nfnetlink_queue.c

linux 内核源代码

/*
 * This is a module which is used for queueing packets and communicating with
 * userspace via nfetlink.
 *
 * (C) 2005 by Harald Welte <[e-mail]>
 *
 * Based on the old ipv4-only ip_queue.c:
 * (C) 2000-2002 James Morris <[e-mail]>
 * (C) 2003-2005 Netfilter Core Team <[e-mail]>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 */
#include <linux/module.h>
#include <linux/skbuff.h>
#include <linux/init.h>
#include <linux/spinlock.h>
#include <linux/notifier.h>
#include <linux/netdevice.h>
#include <linux/netfilter.h>
#include <linux/proc_fs.h>
#include <linux/netfilter_ipv4.h>
#include <linux/netfilter_ipv6.h>
#include <linux/netfilter/nfnetlink.h>
#include <linux/netfilter/nfnetlink_queue.h>
#include <linux/list.h>
#include <net/sock.h>

#include <asm/atomic.h>

#ifdef CONFIG_BRIDGE_NETFILTER
#include "../bridge/br_private.h"
#endif

#define NFQNL_QMAX_DEFAULT 1024

#if 0
#define QDEBUG(x, args ...)	printk(KERN_DEBUG "%s(%d):%s():	" x, 	   \
					__FILE__, __LINE__, __FUNCTION__,  \
					## args)
#else
#define QDEBUG(x, ...)
#endif

struct nfqnl_queue_entry {
	struct list_head list;
	struct nf_info *info;
	struct sk_buff *skb;
	unsigned int id;
};

struct nfqnl_instance {
	struct hlist_node hlist;		/* global list of queues */
	atomic_t use;

	int peer_pid;
	unsigned int queue_maxlen;
	unsigned int copy_range;
	unsigned int queue_total;
	unsigned int queue_dropped;
	unsigned int queue_user_dropped;

	atomic_t id_sequence;			/* 'sequence' of pkt ids */

	u_int16_t queue_num;			/* number of this queue */
	u_int8_t copy_mode;

	spinlock_t lock;

	struct list_head queue_list;		/* packets in queue */
};

typedef int (*nfqnl_cmpfn)(struct nfqnl_queue_entry *, unsigned long);

static DEFINE_RWLOCK(instances_lock);

#define INSTANCE_BUCKETS	16
static struct hlist_head instance_table[INSTANCE_BUCKETS];

static inline u_int8_t instance_hashfn(u_int16_t queue_num)
{
	return ((queue_num >> 8) | queue_num) % INSTANCE_BUCKETS;
}

static struct nfqnl_instance *
__instance_lookup(u_int16_t queue_num)
{
	struct hlist_head *head;
	struct hlist_node *pos;
	struct nfqnl_instance *inst;

	head = &instance_table[instance_hashfn(queue_num)];
	hlist_for_each_entry(inst, pos, head, hlist) {
		if (inst->queue_num == queue_num)
			return inst;
	}
	return NULL;
}

static struct nfqnl_instance *
instance_lookup_get(u_int16_t queue_num)
{
	struct nfqnl_instance *inst;

	read_lock_bh(&instances_lock);
	inst = __instance_lookup(queue_num);
	if (inst)
		atomic_inc(&inst->use);
	read_unlock_bh(&instances_lock);

	return inst;
}

static void
instance_put(struct nfqnl_instance *inst)
{
	if (inst && atomic_dec_and_test(&inst->use)) {
		QDEBUG("kfree(inst=%p)\n", inst);
		kfree(inst);
	}
}

static struct nfqnl_instance *
instance_create(u_int16_t queue_num, int pid)
{
	struct nfqnl_instance *inst;

	QDEBUG("entering for queue_num=%u, pid=%d\n", queue_num, pid);

	write_lock_bh(&instances_lock);
	if (__instance_lookup(queue_num)) {
		inst = NULL;
		QDEBUG("aborting, instance already exists\n");
		goto out_unlock;
	}

	inst = kzalloc(sizeof(*inst), GFP_ATOMIC);
	if (!inst)
		goto out_unlock;

	inst->queue_num = queue_num;
	inst->peer_pid = pid;
	inst->queue_maxlen = NFQNL_QMAX_DEFAULT;
	inst->copy_range = 0xfffff;
	inst->copy_mode = NFQNL_COPY_NONE;
	atomic_set(&inst->id_sequence, 0);
	/* needs to be two, since we _put() after creation */
	atomic_set(&inst->use, 2);
	spin_lock_init(&inst->lock);
	INIT_LIST_HEAD(&inst->queue_list);

	if (!try_module_get(THIS_MODULE))
		goto out_free;

	hlist_add_head(&inst->hlist,
		       &instance_table[instance_hashfn(queue_num)]);

	write_unlock_bh(&instances_lock);

	QDEBUG("successfully created new instance\n");

	return inst;

out_free:
	kfree(inst);
out_unlock:
	write_unlock_bh(&instances_lock);
	return NULL;
}

static void nfqnl_flush(struct nfqnl_instance *queue, int verdict);

static void
_instance_destroy2(struct nfqnl_instance *inst, int lock)
{
	/* first pull it out of the global list */
	if (lock)
		write_lock_bh(&instances_lock);

	QDEBUG("removing instance %p (queuenum=%u) from hash\n",
		inst, inst->queue_num);
	hlist_del(&inst->hlist);

	if (lock)
		write_unlock_bh(&instances_lock);

	/* then flush all pending skbs from the queue */
	nfqnl_flush(inst, NF_DROP);

	/* and finally put the refcount */
	instance_put(inst);

	module_put(THIS_MODULE);
}

static inline void
__instance_destroy(struct nfqnl_instance *inst)
{
	_instance_destroy2(inst, 0);
}

static inline void
instance_destroy(struct nfqnl_instance *inst)
{
	_instance_destroy2(inst, 1);
}



static void
issue_verdict(struct nfqnl_queue_entry *entry, int verdict)
{
	QDEBUG("entering for entry %p, verdict %u\n", entry, verdict);

	/* TCP input path (and probably other bits) assume to be called
	 * from softirq context, not from syscall, like issue_verdict is
	 * called.  TCP input path deadlocks with locks taken from timer
	 * softirq, e.g.  We therefore emulate this by local_bh_disable() */

	local_bh_disable();
	nf_reinject(entry->skb, entry->info, verdict);
	local_bh_enable();

	kfree(entry);
}

static inline void
__enqueue_entry(struct nfqnl_instance *queue,
		      struct nfqnl_queue_entry *entry)
{
       list_add(&entry->list, &queue->queue_list);
       queue->queue_total++;
}

/*
 * Find and return a queued entry matched by cmpfn, or return the last
 * entry if cmpfn is NULL.
 */
static inline struct nfqnl_queue_entry *
__find_entry(struct nfqnl_instance *queue, nfqnl_cmpfn cmpfn,
		   unsigned long data)
{
	struct list_head *p;

	list_for_each_prev(p, &queue->queue_list) {
		struct nfqnl_queue_entry *entry = (struct nfqnl_queue_entry *)p;

		if (!cmpfn || cmpfn(entry, data))
			return entry;
	}
	return NULL;
}

static inline void
__dequeue_entry(struct nfqnl_instance *q, struct nfqnl_queue_entry *entry)
{
	list_del(&entry->list);
	q->queue_total--;
}

static inline struct nfqnl_queue_entry *
__find_dequeue_entry(struct nfqnl_instance *queue,
		     nfqnl_cmpfn cmpfn, unsigned long data)
{
	struct nfqnl_queue_entry *entry;

	entry = __find_entry(queue, cmpfn, data);
	if (entry == NULL)
		return NULL;

	__dequeue_entry(queue, entry);
	return entry;
}


static inline void
__nfqnl_flush(struct nfqnl_instance *queue, int verdict)
{
	struct nfqnl_queue_entry *entry;

	while ((entry = __find_dequeue_entry(queue, NULL, 0)))
		issue_verdict(entry, verdict);
}

static inline int
__nfqnl_set_mode(struct nfqnl_instance *queue,
		 unsigned char mode, unsigned int range)
{
	int status = 0;

	switch (mode) {
	case NFQNL_COPY_NONE:
	case NFQNL_COPY_META:
		queue->copy_mode = mode;
		queue->copy_range = 0;
		break;

	case NFQNL_COPY_PACKET:
		queue->copy_mode = mode;
		/* we're using struct nlattr which has 16bit nla_len */
		if (range > 0xffff)
			queue->copy_range = 0xffff;
		else
			queue->copy_range = range;
		break;

	default:
		status = -EINVAL;

	}
	return status;
}

static struct nfqnl_queue_entry *
find_dequeue_entry(struct nfqnl_instance *queue,
			 nfqnl_cmpfn cmpfn, unsigned long data)
{
	struct nfqnl_queue_entry *entry;

	spin_lock_bh(&queue->lock);
	entry = __find_dequeue_entry(queue, cmpfn, data);
	spin_unlock_bh(&queue->lock);

	return entry;
}

static void
nfqnl_flush(struct nfqnl_instance *queue, int verdict)
{
	spin_lock_bh(&queue->lock);
	__nfqnl_flush(queue, verdict);
	spin_unlock_bh(&queue->lock);
}

static struct sk_buff *
nfqnl_build_packet_message(struct nfqnl_instance *queue,
			   struct nfqnl_queue_entry *entry, int *errp)
{
	sk_buff_data_t old_tail;
	size_t size;
	size_t data_len = 0;
	struct sk_buff *skb;
	struct nfqnl_msg_packet_hdr pmsg;
	struct nlmsghdr *nlh;
	struct nfgenmsg *nfmsg;
	struct nf_info *entinf = entry->info;
	struct sk_buff *entskb = entry->skb;
	struct net_device *indev;
	struct net_device *outdev;
	__be32 tmp_uint;

	QDEBUG("entered\n");

	size =    NLMSG_ALIGN(sizeof(struct nfgenmsg))
		+ nla_total_size(sizeof(struct nfqnl_msg_packet_hdr))
		+ nla_total_size(sizeof(u_int32_t))	/* ifindex */
		+ nla_total_size(sizeof(u_int32_t))	/* ifindex */
#ifdef CONFIG_BRIDGE_NETFILTER
		+ nla_total_size(sizeof(u_int32_t))	/* ifindex */
		+ nla_total_size(sizeof(u_int32_t))	/* ifindex */
#endif
		+ nla_total_size(sizeof(u_int32_t))	/* mark */
		+ nla_total_size(sizeof(struct nfqnl_msg_packet_hw))
		+ nla_total_size(sizeof(struct nfqnl_msg_packet_timestamp));

	outdev = entinf->outdev;

	spin_lock_bh(&queue->lock);

	switch (queue->copy_mode) {
	case NFQNL_COPY_META:
	case NFQNL_COPY_NONE:
		data_len = 0;
		break;

	case NFQNL_COPY_PACKET:
		if ((entskb->ip_summed == CHECKSUM_PARTIAL ||
		     entskb->ip_summed == CHECKSUM_COMPLETE) &&
		    (*errp = skb_checksum_help(entskb))) {
			spin_unlock_bh(&queue->lock);
			return NULL;
		}
		if (queue->copy_range == 0
		    || queue->copy_range > entskb->len)
			data_len = entskb->len;
		else
			data_len = queue->copy_range;

		size += nla_total_size(data_len);
		break;

	default:
		*errp = -EINVAL;
		spin_unlock_bh(&queue->lock);
		return NULL;
	}

	spin_unlock_bh(&queue->lock);

	skb = alloc_skb(size, GFP_ATOMIC);
	if (!skb)
		goto nlmsg_failure;

	old_tail = skb->tail;
	nlh = NLMSG_PUT(skb, 0, 0,
			NFNL_SUBSYS_QUEUE << 8 | NFQNL_MSG_PACKET,
			sizeof(struct nfgenmsg));
	nfmsg = NLMSG_DATA(nlh);
	nfmsg->nfgen_family = entinf->pf;
	nfmsg->version = NFNETLINK_V0;
	nfmsg->res_id = htons(queue->queue_num);

	pmsg.packet_id 		= htonl(entry->id);
	pmsg.hw_protocol	= entskb->protocol;
	pmsg.hook		= entinf->hook;

	NLA_PUT(skb, NFQA_PACKET_HDR, sizeof(pmsg), &pmsg);

	indev = entinf->indev;
	if (indev) {
		tmp_uint = htonl(indev->ifindex);
#ifndef CONFIG_BRIDGE_NETFILTER
		NLA_PUT(skb, NFQA_IFINDEX_INDEV, sizeof(tmp_uint), &tmp_uint);
#else
		if (entinf->pf == PF_BRIDGE) {
			/* Case 1: indev is physical input device, we need to
			 * look for bridge group (when called from
			 * netfilter_bridge) */
			NLA_PUT(skb, NFQA_IFINDEX_PHYSINDEV, sizeof(tmp_uint),
				&tmp_uint);
			/* this is the bridge group "brX" */
			tmp_uint = htonl(indev->br_port->br->dev->ifindex);
			NLA_PUT(skb, NFQA_IFINDEX_INDEV, sizeof(tmp_uint),
				&tmp_uint);
		} else {
			/* Case 2: indev is bridge group, we need to look for
			 * physical device (when called from ipv4) */
			NLA_PUT(skb, NFQA_IFINDEX_INDEV, sizeof(tmp_uint),
				&tmp_uint);
			if (entskb->nf_bridge
			    && entskb->nf_bridge->physindev) {
				tmp_uint = htonl(entskb->nf_bridge->physindev->ifindex);
				NLA_PUT(skb, NFQA_IFINDEX_PHYSINDEV,
					sizeof(tmp_uint), &tmp_uint);
			}
		}
#endif
	}

	if (outdev) {
		tmp_uint = htonl(outdev->ifindex);
#ifndef CONFIG_BRIDGE_NETFILTER
		NLA_PUT(skb, NFQA_IFINDEX_OUTDEV, sizeof(tmp_uint), &tmp_uint);
#else
		if (entinf->pf == PF_BRIDGE) {
			/* Case 1: outdev is physical output device, we need to
			 * look for bridge group (when called from
			 * netfilter_bridge) */
			NLA_PUT(skb, NFQA_IFINDEX_PHYSOUTDEV, sizeof(tmp_uint),
				&tmp_uint);
			/* this is the bridge group "brX" */
			tmp_uint = htonl(outdev->br_port->br->dev->ifindex);
			NLA_PUT(skb, NFQA_IFINDEX_OUTDEV, sizeof(tmp_uint),
				&tmp_uint);
		} else {
			/* Case 2: outdev is bridge group, we need to look for
			 * physical output device (when called from ipv4) */
			NLA_PUT(skb, NFQA_IFINDEX_OUTDEV, sizeof(tmp_uint),
				&tmp_uint);
			if (entskb->nf_bridge
			    && entskb->nf_bridge->physoutdev) {
				tmp_uint = htonl(entskb->nf_bridge->physoutdev->ifindex);
				NLA_PUT(skb, NFQA_IFINDEX_PHYSOUTDEV,
					sizeof(tmp_uint), &tmp_uint);
			}
		}
#endif
	}

	if (entskb->mark) {
		tmp_uint = htonl(entskb->mark);
		NLA_PUT(skb, NFQA_MARK, sizeof(u_int32_t), &tmp_uint);
	}

	if (indev && entskb->dev) {
		struct nfqnl_msg_packet_hw phw;
		int len = dev_parse_header(entskb, phw.hw_addr);
		if (len) {
			phw.hw_addrlen = htons(len);
			NLA_PUT(skb, NFQA_HWADDR, sizeof(phw), &phw);
		}
	}

	if (entskb->tstamp.tv64) {
		struct nfqnl_msg_packet_timestamp ts;
		struct timeval tv = ktime_to_timeval(entskb->tstamp);
		ts.sec = cpu_to_be64(tv.tv_sec);
		ts.usec = cpu_to_be64(tv.tv_usec);

		NLA_PUT(skb, NFQA_TIMESTAMP, sizeof(ts), &ts);
	}

	if (data_len) {
		struct nlattr *nla;
		int size = nla_attr_size(data_len);

		if (skb_tailroom(skb) < nla_total_size(data_len)) {
			printk(KERN_WARNING "nf_queue: no tailroom!\n");
			goto nlmsg_failure;
		}

		nla = (struct nlattr *)skb_put(skb, nla_total_size(data_len));
		nla->nla_type = NFQA_PAYLOAD;
		nla->nla_len = size;

		if (skb_copy_bits(entskb, 0, nla_data(nla), data_len))
			BUG();
	}

	nlh->nlmsg_len = skb->tail - old_tail;
	return skb;

nlmsg_failure:
nla_put_failure:
	if (skb)
		kfree_skb(skb);
	*errp = -EINVAL;
	if (net_ratelimit())
		printk(KERN_ERR "nf_queue: error creating packet message\n");
	return NULL;
}

static int
nfqnl_enqueue_packet(struct sk_buff *skb, struct nf_info *info,
		     unsigned int queuenum, void *data)
{
	int status = -EINVAL;
	struct sk_buff *nskb;
	struct nfqnl_instance *queue;
	struct nfqnl_queue_entry *entry;

	QDEBUG("entered\n");

	queue = instance_lookup_get(queuenum);
	if (!queue) {
		QDEBUG("no queue instance matching\n");
		return -EINVAL;
	}

	if (queue->copy_mode == NFQNL_COPY_NONE) {
		QDEBUG("mode COPY_NONE, aborting\n");
		status = -EAGAIN;
		goto err_out_put;
	}

	entry = kmalloc(sizeof(*entry), GFP_ATOMIC);
	if (entry == NULL) {