of the next overlapped fragment
             * and leave the loop. The next ones cannot overlap.
             */
            if (!pskb_pull(next, i))
                goto err;
            FRAG_CB(next)->offset += i;
            qp->q.meat -= i;
            if (next->ip_summed != CHECKSUM_UNNECESSARY)
                next->ip_summed = CHECKSUM_NONE;
            break;
        } else {
            struct sk_buff *free_it = next;

            /* Old fragment is completely overridden with
             * new one drop it.
             */
            next = next->next;

            if (prev)
                prev->next = next;
            else
                qp->q.fragments = next;

            qp->q.meat -= free_it->len;
            sub_frag_mem_limit(qp->q.net, free_it->truesize);
            kfree_skb(free_it);
        }
    }

    FRAG_CB(skb)->offset = offset;//当前片的偏移

    /* Insert this fragment in the chain of fragments.
    当前的片插入到ipq队列中相应的位置*/
    skb->next = next;
    if (!next)
        qp->q.fragments_tail = skb;
    if (prev)
        prev->next = skb;
    else
        qp->q.fragments = skb;

    dev = skb->dev;
    if (dev) {
        qp->iif = dev->ifindex;
        skb->dev = NULL;
    }
    qp->q.stamp = skb->tstamp;//更新时间搓
    qp->q.meat += skb->len;//sum ipq已收到分片的总长度
    qp->ecn |= ecn;
    //分片组装模块的所占内存的总长度
    add_frag_mem_limit(qp->q.net, skb->truesize);
    if (offset == 0)//为第一个片 设置标志
        qp->q.flags |= INET_FRAG_FIRST_IN;

    fragsize = skb->len + ihl;

    if (fragsize > qp->q.max_size)
        qp->q.max_size = fragsize;

    if (ip_hdr(skb)->frag_off & htons(IP_DF) &&
        fragsize > qp->max_df_size)
        qp->max_df_size = fragsize;

    if (qp->q.flags == (INET_FRAG_FIRST_IN | INET_FRAG_LAST_IN) &&
        qp->q.meat == qp->q.len) {//所有报文都到齐则重组
        unsigned long orefdst = skb->_skb_refdst;

        skb->_skb_refdst = 0UL;
        err = ip_frag_reasm(qp, prev, dev);
        skb->_skb_refdst = orefdst;
        return err;
    }

    skb_dst_drop(skb);
    return -EINPROGRESS;

err:
    kfree_skb(skb);
    return err;
}

ip_frag_reasm 重组报文；

* Build a new IP datagram from all its fragments. */
/*
*用于组装已到齐的所有分片，当原始
 * 数据包的所有分片都已到齐时，会调用此函
 * 数组装分片。
 */
static int ip_frag_reasm(struct ipq *qp, struct sk_buff *prev,
             struct net_device *dev)
{
    struct net *net = container_of(qp->q.net, struct net, ipv4.frags);
    struct iphdr *iph;
    struct sk_buff *fp, *head = qp->q.fragments;
    int len;
    int ihlen;
    int err;
    u8 ecn;
/*
     * 要开始组装了，因此调用ipq_kill()将此ipq结点从
     * ipq散列表删除，并删除定时器。
     */
    ipq_kill(qp);

    ecn = ip_frag_ecn_table[qp->ecn];
    if (unlikely(ecn == 0xff)) {
        err = -EINVAL;
        goto out_fail;
    }
    /* Make the one we just received the head. */
    if (prev) {
        head = prev->next;
        fp = skb_clone(head, GFP_ATOMIC);
        if (!fp)
            goto out_nomem;

        fp->next = head->next;
        if (!fp->next)
            qp->q.fragments_tail = fp;
        prev->next = fp;

        skb_morph(head, qp->q.fragments);
        head->next = qp->q.fragments->next;

        consume_skb(qp->q.fragments);
        qp->q.fragments = head;
    }

    WARN_ON(!head);
    WARN_ON(FRAG_CB(head)->offset != 0);

    /* Allocate a new buffer for the datagram.
    计算原始报文的长度 超过64  KB*/
    ihlen = ip_hdrlen(head);
    len = ihlen + qp->q.len;

    err = -E2BIG;
    if (len > 65535)
        goto out_oversize;

    /* Head of list must not be cloned.
     * 在组装分片时，所有的分片都会组装到第一个分片
     * 上，因此第一个分片是不能克隆的，如果是克隆的，
     * 则需为分片组装重新分配一个SKB。
      */
    if (skb_unclone(head, GFP_ATOMIC))
        goto out_nomem;

    /* If the first fragment is fragmented itself, we split
     * it to two chunks: the first with data and paged part
     * and the second, holding only fragments. */
     /*
     * 分片队列的第一个SKB不能既带有数据，又带有分片，即其
     * frag_list上不能有分片skb，如果有则重新分配一个SKB。最终的
     * 效果是，head自身不包括数据，其frag_list上链接着所有分片的
     * SKB。这也是SKB的一种表现形式，不一定是一个连续的数据块，
     * 但最终会调用skb_linearize()将这些数据都复制到一个连续的数据
     * 块中。
     */
    if (skb_has_frag_list(head)) {
        struct sk_buff *clone;
        int i, plen = 0;

        clone = alloc_skb(0, GFP_ATOMIC);
        if (!clone)
            goto out_nomem;
        clone->next = head->next;
        head->next = clone;
        skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list;
        skb_frag_list_init(head);
        for (i = 0; i < skb_shinfo(head)->nr_frags; i++)
            plen += skb_frag_size(&skb_shinfo(head)->frags[i]);
        clone->len = clone->data_len = head->data_len - plen;
        head->data_len -= clone->len;
        head->len -= clone->len;
        clone->csum = 0;
        clone->ip_summed = head->ip_summed;
        add_frag_mem_limit(qp->q.net, clone->truesize);
    }
 /*
     * 把所有分片组装起来即将分片链接到第一个
     * SKB的frag_list上，同时还需要遍历所有分片，
     * 重新计算IP数据包长度以及校验和等。
     */
    skb_shinfo(head)->frag_list = head->next;
    skb_push(head, head->data - skb_network_header(head));

    for (fp=head->next; fp; fp = fp->next) {
        head->data_len += fp->len;
        head->len += fp->len;
        if (head->ip_summed != fp->ip_summed)
            head->ip_summed = CHECKSUM_NONE;
        else if (head->ip_summed == CHECKSUM_COMPLETE)
            head->csum = csum_add(head->csum, fp->csum);
        head->truesize += fp->truesize;
    }
     /*
     * 重置首部长度、片偏移、标志位和总长度。
     */
    sub_frag_mem_limit(qp->q.net, head->truesize);

    head->next = NULL;
    head->dev = dev;
    head->tstamp = qp->q.stamp;
    IPCB(head)->frag_max_size = max(qp->max_df_size, qp->q.max_size);

    iph = ip_hdr(head);
    iph->tot_len = htons(len);
    iph->tos |= ecn;

    /* When we set IP_DF on a refragmented skb we must also force a
     * call to ip_fragment to avoid forwarding a DF-skb of size s while
     * original sender only sent fragments of size f (where f < s).
     *
     * We only set DF/IPSKB_FRAG_PMTU if such DF fragment was the largest
     * frag seen to avoid sending tiny DF-fragments in case skb was built
     * from one very small df-fragment and one large non-df frag.
     */
    if (qp->max_df_size == qp->q.max_size) {
        IPCB(head)->flags |= IPSKB_FRAG_PMTU;
        iph->frag_off = htons(IP_DF);
    } else {
        iph->frag_off = 0;
    }

    ip_send_check(iph);

    __IP_INC_STATS(net, IPSTATS_MIB_REASMOKS);
     /*
     * 既然各分片都已处理完，释放ipq的分片队列。
     */
    qp->q.fragments = NULL;
    qp->q.fragments_tail = NULL;
    return 0;

out_nomem:
    net_dbg_ratelimited("queue_glue: no memory for gluing queue %p\n", qp);
    err = -ENOMEM;
    goto out_fail;
out_oversize:
    net_info_ratelimited("Oversized IP packet from %pI4\n", &qp->saddr);
out_fail:
    __IP_INC_STATS(net, IPSTATS_MIB_REASMFAILS);
    return err;
}

1/4/4 ipq散列表重组

static void inet_frag_secret_rebuild(struct inet_frags *f)
{
    int i;

    write_seqlock_bh(&f->rnd_seqlock);//顺序锁

    if (!inet_frag_may_rebuild(f))
        goto out;
/* 获取新的用于计算hash的随机值 */
    get_random_bytes(&f->rnd, sizeof(u32));

    for (i = 0; i < INETFRAGS_HASHSZ; i++) {
        struct inet_frag_bucket *hb;
        struct inet_frag_queue *q;
        struct hlist_node *n;

        hb = &f->hash[i]; /* 取的桶节点 */
        spin_lock(&hb->chain_lock);

        hlist_for_each_entry_safe(q, n, &hb->chain, list) {
            unsigned int hval = inet_frag_hashfn(f, q);

            if (hval != i) {/* 节点不属于当前桶 */
                struct inet_frag_bucket *hb_dest;

                hlist_del(&q->list); /* 从当前桶中删除该节点 */

                /* Relink to new hash chain. */
                hb_dest = &f->hash[hval]; /* 找到目标桶 */

                /* This is the only place where we take
                 * another chain_lock while already holding
                 * one.  As this will not run concurrently,
                 * we cannot deadlock on hb_dest lock below, if its
                 * already locked it will be released soon since
                 * other caller cannot be waiting for hb lock
                 * that we‘ve taken above.
                 */
                spin_lock_nested(&hb_dest->chain_lock,
                         SINGLE_DEPTH_NESTING);/* 节点加入目标桶的链表中 */
                hlist_add_head(&q->list, &hb_dest->chain);
                spin_unlock(&hb_dest->chain_lock);
            }
        }
        spin_unlock(&hb->chain_lock);
    }
 /* 设置重建标记和重建时间 */
    f->rebuild = false;
    f->last_rebuild_jiffies = jiffies;
out:
    write_sequnlock_bh(&f->rnd_seqlock);
}

1/4/5 超时IP分片的清除

会定时清除规定 时间内没有完成重组的upq及其所有的分片

/*
 * Oops, a fragment queue timed out.  Kill it and send an ICMP reply.
 */
static void ip_expire(unsigned long arg)
{
    struct ipq *qp;
    struct net *net;

    qp = container_of((struct inet_frag_queue *) arg, struct ipq, q);
    net = container_of(qp->q.net, struct net, ipv4.frags);

    spin_lock(&qp->q.lock);
//ipq 已经是complete状态不处理 直接释放ipq以及其所有的分片
    if (qp->q.flags & INET_FRAG_COMPLETE)
        goto out;

    ipq_kill(qp);//将其从散列表移除
    __IP_INC_STATS(net, IPSTATS_MIB_REASMFAILS);//数据统计

    if (!inet_frag_evicting(&qp->q)) {//在回收队列中
        struct sk_buff *head = qp->q.fragments;
        const struct iphdr *iph;
        int err;

        __IP_INC_STATS(net, IPSTATS_MIB_REASMTIMEOUT);

        if (!(qp->q.flags & INET_FRAG_FIRST_IN) || !qp->q.fragments)
            goto out;

        rcu_read_lock();
        head->dev = dev_get_by_index_rcu(net, qp->iif);
        if (!head->dev)
            goto out_rcu_unlock;

        /* skb has no dst, perform route lookup again */
        iph = ip_hdr(head);
        err = ip_route_input_noref(head, iph->daddr, iph->saddr,
                       iph->tos, head->dev);
        if (err)
            goto out_rcu_unlock;

        /* Only an end host needs to send an ICMP
         * "Fragment Reassembly Timeout" message, per RFC792.
         */
        if (frag_expire_skip_icmp(qp->user) &&
            (skb_rtable(head)->rt_type != RTN_LOCAL))
            goto out_rcu_unlock;

        /* Send an ICMP "Fragment Reassembly Timeout" message. 发送ICMP 报文*/
        icmp_send(head, ICMP_TIME_EXCEEDED, ICMP_EXC_FRAGTIME, 0);
out_rcu_unlock:
        rcu_read_unlock();
    }
out:
    spin_unlock(&qp->q.lock);
    ipq_put(qp);
}

1/4/6 进行节点回收工作队列

为了控制ip组装所占用的内存，设置了两个阈值low_thresh 、high_thresh 当前ipq散列表所占用的内存存储在 mem变量中，这些全局变量存在如下结构中（netns_frags）

struct netns_frags {
    /* The percpu_counter "mem" need to be cacheline aligned.
     *  mem.count must not share cacheline with other writers
     */
    struct percpu_counter   mem ____cacheline_aligned_in_smp;

    /* sysctls */
    int            timeout;
    int            high_thresh;
    int            low_thresh;
    int            max_dist;
};

当mem大于high_thres 时，需要对散列表清理，直到mem值降低到low_thres。这两个值可以通过proc修改

static unsigned int
inet_evict_bucket(struct inet_frags *f, struct inet_frag_bucket *hb)
{
    struct inet_frag_queue *fq;
    struct hlist_node *n;
    unsigned int evicted = 0;
    HLIST_HEAD(expired);

    spin_lock(&hb->chain_lock);
 /* 遍历桶下的链表 */
    hlist_for_each_entry_safe(fq, n, &hb->chain, list) {
        if (!inet_fragq_should_evict(fq))/* 未超过限定，无需回收 */
            continue;

        if (!del_timer(&fq->timer)) /* 定时器无法删除 */
            continue;
/* 能够回收的节点加入到临时hash */
        hlist_add_head(&fq->list_evictor, &expired);
        ++evicted;
    }

    spin_unlock(&hb->chain_lock);
 /* 依次调用回收函数进行回收 */
    hlist_for_each_entry_safe(fq, n, &expired, list_evictor)
        f->frag_expire((unsigned long) fq);

    return evicted;
}

static void inet_frag_worker(struct work_struct *work)
{
 /*  本次回收的桶节点数 */
    unsigned int budget = INETFRAGS_EVICT_BUCKETS;
    unsigned int i, evicted = 0;
    struct inet_frags *f;

    f = container_of(work, struct inet_frags, frags_work);

    BUILD_BUG_ON(INETFRAGS_EVICT_BUCKETS >= INETFRAGS_HASHSZ);

    local_bh_disable();
 /* 从上次回收完的下一个节点开始，进行回收 */
    for (i = ACCESS_ONCE(f->next_bucket); budget; --budget) {
        evicted += inet_evict_bucket(f, &f->hash[i]);
        /* 回收并统计回收数量 */
        i = (i + 1) & (INETFRAGS_HASHSZ - 1);
        /* 回收节点数超过最大值，停止 */
        if (evicted > INETFRAGS_EVICT_MAX)
            break;
    }

    f->next_bucket = i;  /* 记录下次需要开始回收的桶节点 */

    local_bh_enable();
 /* 如果需要重建，则重建 */
    if (f->rebuild && inet_frag_may_rebuild(f))
        inet_frag_secret_rebuild(f);
}

原文地址：https://www.cnblogs.com/codestack/p/9194980.html