[linux.git] / net / xfrm / xfrm_input.c

/*
 * xfrm_input.c
 *
 * Changes:
 * 	YOSHIFUJI Hideaki @USAGI
 * 		Split up af-specific portion
 *
 */

#include <linux/slab.h>
#include <linux/module.h>
#include <linux/netdevice.h>
#include <net/dst.h>
#include <net/ip.h>
#include <net/xfrm.h>

static struct kmem_cache *secpath_cachep __read_mostly;

void __secpath_destroy(struct sec_path *sp)
{
	int i;
	for (i = 0; i < sp->len; i++)
		xfrm_state_put(sp->xvec[i]);
	kmem_cache_free(secpath_cachep, sp);
}
EXPORT_SYMBOL(__secpath_destroy);

struct sec_path *secpath_dup(struct sec_path *src)
{
	struct sec_path *sp;

	sp = kmem_cache_alloc(secpath_cachep, GFP_ATOMIC);
	if (!sp)
		return NULL;

	sp->len = 0;
	if (src) {
		int i;

		memcpy(sp, src, sizeof(*sp));
		for (i = 0; i < sp->len; i++)
			xfrm_state_hold(sp->xvec[i]);
	}
	atomic_set(&sp->refcnt, 1);
	return sp;
}
EXPORT_SYMBOL(secpath_dup);

/* Fetch spi and seq from ipsec header */

int xfrm_parse_spi(struct sk_buff *skb, u8 nexthdr, __be32 *spi, __be32 *seq)
{
	int offset, offset_seq;
	int hlen;

	switch (nexthdr) {
	case IPPROTO_AH:
		hlen = sizeof(struct ip_auth_hdr);
		offset = offsetof(struct ip_auth_hdr, spi);
		offset_seq = offsetof(struct ip_auth_hdr, seq_no);
		break;
	case IPPROTO_ESP:
		hlen = sizeof(struct ip_esp_hdr);
		offset = offsetof(struct ip_esp_hdr, spi);
		offset_seq = offsetof(struct ip_esp_hdr, seq_no);
		break;
	case IPPROTO_COMP:
		if (!pskb_may_pull(skb, sizeof(struct ip_comp_hdr)))
			return -EINVAL;
		*spi = htonl(ntohs(*(__be16*)(skb_transport_header(skb) + 2)));
		*seq = 0;
		return 0;
	default:
		return 1;
	}

	if (!pskb_may_pull(skb, hlen))
		return -EINVAL;

	*spi = *(__be32*)(skb_transport_header(skb) + offset);
	*seq = *(__be32*)(skb_transport_header(skb) + offset_seq);
	return 0;
}

int xfrm_prepare_input(struct xfrm_state *x, struct sk_buff *skb)
{
	struct xfrm_mode *inner_mode = x->inner_mode;
	int err;

	err = x->outer_mode->afinfo->extract_input(x, skb);
	if (err)
		return err;

	if (x->sel.family == AF_UNSPEC) {
		inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
		if (inner_mode == NULL)
			return -EAFNOSUPPORT;
	}

	skb->protocol = inner_mode->afinfo->eth_proto;
	return inner_mode->input2(x, skb);
}
EXPORT_SYMBOL(xfrm_prepare_input);

int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
{
	struct net *net = dev_net(skb->dev);
	int err;
	__be32 seq;
	__be32 seq_hi;
	struct xfrm_state *x;
	xfrm_address_t *daddr;
	struct xfrm_mode *inner_mode;
	unsigned int family;
	int decaps = 0;
	int async = 0;

	/* A negative encap_type indicates async resumption. */
	if (encap_type < 0) {
		async = 1;
		x = xfrm_input_state(skb);
		seq = XFRM_SKB_CB(skb)->seq.input.low;
		goto resume;
	}

	/* Allocate new secpath or COW existing one. */
	if (!skb->sp || atomic_read(&skb->sp->refcnt) != 1) {
		struct sec_path *sp;

		sp = secpath_dup(skb->sp);
		if (!sp) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
			goto drop;
		}
		if (skb->sp)
			secpath_put(skb->sp);
		skb->sp = sp;
	}

	daddr = (xfrm_address_t *)(skb_network_header(skb) +
				   XFRM_SPI_SKB_CB(skb)->daddroff);
	family = XFRM_SPI_SKB_CB(skb)->family;

	seq = 0;
	if (!spi && (err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
		XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
		goto drop;
	}

	do {
		if (skb->sp->len == XFRM_MAX_DEPTH) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
			goto drop;
		}

		x = xfrm_state_lookup(net, skb->mark, daddr, spi, nexthdr, family);
		if (x == NULL) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINNOSTATES);
			xfrm_audit_state_notfound(skb, family, spi, seq);
			goto drop;
		}

		skb->sp->xvec[skb->sp->len++] = x;

		spin_lock(&x->lock);
		if (unlikely(x->km.state != XFRM_STATE_VALID)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEINVALID);
			goto drop_unlock;
		}

		if ((x->encap ? x->encap->encap_type : 0) != encap_type) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMISMATCH);
			goto drop_unlock;
		}

		if (x->repl->check(x, skb, seq)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
			goto drop_unlock;
		}

		if (xfrm_state_check_expire(x)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEEXPIRED);
			goto drop_unlock;
		}

		spin_unlock(&x->lock);

		seq_hi = htonl(xfrm_replay_seqhi(x, seq));

		XFRM_SKB_CB(skb)->seq.input.low = seq;
		XFRM_SKB_CB(skb)->seq.input.hi = seq_hi;

		skb_dst_force(skb);

		nexthdr = x->type->input(x, skb);

		if (nexthdr == -EINPROGRESS)
			return 0;

resume:
		spin_lock(&x->lock);
		if (nexthdr <= 0) {
			if (nexthdr == -EBADMSG) {
				xfrm_audit_state_icvfail(x, skb,
							 x->type->proto);
				x->stats.integrity_failed++;
			}
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
			goto drop_unlock;
		}

		/* only the first xfrm gets the encap type */
		encap_type = 0;

		if (async && x->repl->recheck(x, skb, seq)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
			goto drop_unlock;
		}

		x->repl->advance(x, seq);

		x->curlft.bytes += skb->len;
		x->curlft.packets++;

		spin_unlock(&x->lock);

		XFRM_MODE_SKB_CB(skb)->protocol = nexthdr;

		inner_mode = x->inner_mode;

		if (x->sel.family == AF_UNSPEC) {
			inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
			if (inner_mode == NULL)
				goto drop;
		}

		if (inner_mode->input(x, skb)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
			goto drop;
		}

		if (x->outer_mode->flags & XFRM_MODE_FLAG_TUNNEL) {
			decaps = 1;
			break;
		}

		/*
		 * We need the inner address.  However, we only get here for
		 * transport mode so the outer address is identical.
		 */
		daddr = &x->id.daddr;
		family = x->outer_mode->afinfo->family;

		err = xfrm_parse_spi(skb, nexthdr, &spi, &seq);
		if (err < 0) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
			goto drop;
		}
	} while (!err);

	nf_reset(skb);

	if (decaps) {
		skb_dst_drop(skb);
		netif_rx(skb);
		return 0;
	} else {
		return x->inner_mode->afinfo->transport_finish(skb, async);
	}

drop_unlock:
	spin_unlock(&x->lock);
drop:
	kfree_skb(skb);
	return 0;
}
EXPORT_SYMBOL(xfrm_input);

int xfrm_input_resume(struct sk_buff *skb, int nexthdr)
{
	return xfrm_input(skb, nexthdr, 0, -1);
}
EXPORT_SYMBOL(xfrm_input_resume);

void __init xfrm_input_init(void)
{
	secpath_cachep = kmem_cache_create("secpath_cache",
					   sizeof(struct sec_path),
					   0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
					   NULL);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* xfrm_input.c
	3	*
	4	* Changes:
	5	* YOSHIFUJI Hideaki @USAGI
	6	* Split up af-specific portion
a716c119	7	*
1da177e4 LT	8	*/
	9
	10	#include <linux/slab.h>
	11	#include <linux/module.h>
716062fd HX	12	#include <linux/netdevice.h>
716062fd HX	13	#include <net/dst.h>
1da177e4 LT	14	#include <net/ip.h>
	15	#include <net/xfrm.h>
	16
e18b890b	17	static struct kmem_cache *secpath_cachep __read_mostly;
1da177e4 LT	18
	19	void __secpath_destroy(struct sec_path *sp)
	20	{
	21	int i;
	22	for (i = 0; i < sp->len; i++)
dbe5b4aa	23	xfrm_state_put(sp->xvec[i]);
1da177e4 LT	24	kmem_cache_free(secpath_cachep, sp);
	25	}
	26	EXPORT_SYMBOL(__secpath_destroy);
	27
	28	struct sec_path secpath_dup(struct sec_path src)
	29	{
	30	struct sec_path *sp;
	31
54e6ecb2	32	sp = kmem_cache_alloc(secpath_cachep, GFP_ATOMIC);
1da177e4 LT	33	if (!sp)
	34	return NULL;
	35
	36	sp->len = 0;
	37	if (src) {
	38	int i;
	39
	40	memcpy(sp, src, sizeof(*sp));
	41	for (i = 0; i < sp->len; i++)
dbe5b4aa	42	xfrm_state_hold(sp->xvec[i]);
1da177e4 LT	43	}
	44	atomic_set(&sp->refcnt, 1);
	45	return sp;
	46	}
	47	EXPORT_SYMBOL(secpath_dup);
	48
	49	/* Fetch spi and seq from ipsec header */
	50
6067b2ba	51	int xfrm_parse_spi(struct sk_buff skb, u8 nexthdr, __be32 spi, __be32 *seq)
1da177e4 LT	52	{
1da177e4 LT	53	int offset, offset_seq;
44072500	54	int hlen;
1da177e4 LT	55
	56	switch (nexthdr) {
	57	case IPPROTO_AH:
44072500	58	hlen = sizeof(struct ip_auth_hdr);
1da177e4 LT	59	offset = offsetof(struct ip_auth_hdr, spi);
	60	offset_seq = offsetof(struct ip_auth_hdr, seq_no);
	61	break;
	62	case IPPROTO_ESP:
44072500	63	hlen = sizeof(struct ip_esp_hdr);
1da177e4 LT	64	offset = offsetof(struct ip_esp_hdr, spi);
	65	offset_seq = offsetof(struct ip_esp_hdr, seq_no);
	66	break;
	67	case IPPROTO_COMP:
	68	if (!pskb_may_pull(skb, sizeof(struct ip_comp_hdr)))
	69	return -EINVAL;
9c70220b	70	spi = htonl(ntohs((__be16*)(skb_transport_header(skb) + 2)));
1da177e4 LT	71	*seq = 0;
	72	return 0;
	73	default:
	74	return 1;
	75	}
	76
44072500	77	if (!pskb_may_pull(skb, hlen))
1da177e4 LT	78	return -EINVAL;
1da177e4 LT	79
9c70220b ACM	80	spi = (__be32*)(skb_transport_header(skb) + offset);
9c70220b ACM	81	seq = (__be32*)(skb_transport_header(skb) + offset_seq);
1da177e4 LT	82	return 0;
1da177e4 LT	83	}
1da177e4	84
227620e2 HX	85	int xfrm_prepare_input(struct xfrm_state x, struct sk_buff skb)
227620e2 HX	86	{
df9dcb45	87	struct xfrm_mode *inner_mode = x->inner_mode;
227620e2 HX	88	int err;
	89
	90	err = x->outer_mode->afinfo->extract_input(x, skb);
	91	if (err)
	92	return err;
	93
df9dcb45 KM	94	if (x->sel.family == AF_UNSPEC) {
	95	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
	96	if (inner_mode == NULL)
	97	return -EAFNOSUPPORT;
	98	}
	99
	100	skb->protocol = inner_mode->afinfo->eth_proto;
	101	return inner_mode->input2(x, skb);
227620e2 HX	102	}
	103	EXPORT_SYMBOL(xfrm_prepare_input);
	104
716062fd HX	105	int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
716062fd HX	106	{
bd235e3c	107	struct net *net = dev_net(skb->dev);
716062fd HX	108	int err;
716062fd HX	109	__be32 seq;
2cd08467	110	__be32 seq_hi;
716062fd	111	struct xfrm_state *x;
1bf06cd2	112	xfrm_address_t *daddr;
df9dcb45	113	struct xfrm_mode *inner_mode;
2fcb45b6	114	unsigned int family;
716062fd	115	int decaps = 0;
1bf06cd2 HX	116	int async = 0;
	117
	118	/* A negative encap_type indicates async resumption. */
	119	if (encap_type < 0) {
	120	async = 1;
00501121	121	x = xfrm_input_state(skb);
1ce3644a	122	seq = XFRM_SKB_CB(skb)->seq.input.low;
1bf06cd2 HX	123	goto resume;
1bf06cd2 HX	124	}
716062fd	125
b2aa5e9d HX	126	/* Allocate new secpath or COW existing one. */
	127	if (!skb->sp \|\| atomic_read(&skb->sp->refcnt) != 1) {
	128	struct sec_path *sp;
	129
	130	sp = secpath_dup(skb->sp);
0aa64774	131	if (!sp) {
59c9940e	132	XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
b2aa5e9d	133	goto drop;
0aa64774	134	}
b2aa5e9d HX	135	if (skb->sp)
	136	secpath_put(skb->sp);
	137	skb->sp = sp;
	138	}
	139
1bf06cd2 HX	140	daddr = (xfrm_address_t *)(skb_network_header(skb) +
1bf06cd2 HX	141	XFRM_SPI_SKB_CB(skb)->daddroff);
2fcb45b6	142	family = XFRM_SPI_SKB_CB(skb)->family;
1bf06cd2	143
716062fd	144	seq = 0;
0aa64774	145	if (!spi && (err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
59c9940e	146	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
716062fd	147	goto drop;
0aa64774	148	}
716062fd HX	149
716062fd HX	150	do {
0aa64774	151	if (skb->sp->len == XFRM_MAX_DEPTH) {
59c9940e	152	XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
716062fd	153	goto drop;
0aa64774	154	}
716062fd	155
bd55775c	156	x = xfrm_state_lookup(net, skb->mark, daddr, spi, nexthdr, family);
0aa64774	157	if (x == NULL) {
59c9940e	158	XFRM_INC_STATS(net, LINUX_MIB_XFRMINNOSTATES);
afeb14b4	159	xfrm_audit_state_notfound(skb, family, spi, seq);
716062fd	160	goto drop;
0aa64774	161	}
716062fd	162
b2aa5e9d HX	163	skb->sp->xvec[skb->sp->len++] = x;
b2aa5e9d HX	164
716062fd	165	spin_lock(&x->lock);
0aa64774	166	if (unlikely(x->km.state != XFRM_STATE_VALID)) {
59c9940e	167	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEINVALID);
716062fd	168	goto drop_unlock;
0aa64774	169	}
716062fd	170
3de77cf2 DM	171	if ((x->encap ? x->encap->encap_type : 0) != encap_type) {
	172	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMISMATCH);
	173	goto drop_unlock;
	174	}
	175
36ae0148	176	if (x->repl->check(x, skb, seq)) {
59c9940e	177	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
716062fd	178	goto drop_unlock;
0aa64774	179	}
716062fd	180
0aa64774	181	if (xfrm_state_check_expire(x)) {
59c9940e	182	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEEXPIRED);
716062fd	183	goto drop_unlock;
0aa64774	184	}
716062fd	185
0ebea8ef HX	186	spin_unlock(&x->lock);
0ebea8ef HX	187
2cd08467 SK	188	seq_hi = htonl(xfrm_replay_seqhi(x, seq));
2cd08467 SK	189
1ce3644a	190	XFRM_SKB_CB(skb)->seq.input.low = seq;
2cd08467	191	XFRM_SKB_CB(skb)->seq.input.hi = seq_hi;
1bf06cd2	192
3bc07321 SK	193	skb_dst_force(skb);
3bc07321 SK	194
716062fd	195	nexthdr = x->type->input(x, skb);
0ebea8ef	196
1bf06cd2 HX	197	if (nexthdr == -EINPROGRESS)
	198	return 0;
	199
	200	resume:
0ebea8ef	201	spin_lock(&x->lock);
668dc8af	202	if (nexthdr <= 0) {
9dd3245a HX	203	if (nexthdr == -EBADMSG) {
	204	xfrm_audit_state_icvfail(x, skb,
	205	x->type->proto);
668dc8af	206	x->stats.integrity_failed++;
9dd3245a	207	}
59c9940e	208	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
716062fd	209	goto drop_unlock;
668dc8af	210	}
716062fd	211
716062fd HX	212	/* only the first xfrm gets the encap type */
	213	encap_type = 0;
	214
3b59df46	215	if (async && x->repl->recheck(x, skb, seq)) {
bcf66bf5 SK	216	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
	217	goto drop_unlock;
	218	}
	219
9fdc4883	220	x->repl->advance(x, seq);
716062fd HX	221
	222	x->curlft.bytes += skb->len;
	223	x->curlft.packets++;
	224
	225	spin_unlock(&x->lock);
	226
60d5fcfb HX	227	XFRM_MODE_SKB_CB(skb)->protocol = nexthdr;
60d5fcfb HX	228
df9dcb45 KM	229	inner_mode = x->inner_mode;
	230
	231	if (x->sel.family == AF_UNSPEC) {
	232	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
	233	if (inner_mode == NULL)
	234	goto drop;
	235	}
	236
	237	if (inner_mode->input(x, skb)) {
59c9940e	238	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
716062fd	239	goto drop;
0aa64774	240	}
716062fd HX	241
	242	if (x->outer_mode->flags & XFRM_MODE_FLAG_TUNNEL) {
	243	decaps = 1;
	244	break;
	245	}
	246
1bf06cd2 HX	247	/*
	248	* We need the inner address. However, we only get here for
	249	* transport mode so the outer address is identical.
	250	*/
	251	daddr = &x->id.daddr;
2fcb45b6	252	family = x->outer_mode->afinfo->family;
1bf06cd2	253
716062fd	254	err = xfrm_parse_spi(skb, nexthdr, &spi, &seq);
0aa64774	255	if (err < 0) {
59c9940e	256	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
716062fd	257	goto drop;
0aa64774	258	}
716062fd HX	259	} while (!err);
716062fd HX	260
716062fd HX	261	nf_reset(skb);
	262
	263	if (decaps) {
adf30907	264	skb_dst_drop(skb);
716062fd HX	265	netif_rx(skb);
	266	return 0;
	267	} else {
1bf06cd2	268	return x->inner_mode->afinfo->transport_finish(skb, async);
716062fd HX	269	}
	270
	271	drop_unlock:
	272	spin_unlock(&x->lock);
716062fd	273	drop:
716062fd HX	274	kfree_skb(skb);
	275	return 0;
	276	}
	277	EXPORT_SYMBOL(xfrm_input);
	278
1bf06cd2 HX	279	int xfrm_input_resume(struct sk_buff *skb, int nexthdr)
	280	{
	281	return xfrm_input(skb, nexthdr, 0, -1);
	282	}
	283	EXPORT_SYMBOL(xfrm_input_resume);
	284
1da177e4 LT	285	void __init xfrm_input_init(void)
	286	{
	287	secpath_cachep = kmem_cache_create("secpath_cache",
	288	sizeof(struct sec_path),
e5d679f3	289	0, SLAB_HWCACHE_ALIGN\|SLAB_PANIC,
20c2df83	290	NULL);
1da177e4	291	}