[linux-2.6-block.git] / net / ieee802154 / 6lowpan / reassembly.c

/*	6LoWPAN fragment reassembly
 *
 *
 *	Authors:
 *	Alexander Aring		<aar@pengutronix.de>
 *
 *	Based on: net/ipv6/reassembly.c
 *
 *	This program is free software; you can redistribute it and/or
 *	modify it under the terms of the GNU General Public License
 *	as published by the Free Software Foundation; either version
 *	2 of the License, or (at your option) any later version.
 */

#define pr_fmt(fmt) "6LoWPAN: " fmt

#include <linux/net.h>
#include <linux/list.h>
#include <linux/netdevice.h>
#include <linux/random.h>
#include <linux/jhash.h>
#include <linux/skbuff.h>
#include <linux/slab.h>
#include <linux/export.h>

#include <net/ieee802154_netdev.h>
#include <net/6lowpan.h>
#include <net/ipv6_frag.h>
#include <net/inet_frag.h>

#include "6lowpan_i.h"

static const char lowpan_frags_cache_name[] = "lowpan-frags";

static struct inet_frags lowpan_frags;

static int lowpan_frag_reasm(struct lowpan_frag_queue *fq,
			     struct sk_buff *prev, struct net_device *ldev);

static void lowpan_frag_init(struct inet_frag_queue *q, const void *a)
{
	const struct frag_lowpan_compare_key *key = a;

	BUILD_BUG_ON(sizeof(*key) > sizeof(q->key));
	memcpy(&q->key, key, sizeof(*key));
}

static void lowpan_frag_expire(struct timer_list *t)
{
	struct inet_frag_queue *frag = from_timer(frag, t, timer);
	struct frag_queue *fq;

	fq = container_of(frag, struct frag_queue, q);

	spin_lock(&fq->q.lock);

	if (fq->q.flags & INET_FRAG_COMPLETE)
		goto out;

	inet_frag_kill(&fq->q);
out:
	spin_unlock(&fq->q.lock);
	inet_frag_put(&fq->q);
}

static inline struct lowpan_frag_queue *
fq_find(struct net *net, const struct lowpan_802154_cb *cb,
	const struct ieee802154_addr *src,
	const struct ieee802154_addr *dst)
{
	struct netns_ieee802154_lowpan *ieee802154_lowpan =
		net_ieee802154_lowpan(net);
	struct frag_lowpan_compare_key key = {};
	struct inet_frag_queue *q;

	key.tag = cb->d_tag;
	key.d_size = cb->d_size;
	key.src = *src;
	key.dst = *dst;

	q = inet_frag_find(&ieee802154_lowpan->frags, &key);
	if (!q)
		return NULL;

	return container_of(q, struct lowpan_frag_queue, q);
}

static int lowpan_frag_queue(struct lowpan_frag_queue *fq,
			     struct sk_buff *skb, u8 frag_type)
{
	struct sk_buff *prev, *next;
	struct net_device *ldev;
	int end, offset;

	if (fq->q.flags & INET_FRAG_COMPLETE)
		goto err;

	offset = lowpan_802154_cb(skb)->d_offset << 3;
	end = lowpan_802154_cb(skb)->d_size;

	/* Is this the final fragment? */
	if (offset + skb->len == end) {
		/* If we already have some bits beyond end
		 * or have different end, the segment is corrupted.
		 */
		if (end < fq->q.len ||
		    ((fq->q.flags & INET_FRAG_LAST_IN) && end != fq->q.len))
			goto err;
		fq->q.flags |= INET_FRAG_LAST_IN;
		fq->q.len = end;
	} else {
		if (end > fq->q.len) {
			/* Some bits beyond end -> corruption. */
			if (fq->q.flags & INET_FRAG_LAST_IN)
				goto err;
			fq->q.len = end;
		}
	}

	/* 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 = fq->q.fragments_tail;
	if (!prev ||
	    lowpan_802154_cb(prev)->d_offset <
	    lowpan_802154_cb(skb)->d_offset) {
		next = NULL;
		goto found;
	}
	prev = NULL;
	for (next = fq->q.fragments; next != NULL; next = next->next) {
		if (lowpan_802154_cb(next)->d_offset >=
		    lowpan_802154_cb(skb)->d_offset)
			break;	/* bingo! */
		prev = next;
	}

found:
	/* Insert this fragment in the chain of fragments. */
	skb->next = next;
	if (!next)
		fq->q.fragments_tail = skb;
	if (prev)
		prev->next = skb;
	else
		fq->q.fragments = skb;

	ldev = skb->dev;
	if (ldev)
		skb->dev = NULL;

	fq->q.stamp = skb->tstamp;
	if (frag_type == LOWPAN_DISPATCH_FRAG1)
		fq->q.flags |= INET_FRAG_FIRST_IN;

	fq->q.meat += skb->len;
	add_frag_mem_limit(fq->q.net, skb->truesize);

	if (fq->q.flags == (INET_FRAG_FIRST_IN | INET_FRAG_LAST_IN) &&
	    fq->q.meat == fq->q.len) {
		int res;
		unsigned long orefdst = skb->_skb_refdst;

		skb->_skb_refdst = 0UL;
		res = lowpan_frag_reasm(fq, prev, ldev);
		skb->_skb_refdst = orefdst;
		return res;
	}

	return -1;
err:
	kfree_skb(skb);
	return -1;
}

/*	Check if this packet is complete.
 *	Returns NULL on failure by any reason, and pointer
 *	to current nexthdr field in reassembled frame.
 *
 *	It is called with locked fq, and caller must check that
 *	queue is eligible for reassembly i.e. it is not COMPLETE,
 *	the last and the first frames arrived and all the bits are here.
 */
static int lowpan_frag_reasm(struct lowpan_frag_queue *fq, struct sk_buff *prev,
			     struct net_device *ldev)
{
	struct sk_buff *fp, *head = fq->q.fragments;
	int sum_truesize;

	inet_frag_kill(&fq->q);

	/* Make the one we just received the head. */
	if (prev) {
		head = prev->next;
		fp = skb_clone(head, GFP_ATOMIC);

		if (!fp)
			goto out_oom;

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

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

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

	/* Head of list must not be cloned. */
	if (skb_unclone(head, GFP_ATOMIC))
		goto out_oom;

	/* 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.
	 */
	if (skb_has_frag_list(head)) {
		struct sk_buff *clone;
		int i, plen = 0;

		clone = alloc_skb(0, GFP_ATOMIC);
		if (!clone)
			goto out_oom;
		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 = head->data_len - plen;
		clone->data_len = clone->len;
		head->data_len -= clone->len;
		head->len -= clone->len;
		add_frag_mem_limit(fq->q.net, clone->truesize);
	}

	WARN_ON(head == NULL);

	sum_truesize = head->truesize;
	for (fp = head->next; fp;) {
		bool headstolen;
		int delta;
		struct sk_buff *next = fp->next;

		sum_truesize += fp->truesize;
		if (skb_try_coalesce(head, fp, &headstolen, &delta)) {
			kfree_skb_partial(fp, headstolen);
		} else {
			if (!skb_shinfo(head)->frag_list)
				skb_shinfo(head)->frag_list = fp;
			head->data_len += fp->len;
			head->len += fp->len;
			head->truesize += fp->truesize;
		}
		fp = next;
	}
	sub_frag_mem_limit(fq->q.net, sum_truesize);

	skb_mark_not_on_list(head);
	head->dev = ldev;
	head->tstamp = fq->q.stamp;

	fq->q.fragments = NULL;
	fq->q.fragments_tail = NULL;

	return 1;
out_oom:
	net_dbg_ratelimited("lowpan_frag_reasm: no memory for reassembly\n");
	return -1;
}

static int lowpan_frag_rx_handlers_result(struct sk_buff *skb,
					  lowpan_rx_result res)
{
	switch (res) {
	case RX_QUEUED:
		return NET_RX_SUCCESS;
	case RX_CONTINUE:
		/* nobody cared about this packet */
		net_warn_ratelimited("%s: received unknown dispatch\n",
				     __func__);

		/* fall-through */
	default:
		/* all others failure */
		return NET_RX_DROP;
	}
}

static lowpan_rx_result lowpan_frag_rx_h_iphc(struct sk_buff *skb)
{
	int ret;

	if (!lowpan_is_iphc(*skb_network_header(skb)))
		return RX_CONTINUE;

	ret = lowpan_iphc_decompress(skb);
	if (ret < 0)
		return RX_DROP;

	return RX_QUEUED;
}

static int lowpan_invoke_frag_rx_handlers(struct sk_buff *skb)
{
	lowpan_rx_result res;

#define CALL_RXH(rxh)			\
	do {				\
		res = rxh(skb);	\
		if (res != RX_CONTINUE)	\
			goto rxh_next;	\
	} while (0)

	/* likely at first */
	CALL_RXH(lowpan_frag_rx_h_iphc);
	CALL_RXH(lowpan_rx_h_ipv6);

rxh_next:
	return lowpan_frag_rx_handlers_result(skb, res);
#undef CALL_RXH
}

#define LOWPAN_FRAG_DGRAM_SIZE_HIGH_MASK	0x07
#define LOWPAN_FRAG_DGRAM_SIZE_HIGH_SHIFT	8

static int lowpan_get_cb(struct sk_buff *skb, u8 frag_type,
			 struct lowpan_802154_cb *cb)
{
	bool fail;
	u8 high = 0, low = 0;
	__be16 d_tag = 0;

	fail = lowpan_fetch_skb(skb, &high, 1);
	fail |= lowpan_fetch_skb(skb, &low, 1);
	/* remove the dispatch value and use first three bits as high value
	 * for the datagram size
	 */
	cb->d_size = (high & LOWPAN_FRAG_DGRAM_SIZE_HIGH_MASK) <<
		LOWPAN_FRAG_DGRAM_SIZE_HIGH_SHIFT | low;
	fail |= lowpan_fetch_skb(skb, &d_tag, 2);
	cb->d_tag = ntohs(d_tag);

	if (frag_type == LOWPAN_DISPATCH_FRAGN) {
		fail |= lowpan_fetch_skb(skb, &cb->d_offset, 1);
	} else {
		skb_reset_network_header(skb);
		cb->d_offset = 0;
		/* check if datagram_size has ipv6hdr on FRAG1 */
		fail |= cb->d_size < sizeof(struct ipv6hdr);
		/* check if we can dereference the dispatch value */
		fail |= !skb->len;
	}

	if (unlikely(fail))
		return -EIO;

	return 0;
}

int lowpan_frag_rcv(struct sk_buff *skb, u8 frag_type)
{
	struct lowpan_frag_queue *fq;
	struct net *net = dev_net(skb->dev);
	struct lowpan_802154_cb *cb = lowpan_802154_cb(skb);
	struct ieee802154_hdr hdr = {};
	int err;

	if (ieee802154_hdr_peek_addrs(skb, &hdr) < 0)
		goto err;

	err = lowpan_get_cb(skb, frag_type, cb);
	if (err < 0)
		goto err;

	if (frag_type == LOWPAN_DISPATCH_FRAG1) {
		err = lowpan_invoke_frag_rx_handlers(skb);
		if (err == NET_RX_DROP)
			goto err;
	}

	if (cb->d_size > IPV6_MIN_MTU) {
		net_warn_ratelimited("lowpan_frag_rcv: datagram size exceeds MTU\n");
		goto err;
	}

	fq = fq_find(net, cb, &hdr.source, &hdr.dest);
	if (fq != NULL) {
		int ret;

		spin_lock(&fq->q.lock);
		ret = lowpan_frag_queue(fq, skb, frag_type);
		spin_unlock(&fq->q.lock);

		inet_frag_put(&fq->q);
		return ret;
	}

err:
	kfree_skb(skb);
	return -1;
}

#ifdef CONFIG_SYSCTL

static struct ctl_table lowpan_frags_ns_ctl_table[] = {
	{
		.procname	= "6lowpanfrag_high_thresh",
		.data		= &init_net.ieee802154_lowpan.frags.high_thresh,
		.maxlen		= sizeof(unsigned long),
		.mode		= 0644,
		.proc_handler	= proc_doulongvec_minmax,
		.extra1		= &init_net.ieee802154_lowpan.frags.low_thresh
	},
	{
		.procname	= "6lowpanfrag_low_thresh",
		.data		= &init_net.ieee802154_lowpan.frags.low_thresh,
		.maxlen		= sizeof(unsigned long),
		.mode		= 0644,
		.proc_handler	= proc_doulongvec_minmax,
		.extra2		= &init_net.ieee802154_lowpan.frags.high_thresh
	},
	{
		.procname	= "6lowpanfrag_time",
		.data		= &init_net.ieee802154_lowpan.frags.timeout,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
	{ }
};

/* secret interval has been deprecated */
static int lowpan_frags_secret_interval_unused;
static struct ctl_table lowpan_frags_ctl_table[] = {
	{
		.procname	= "6lowpanfrag_secret_interval",
		.data		= &lowpan_frags_secret_interval_unused,
		.maxlen		= sizeof(int),
		.mode		= 0644,
		.proc_handler	= proc_dointvec_jiffies,
	},
	{ }
};

static int __net_init lowpan_frags_ns_sysctl_register(struct net *net)
{
	struct ctl_table *table;
	struct ctl_table_header *hdr;
	struct netns_ieee802154_lowpan *ieee802154_lowpan =
		net_ieee802154_lowpan(net);

	table = lowpan_frags_ns_ctl_table;
	if (!net_eq(net, &init_net)) {
		table = kmemdup(table, sizeof(lowpan_frags_ns_ctl_table),
				GFP_KERNEL);
		if (table == NULL)
			goto err_alloc;

		table[0].data = &ieee802154_lowpan->frags.high_thresh;
		table[0].extra1 = &ieee802154_lowpan->frags.low_thresh;
		table[1].data = &ieee802154_lowpan->frags.low_thresh;
		table[1].extra2 = &ieee802154_lowpan->frags.high_thresh;
		table[2].data = &ieee802154_lowpan->frags.timeout;

		/* Don't export sysctls to unprivileged users */
		if (net->user_ns != &init_user_ns)
			table[0].procname = NULL;
	}

	hdr = register_net_sysctl(net, "net/ieee802154/6lowpan", table);
	if (hdr == NULL)
		goto err_reg;

	ieee802154_lowpan->sysctl.frags_hdr = hdr;
	return 0;

err_reg:
	if (!net_eq(net, &init_net))
		kfree(table);
err_alloc:
	return -ENOMEM;
}

static void __net_exit lowpan_frags_ns_sysctl_unregister(struct net *net)
{
	struct ctl_table *table;
	struct netns_ieee802154_lowpan *ieee802154_lowpan =
		net_ieee802154_lowpan(net);

	table = ieee802154_lowpan->sysctl.frags_hdr->ctl_table_arg;
	unregister_net_sysctl_table(ieee802154_lowpan->sysctl.frags_hdr);
	if (!net_eq(net, &init_net))
		kfree(table);
}

static struct ctl_table_header *lowpan_ctl_header;

static int __init lowpan_frags_sysctl_register(void)
{
	lowpan_ctl_header = register_net_sysctl(&init_net,
						"net/ieee802154/6lowpan",
						lowpan_frags_ctl_table);
	return lowpan_ctl_header == NULL ? -ENOMEM : 0;
}

static void lowpan_frags_sysctl_unregister(void)
{
	unregister_net_sysctl_table(lowpan_ctl_header);
}
#else
static inline int lowpan_frags_ns_sysctl_register(struct net *net)
{
	return 0;
}

static inline void lowpan_frags_ns_sysctl_unregister(struct net *net)
{
}

static inline int __init lowpan_frags_sysctl_register(void)
{
	return 0;
}

static inline void lowpan_frags_sysctl_unregister(void)
{
}
#endif

static int __net_init lowpan_frags_init_net(struct net *net)
{
	struct netns_ieee802154_lowpan *ieee802154_lowpan =
		net_ieee802154_lowpan(net);
	int res;

	ieee802154_lowpan->frags.high_thresh = IPV6_FRAG_HIGH_THRESH;
	ieee802154_lowpan->frags.low_thresh = IPV6_FRAG_LOW_THRESH;
	ieee802154_lowpan->frags.timeout = IPV6_FRAG_TIMEOUT;
	ieee802154_lowpan->frags.f = &lowpan_frags;

	res = inet_frags_init_net(&ieee802154_lowpan->frags);
	if (res < 0)
		return res;
	res = lowpan_frags_ns_sysctl_register(net);
	if (res < 0)
		inet_frags_exit_net(&ieee802154_lowpan->frags);
	return res;
}

static void __net_exit lowpan_frags_exit_net(struct net *net)
{
	struct netns_ieee802154_lowpan *ieee802154_lowpan =
		net_ieee802154_lowpan(net);

	lowpan_frags_ns_sysctl_unregister(net);
	inet_frags_exit_net(&ieee802154_lowpan->frags);
}

static struct pernet_operations lowpan_frags_ops = {
	.init = lowpan_frags_init_net,
	.exit = lowpan_frags_exit_net,
};

static u32 lowpan_key_hashfn(const void *data, u32 len, u32 seed)
{
	return jhash2(data,
		      sizeof(struct frag_lowpan_compare_key) / sizeof(u32), seed);
}

static u32 lowpan_obj_hashfn(const void *data, u32 len, u32 seed)
{
	const struct inet_frag_queue *fq = data;

	return jhash2((const u32 *)&fq->key,
		      sizeof(struct frag_lowpan_compare_key) / sizeof(u32), seed);
}

static int lowpan_obj_cmpfn(struct rhashtable_compare_arg *arg, const void *ptr)
{
	const struct frag_lowpan_compare_key *key = arg->key;
	const struct inet_frag_queue *fq = ptr;

	return !!memcmp(&fq->key, key, sizeof(*key));
}

static const struct rhashtable_params lowpan_rhash_params = {
	.head_offset		= offsetof(struct inet_frag_queue, node),
	.hashfn			= lowpan_key_hashfn,
	.obj_hashfn		= lowpan_obj_hashfn,
	.obj_cmpfn		= lowpan_obj_cmpfn,
	.automatic_shrinking	= true,
};

int __init lowpan_net_frag_init(void)
{
	int ret;

	lowpan_frags.constructor = lowpan_frag_init;
	lowpan_frags.destructor = NULL;
	lowpan_frags.qsize = sizeof(struct frag_queue);
	lowpan_frags.frag_expire = lowpan_frag_expire;
	lowpan_frags.frags_cache_name = lowpan_frags_cache_name;
	lowpan_frags.rhash_params = lowpan_rhash_params;
	ret = inet_frags_init(&lowpan_frags);
	if (ret)
		goto out;

	ret = lowpan_frags_sysctl_register();
	if (ret)
		goto err_sysctl;

	ret = register_pernet_subsys(&lowpan_frags_ops);
	if (ret)
		goto err_pernet;
out:
	return ret;
err_pernet:
	lowpan_frags_sysctl_unregister();
err_sysctl:
	inet_frags_fini(&lowpan_frags);
	return ret;
}

void lowpan_net_frag_exit(void)
{
	inet_frags_fini(&lowpan_frags);
	lowpan_frags_sysctl_unregister();
	unregister_pernet_subsys(&lowpan_frags_ops);
}
Commit	Line	Data
7240cdec AA	1	/* 6LoWPAN fragment reassembly
	2	*
	3	*
	4	* Authors:
	5	* Alexander Aring <aar@pengutronix.de>
	6	*
	7	* Based on: net/ipv6/reassembly.c
	8	*
	9	* This program is free software; you can redistribute it and/or
	10	* modify it under the terms of the GNU General Public License
	11	* as published by the Free Software Foundation; either version
	12	* 2 of the License, or (at your option) any later version.
	13	*/
	14
	15	#define pr_fmt(fmt) "6LoWPAN: " fmt
	16
	17	#include <linux/net.h>
	18	#include <linux/list.h>
	19	#include <linux/netdevice.h>
	20	#include <linux/random.h>
	21	#include <linux/jhash.h>
	22	#include <linux/skbuff.h>
	23	#include <linux/slab.h>
	24	#include <linux/export.h>
	25
	26	#include <net/ieee802154_netdev.h>
cefc8c8a	27	#include <net/6lowpan.h>
70b095c8	28	#include <net/ipv6_frag.h>
7240cdec AA	29	#include <net/inet_frag.h>
7240cdec AA	30
8691ee59	31	#include "6lowpan_i.h"
7240cdec	32
d4ad4d22 NA	33	static const char lowpan_frags_cache_name[] = "lowpan-frags";
d4ad4d22 NA	34
7240cdec AA	35	static struct inet_frags lowpan_frags;
	36
	37	static int lowpan_frag_reasm(struct lowpan_frag_queue *fq,
f4606583	38	struct sk_buff prev, struct net_device ldev);
7240cdec	39
36c77782	40	static void lowpan_frag_init(struct inet_frag_queue q, const void a)
7240cdec	41	{
648700f7	42	const struct frag_lowpan_compare_key *key = a;
7240cdec	43
648700f7 ED	44	BUILD_BUG_ON(sizeof(*key) > sizeof(q->key));
648700f7 ED	45	memcpy(&q->key, key, sizeof(*key));
7240cdec	46	}
7240cdec	47
78802011	48	static void lowpan_frag_expire(struct timer_list *t)
7240cdec	49	{
78802011	50	struct inet_frag_queue *frag = from_timer(frag, t, timer);
7240cdec	51	struct frag_queue *fq;
7240cdec	52
78802011	53	fq = container_of(frag, struct frag_queue, q);
7240cdec	54
17794326 FW	55	spin_lock(&fq->q.lock);
17794326 FW	56
06aa8b8a	57	if (fq->q.flags & INET_FRAG_COMPLETE)
17794326 FW	58	goto out;
17794326 FW	59
093ba729	60	inet_frag_kill(&fq->q);
17794326 FW	61	out:
17794326 FW	62	spin_unlock(&fq->q.lock);
093ba729	63	inet_frag_put(&fq->q);
7240cdec AA	64	}
	65
	66	static inline struct lowpan_frag_queue *
72a5e6bb	67	fq_find(struct net net, const struct lowpan_802154_cb cb,
ae531b94 PB	68	const struct ieee802154_addr *src,
ae531b94 PB	69	const struct ieee802154_addr *dst)
7240cdec	70	{
599018a7 LR	71	struct netns_ieee802154_lowpan *ieee802154_lowpan =
599018a7 LR	72	net_ieee802154_lowpan(net);
f18fa5de	73	struct frag_lowpan_compare_key key = {};
648700f7	74	struct inet_frag_queue *q;
7240cdec	75
f18fa5de AA	76	key.tag = cb->d_tag;
	77	key.d_size = cb->d_size;
	78	key.src = *src;
	79	key.dst = *dst;
	80
648700f7	81	q = inet_frag_find(&ieee802154_lowpan->frags, &key);
2d44ed22	82	if (!q)
7240cdec	83	return NULL;
2d44ed22	84
7240cdec AA	85	return container_of(q, struct lowpan_frag_queue, q);
	86	}
	87
	88	static int lowpan_frag_queue(struct lowpan_frag_queue *fq,
72a5e6bb	89	struct sk_buff *skb, u8 frag_type)
7240cdec AA	90	{
7240cdec AA	91	struct sk_buff prev, next;
f4606583	92	struct net_device *ldev;
7240cdec AA	93	int end, offset;
7240cdec AA	94
06aa8b8a	95	if (fq->q.flags & INET_FRAG_COMPLETE)
7240cdec AA	96	goto err;
7240cdec AA	97
72a5e6bb AA	98	offset = lowpan_802154_cb(skb)->d_offset << 3;
72a5e6bb AA	99	end = lowpan_802154_cb(skb)->d_size;
7240cdec AA	100
	101	/* Is this the final fragment? */
	102	if (offset + skb->len == end) {
	103	/* If we already have some bits beyond end
	104	* or have different end, the segment is corrupted.
	105	*/
	106	if (end < fq->q.len \|\|
06aa8b8a	107	((fq->q.flags & INET_FRAG_LAST_IN) && end != fq->q.len))
7240cdec	108	goto err;
06aa8b8a	109	fq->q.flags \|= INET_FRAG_LAST_IN;
7240cdec AA	110	fq->q.len = end;
	111	} else {
	112	if (end > fq->q.len) {
	113	/* Some bits beyond end -> corruption. */
06aa8b8a	114	if (fq->q.flags & INET_FRAG_LAST_IN)
7240cdec AA	115	goto err;
	116	fq->q.len = end;
	117	}
	118	}
	119
	120	/* Find out which fragments are in front and at the back of us
	121	* in the chain of fragments so far. We must know where to put
	122	* this fragment, right?
	123	*/
	124	prev = fq->q.fragments_tail;
72a5e6bb AA	125	if (!prev \|\|
	126	lowpan_802154_cb(prev)->d_offset <
	127	lowpan_802154_cb(skb)->d_offset) {
7240cdec AA	128	next = NULL;
	129	goto found;
	130	}
	131	prev = NULL;
	132	for (next = fq->q.fragments; next != NULL; next = next->next) {
72a5e6bb AA	133	if (lowpan_802154_cb(next)->d_offset >=
72a5e6bb AA	134	lowpan_802154_cb(skb)->d_offset)
7240cdec AA	135	break; /* bingo! */
	136	prev = next;
	137	}
	138
	139	found:
	140	/* Insert this fragment in the chain of fragments. */
	141	skb->next = next;
	142	if (!next)
	143	fq->q.fragments_tail = skb;
	144	if (prev)
	145	prev->next = skb;
	146	else
	147	fq->q.fragments = skb;
	148
f4606583 AA	149	ldev = skb->dev;
f4606583 AA	150	if (ldev)
7240cdec AA	151	skb->dev = NULL;
	152
	153	fq->q.stamp = skb->tstamp;
72a5e6bb	154	if (frag_type == LOWPAN_DISPATCH_FRAG1)
06aa8b8a	155	fq->q.flags \|= INET_FRAG_FIRST_IN;
72a5e6bb AA	156
72a5e6bb AA	157	fq->q.meat += skb->len;
0e60d245	158	add_frag_mem_limit(fq->q.net, skb->truesize);
7240cdec	159
06aa8b8a	160	if (fq->q.flags == (INET_FRAG_FIRST_IN \| INET_FRAG_LAST_IN) &&
7240cdec AA	161	fq->q.meat == fq->q.len) {
	162	int res;
	163	unsigned long orefdst = skb->_skb_refdst;
	164
	165	skb->_skb_refdst = 0UL;
f4606583	166	res = lowpan_frag_reasm(fq, prev, ldev);
7240cdec AA	167	skb->_skb_refdst = orefdst;
	168	return res;
	169	}
	170
7240cdec AA	171	return -1;
	172	err:
	173	kfree_skb(skb);
	174	return -1;
	175	}
	176
	177	/* Check if this packet is complete.
	178	* Returns NULL on failure by any reason, and pointer
	179	* to current nexthdr field in reassembled frame.
	180	*
	181	* It is called with locked fq, and caller must check that
	182	* queue is eligible for reassembly i.e. it is not COMPLETE,
	183	* the last and the first frames arrived and all the bits are here.
	184	*/
	185	static int lowpan_frag_reasm(struct lowpan_frag_queue fq, struct sk_buff prev,
f4606583	186	struct net_device *ldev)
7240cdec AA	187	{
	188	struct sk_buff fp, head = fq->q.fragments;
	189	int sum_truesize;
	190
093ba729	191	inet_frag_kill(&fq->q);
7240cdec AA	192
	193	/* Make the one we just received the head. */
	194	if (prev) {
	195	head = prev->next;
	196	fp = skb_clone(head, GFP_ATOMIC);
	197
	198	if (!fp)
	199	goto out_oom;
	200
	201	fp->next = head->next;
	202	if (!fp->next)
	203	fq->q.fragments_tail = fp;
	204	prev->next = fp;
	205
	206	skb_morph(head, fq->q.fragments);
	207	head->next = fq->q.fragments->next;
	208
	209	consume_skb(fq->q.fragments);
	210	fq->q.fragments = head;
	211	}
	212
	213	/* Head of list must not be cloned. */
	214	if (skb_unclone(head, GFP_ATOMIC))
	215	goto out_oom;
	216
	217	/* If the first fragment is fragmented itself, we split
	218	* it to two chunks: the first with data and paged part
	219	* and the second, holding only fragments.
	220	*/
	221	if (skb_has_frag_list(head)) {
	222	struct sk_buff *clone;
	223	int i, plen = 0;
	224
	225	clone = alloc_skb(0, GFP_ATOMIC);
	226	if (!clone)
	227	goto out_oom;
	228	clone->next = head->next;
	229	head->next = clone;
	230	skb_shinfo(clone)->frag_list = skb_shinfo(head)->frag_list;
	231	skb_frag_list_init(head);
	232	for (i = 0; i < skb_shinfo(head)->nr_frags; i++)
	233	plen += skb_frag_size(&skb_shinfo(head)->frags[i]);
	234	clone->len = head->data_len - plen;
	235	clone->data_len = clone->len;
	236	head->data_len -= clone->len;
	237	head->len -= clone->len;
0e60d245	238	add_frag_mem_limit(fq->q.net, clone->truesize);
7240cdec AA	239	}
	240
	241	WARN_ON(head == NULL);
	242
	243	sum_truesize = head->truesize;
	244	for (fp = head->next; fp;) {
	245	bool headstolen;
	246	int delta;
	247	struct sk_buff *next = fp->next;
	248
	249	sum_truesize += fp->truesize;
	250	if (skb_try_coalesce(head, fp, &headstolen, &delta)) {
	251	kfree_skb_partial(fp, headstolen);
	252	} else {
	253	if (!skb_shinfo(head)->frag_list)
	254	skb_shinfo(head)->frag_list = fp;
	255	head->data_len += fp->len;
	256	head->len += fp->len;
	257	head->truesize += fp->truesize;
	258	}
	259	fp = next;
	260	}
0e60d245	261	sub_frag_mem_limit(fq->q.net, sum_truesize);
7240cdec	262
a8305bff	263	skb_mark_not_on_list(head);
f4606583	264	head->dev = ldev;
7240cdec AA	265	head->tstamp = fq->q.stamp;
	266
	267	fq->q.fragments = NULL;
	268	fq->q.fragments_tail = NULL;
	269
	270	return 1;
	271	out_oom:
	272	net_dbg_ratelimited("lowpan_frag_reasm: no memory for reassembly\n");
	273	return -1;
	274	}
	275
72a5e6bb AA	276	static int lowpan_frag_rx_handlers_result(struct sk_buff *skb,
	277	lowpan_rx_result res)
	278	{
	279	switch (res) {
	280	case RX_QUEUED:
	281	return NET_RX_SUCCESS;
	282	case RX_CONTINUE:
	283	/* nobody cared about this packet */
	284	net_warn_ratelimited("%s: received unknown dispatch\n",
	285	__func__);
	286
	287	/* fall-through */
	288	default:
	289	/* all others failure */
	290	return NET_RX_DROP;
	291	}
	292	}
	293
	294	static lowpan_rx_result lowpan_frag_rx_h_iphc(struct sk_buff *skb)
	295	{
	296	int ret;
	297
	298	if (!lowpan_is_iphc(*skb_network_header(skb)))
	299	return RX_CONTINUE;
	300
	301	ret = lowpan_iphc_decompress(skb);
	302	if (ret < 0)
	303	return RX_DROP;
	304
	305	return RX_QUEUED;
	306	}
	307
	308	static int lowpan_invoke_frag_rx_handlers(struct sk_buff *skb)
	309	{
	310	lowpan_rx_result res;
	311
	312	#define CALL_RXH(rxh) \
	313	do { \
	314	res = rxh(skb); \
	315	if (res != RX_CONTINUE) \
	316	goto rxh_next; \
	317	} while (0)
	318
	319	/* likely at first */
	320	CALL_RXH(lowpan_frag_rx_h_iphc);
	321	CALL_RXH(lowpan_rx_h_ipv6);
	322
	323	rxh_next:
	324	return lowpan_frag_rx_handlers_result(skb, res);
	325	#undef CALL_RXH
	326	}
	327
	328	#define LOWPAN_FRAG_DGRAM_SIZE_HIGH_MASK 0x07
	329	#define LOWPAN_FRAG_DGRAM_SIZE_HIGH_SHIFT 8
	330
	331	static int lowpan_get_cb(struct sk_buff *skb, u8 frag_type,
	332	struct lowpan_802154_cb *cb)
7240cdec AA	333	{
7240cdec AA	334	bool fail;
72a5e6bb	335	u8 high = 0, low = 0;
f870b8c6	336	__be16 d_tag = 0;
7240cdec	337
72a5e6bb	338	fail = lowpan_fetch_skb(skb, &high, 1);
7240cdec	339	fail \|= lowpan_fetch_skb(skb, &low, 1);
72a5e6bb AA	340	/* remove the dispatch value and use first three bits as high value
	341	* for the datagram size
	342	*/
	343	cb->d_size = (high & LOWPAN_FRAG_DGRAM_SIZE_HIGH_MASK) <<
	344	LOWPAN_FRAG_DGRAM_SIZE_HIGH_SHIFT \| low;
f870b8c6	345	fail \|= lowpan_fetch_skb(skb, &d_tag, 2);
72a5e6bb	346	cb->d_tag = ntohs(d_tag);
7240cdec AA	347
7240cdec AA	348	if (frag_type == LOWPAN_DISPATCH_FRAGN) {
72a5e6bb	349	fail \|= lowpan_fetch_skb(skb, &cb->d_offset, 1);
7240cdec AA	350	} else {
7240cdec AA	351	skb_reset_network_header(skb);
72a5e6bb AA	352	cb->d_offset = 0;
	353	/* check if datagram_size has ipv6hdr on FRAG1 */
	354	fail \|= cb->d_size < sizeof(struct ipv6hdr);
	355	/* check if we can dereference the dispatch value */
	356	fail \|= !skb->len;
7240cdec AA	357	}
	358
	359	if (unlikely(fail))
	360	return -EIO;
	361
	362	return 0;
	363	}
	364
72a5e6bb	365	int lowpan_frag_rcv(struct sk_buff *skb, u8 frag_type)
7240cdec AA	366	{
	367	struct lowpan_frag_queue *fq;
	368	struct net *net = dev_net(skb->dev);
72a5e6bb	369	struct lowpan_802154_cb *cb = lowpan_802154_cb(skb);
f18fa5de	370	struct ieee802154_hdr hdr = {};
7240cdec AA	371	int err;
7240cdec AA	372
72a5e6bb AA	373	if (ieee802154_hdr_peek_addrs(skb, &hdr) < 0)
72a5e6bb AA	374	goto err;
ae531b94	375
72a5e6bb	376	err = lowpan_get_cb(skb, frag_type, cb);
7240cdec AA	377	if (err < 0)
	378	goto err;
	379
72a5e6bb AA	380	if (frag_type == LOWPAN_DISPATCH_FRAG1) {
	381	err = lowpan_invoke_frag_rx_handlers(skb);
	382	if (err == NET_RX_DROP)
	383	goto err;
	384	}
	385
	386	if (cb->d_size > IPV6_MIN_MTU) {
6697dabe	387	net_warn_ratelimited("lowpan_frag_rcv: datagram size exceeds MTU\n");
7240cdec	388	goto err;
6697dabe	389	}
7240cdec	390
72a5e6bb	391	fq = fq_find(net, cb, &hdr.source, &hdr.dest);
7240cdec AA	392	if (fq != NULL) {
7240cdec AA	393	int ret;
4710d806	394
7240cdec AA	395	spin_lock(&fq->q.lock);
	396	ret = lowpan_frag_queue(fq, skb, frag_type);
	397	spin_unlock(&fq->q.lock);
	398
093ba729	399	inet_frag_put(&fq->q);
7240cdec AA	400	return ret;
	401	}
	402
	403	err:
	404	kfree_skb(skb);
	405	return -1;
	406	}
7240cdec AA	407
7240cdec AA	408	#ifdef CONFIG_SYSCTL
1bab4c75	409
7240cdec AA	410	static struct ctl_table lowpan_frags_ns_ctl_table[] = {
	411	{
	412	.procname = "6lowpanfrag_high_thresh",
	413	.data = &init_net.ieee802154_lowpan.frags.high_thresh,
3e67f106	414	.maxlen = sizeof(unsigned long),
7240cdec	415	.mode = 0644,
3e67f106	416	.proc_handler = proc_doulongvec_minmax,
1bab4c75	417	.extra1 = &init_net.ieee802154_lowpan.frags.low_thresh
7240cdec AA	418	},
	419	{
	420	.procname = "6lowpanfrag_low_thresh",
	421	.data = &init_net.ieee802154_lowpan.frags.low_thresh,
3e67f106	422	.maxlen = sizeof(unsigned long),
7240cdec	423	.mode = 0644,
3e67f106	424	.proc_handler = proc_doulongvec_minmax,
1bab4c75	425	.extra2 = &init_net.ieee802154_lowpan.frags.high_thresh
7240cdec AA	426	},
	427	{
	428	.procname = "6lowpanfrag_time",
	429	.data = &init_net.ieee802154_lowpan.frags.timeout,
	430	.maxlen = sizeof(int),
	431	.mode = 0644,
	432	.proc_handler = proc_dointvec_jiffies,
	433	},
7240cdec AA	434	{ }
	435	};
	436
e3a57d18 FW	437	/* secret interval has been deprecated */
e3a57d18 FW	438	static int lowpan_frags_secret_interval_unused;
7240cdec AA	439	static struct ctl_table lowpan_frags_ctl_table[] = {
	440	{
	441	.procname = "6lowpanfrag_secret_interval",
e3a57d18	442	.data = &lowpan_frags_secret_interval_unused,
7240cdec AA	443	.maxlen = sizeof(int),
	444	.mode = 0644,
	445	.proc_handler = proc_dointvec_jiffies,
	446	},
	447	{ }
	448	};
	449
	450	static int __net_init lowpan_frags_ns_sysctl_register(struct net *net)
	451	{
	452	struct ctl_table *table;
	453	struct ctl_table_header *hdr;
599018a7 LR	454	struct netns_ieee802154_lowpan *ieee802154_lowpan =
599018a7 LR	455	net_ieee802154_lowpan(net);
7240cdec AA	456
	457	table = lowpan_frags_ns_ctl_table;
	458	if (!net_eq(net, &init_net)) {
	459	table = kmemdup(table, sizeof(lowpan_frags_ns_ctl_table),
	460	GFP_KERNEL);
	461	if (table == NULL)
	462	goto err_alloc;
	463
599018a7	464	table[0].data = &ieee802154_lowpan->frags.high_thresh;
1bab4c75	465	table[0].extra1 = &ieee802154_lowpan->frags.low_thresh;
599018a7	466	table[1].data = &ieee802154_lowpan->frags.low_thresh;
1bab4c75	467	table[1].extra2 = &ieee802154_lowpan->frags.high_thresh;
599018a7	468	table[2].data = &ieee802154_lowpan->frags.timeout;
7240cdec AA	469
	470	/* Don't export sysctls to unprivileged users */
	471	if (net->user_ns != &init_user_ns)
	472	table[0].procname = NULL;
	473	}
	474
	475	hdr = register_net_sysctl(net, "net/ieee802154/6lowpan", table);
	476	if (hdr == NULL)
	477	goto err_reg;
	478
599018a7	479	ieee802154_lowpan->sysctl.frags_hdr = hdr;
7240cdec AA	480	return 0;
	481
	482	err_reg:
	483	if (!net_eq(net, &init_net))
	484	kfree(table);
	485	err_alloc:
	486	return -ENOMEM;
	487	}
	488
	489	static void __net_exit lowpan_frags_ns_sysctl_unregister(struct net *net)
	490	{
	491	struct ctl_table *table;
599018a7 LR	492	struct netns_ieee802154_lowpan *ieee802154_lowpan =
599018a7 LR	493	net_ieee802154_lowpan(net);
7240cdec	494
599018a7 LR	495	table = ieee802154_lowpan->sysctl.frags_hdr->ctl_table_arg;
599018a7 LR	496	unregister_net_sysctl_table(ieee802154_lowpan->sysctl.frags_hdr);
7240cdec AA	497	if (!net_eq(net, &init_net))
	498	kfree(table);
	499	}
	500
	501	static struct ctl_table_header *lowpan_ctl_header;
	502
3243acd3	503	static int __init lowpan_frags_sysctl_register(void)
7240cdec AA	504	{
	505	lowpan_ctl_header = register_net_sysctl(&init_net,
	506	"net/ieee802154/6lowpan",
	507	lowpan_frags_ctl_table);
	508	return lowpan_ctl_header == NULL ? -ENOMEM : 0;
	509	}
	510
	511	static void lowpan_frags_sysctl_unregister(void)
	512	{
	513	unregister_net_sysctl_table(lowpan_ctl_header);
	514	}
	515	#else
f0a0c1ce	516	static inline int lowpan_frags_ns_sysctl_register(struct net *net)
7240cdec AA	517	{
	518	return 0;
	519	}
	520
	521	static inline void lowpan_frags_ns_sysctl_unregister(struct net *net)
	522	{
	523	}
	524
f0a0c1ce	525	static inline int __init lowpan_frags_sysctl_register(void)
7240cdec AA	526	{
	527	return 0;
	528	}
	529
	530	static inline void lowpan_frags_sysctl_unregister(void)
	531	{
	532	}
	533	#endif
	534
	535	static int __net_init lowpan_frags_init_net(struct net *net)
	536	{
599018a7 LR	537	struct netns_ieee802154_lowpan *ieee802154_lowpan =
599018a7 LR	538	net_ieee802154_lowpan(net);
787bea77	539	int res;
7240cdec	540
599018a7 LR	541	ieee802154_lowpan->frags.high_thresh = IPV6_FRAG_HIGH_THRESH;
	542	ieee802154_lowpan->frags.low_thresh = IPV6_FRAG_LOW_THRESH;
	543	ieee802154_lowpan->frags.timeout = IPV6_FRAG_TIMEOUT;
093ba729	544	ieee802154_lowpan->frags.f = &lowpan_frags;
599018a7	545
787bea77 ED	546	res = inet_frags_init_net(&ieee802154_lowpan->frags);
	547	if (res < 0)
	548	return res;
	549	res = lowpan_frags_ns_sysctl_register(net);
	550	if (res < 0)
093ba729	551	inet_frags_exit_net(&ieee802154_lowpan->frags);
787bea77	552	return res;
7240cdec AA	553	}
	554
	555	static void __net_exit lowpan_frags_exit_net(struct net *net)
	556	{
599018a7 LR	557	struct netns_ieee802154_lowpan *ieee802154_lowpan =
	558	net_ieee802154_lowpan(net);
	559
7240cdec	560	lowpan_frags_ns_sysctl_unregister(net);
093ba729	561	inet_frags_exit_net(&ieee802154_lowpan->frags);
7240cdec AA	562	}
	563
	564	static struct pernet_operations lowpan_frags_ops = {
	565	.init = lowpan_frags_init_net,
	566	.exit = lowpan_frags_exit_net,
	567	};
	568
648700f7 ED	569	static u32 lowpan_key_hashfn(const void *data, u32 len, u32 seed)
	570	{
	571	return jhash2(data,
	572	sizeof(struct frag_lowpan_compare_key) / sizeof(u32), seed);
	573	}
	574
	575	static u32 lowpan_obj_hashfn(const void *data, u32 len, u32 seed)
	576	{
	577	const struct inet_frag_queue *fq = data;
	578
	579	return jhash2((const u32 *)&fq->key,
	580	sizeof(struct frag_lowpan_compare_key) / sizeof(u32), seed);
	581	}
	582
	583	static int lowpan_obj_cmpfn(struct rhashtable_compare_arg arg, const void ptr)
	584	{
	585	const struct frag_lowpan_compare_key *key = arg->key;
	586	const struct inet_frag_queue *fq = ptr;
	587
	588	return !!memcmp(&fq->key, key, sizeof(*key));
	589	}
	590
	591	static const struct rhashtable_params lowpan_rhash_params = {
	592	.head_offset = offsetof(struct inet_frag_queue, node),
	593	.hashfn = lowpan_key_hashfn,
	594	.obj_hashfn = lowpan_obj_hashfn,
	595	.obj_cmpfn = lowpan_obj_cmpfn,
	596	.automatic_shrinking = true,
	597	};
	598
7240cdec AA	599	int __init lowpan_net_frag_init(void)
	600	{
	601	int ret;
	602
7240cdec AA	603	lowpan_frags.constructor = lowpan_frag_init;
7240cdec AA	604	lowpan_frags.destructor = NULL;
7240cdec	605	lowpan_frags.qsize = sizeof(struct frag_queue);
7240cdec	606	lowpan_frags.frag_expire = lowpan_frag_expire;
d4ad4d22	607	lowpan_frags.frags_cache_name = lowpan_frags_cache_name;
648700f7	608	lowpan_frags.rhash_params = lowpan_rhash_params;
d4ad4d22 NA	609	ret = inet_frags_init(&lowpan_frags);
d4ad4d22 NA	610	if (ret)
807f1844	611	goto out;
37147652	612
807f1844 ED	613	ret = lowpan_frags_sysctl_register();
	614	if (ret)
	615	goto err_sysctl;
	616
	617	ret = register_pernet_subsys(&lowpan_frags_ops);
	618	if (ret)
	619	goto err_pernet;
	620	out:
37147652	621	return ret;
7240cdec AA	622	err_pernet:
7240cdec AA	623	lowpan_frags_sysctl_unregister();
807f1844 ED	624	err_sysctl:
807f1844 ED	625	inet_frags_fini(&lowpan_frags);
7240cdec AA	626	return ret;
	627	}
	628
	629	void lowpan_net_frag_exit(void)
	630	{
	631	inet_frags_fini(&lowpan_frags);
	632	lowpan_frags_sysctl_unregister();
	633	unregister_pernet_subsys(&lowpan_frags_ops);
	634	}