[linux-2.6-block.git] / net / xfrm / xfrm_input.c

// SPDX-License-Identifier: GPL-2.0
/*
 * xfrm_input.c
 *
 * Changes:
 * 	YOSHIFUJI Hideaki @USAGI
 * 		Split up af-specific portion
 *
 */

#include <linux/bottom_half.h>
#include <linux/cache.h>
#include <linux/interrupt.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/netdevice.h>
#include <linux/percpu.h>
#include <net/dst.h>
#include <net/ip.h>
#include <net/xfrm.h>
#include <net/ip_tunnels.h>
#include <net/ip6_tunnel.h>

struct xfrm_trans_tasklet {
	struct tasklet_struct tasklet;
	struct sk_buff_head queue;
};

struct xfrm_trans_cb {
	union {
		struct inet_skb_parm	h4;
#if IS_ENABLED(CONFIG_IPV6)
		struct inet6_skb_parm	h6;
#endif
	} header;
	int (*finish)(struct net *net, struct sock *sk, struct sk_buff *skb);
};

#define XFRM_TRANS_SKB_CB(__skb) ((struct xfrm_trans_cb *)&((__skb)->cb[0]))

static struct kmem_cache *secpath_cachep __ro_after_init;

static DEFINE_SPINLOCK(xfrm_input_afinfo_lock);
static struct xfrm_input_afinfo const __rcu *xfrm_input_afinfo[AF_INET6 + 1];

static struct gro_cells gro_cells;
static struct net_device xfrm_napi_dev;

static DEFINE_PER_CPU(struct xfrm_trans_tasklet, xfrm_trans_tasklet);

int xfrm_input_register_afinfo(const struct xfrm_input_afinfo *afinfo)
{
	int err = 0;

	if (WARN_ON(afinfo->family >= ARRAY_SIZE(xfrm_input_afinfo)))
		return -EAFNOSUPPORT;

	spin_lock_bh(&xfrm_input_afinfo_lock);
	if (unlikely(xfrm_input_afinfo[afinfo->family] != NULL))
		err = -EEXIST;
	else
		rcu_assign_pointer(xfrm_input_afinfo[afinfo->family], afinfo);
	spin_unlock_bh(&xfrm_input_afinfo_lock);
	return err;
}
EXPORT_SYMBOL(xfrm_input_register_afinfo);

int xfrm_input_unregister_afinfo(const struct xfrm_input_afinfo *afinfo)
{
	int err = 0;

	spin_lock_bh(&xfrm_input_afinfo_lock);
	if (likely(xfrm_input_afinfo[afinfo->family] != NULL)) {
		if (unlikely(xfrm_input_afinfo[afinfo->family] != afinfo))
			err = -EINVAL;
		else
			RCU_INIT_POINTER(xfrm_input_afinfo[afinfo->family], NULL);
	}
	spin_unlock_bh(&xfrm_input_afinfo_lock);
	synchronize_rcu();
	return err;
}
EXPORT_SYMBOL(xfrm_input_unregister_afinfo);

static const struct xfrm_input_afinfo *xfrm_input_get_afinfo(unsigned int family)
{
	const struct xfrm_input_afinfo *afinfo;

	if (WARN_ON_ONCE(family >= ARRAY_SIZE(xfrm_input_afinfo)))
		return NULL;

	rcu_read_lock();
	afinfo = rcu_dereference(xfrm_input_afinfo[family]);
	if (unlikely(!afinfo))
		rcu_read_unlock();
	return afinfo;
}

static int xfrm_rcv_cb(struct sk_buff *skb, unsigned int family, u8 protocol,
		       int err)
{
	int ret;
	const struct xfrm_input_afinfo *afinfo = xfrm_input_get_afinfo(family);

	if (!afinfo)
		return -EAFNOSUPPORT;

	ret = afinfo->callback(skb, protocol, err);
	rcu_read_unlock();

	return ret;
}

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;
	sp->olen = 0;

	memset(sp->ovec, 0, sizeof(sp->ovec));

	if (src) {
		int i;

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

struct sec_path *secpath_set(struct sk_buff *skb)
{
	struct sec_path *sp = skb->sp;

	/* Allocate new secpath or COW existing one. */
	if (!sp || refcount_read(&sp->refcnt) != 1) {
		sp = secpath_dup(skb->sp);
		if (!sp)
			return NULL;

		if (skb->sp)
			secpath_put(skb->sp);
		skb->sp = sp;
	}

	return sp;
}
EXPORT_SYMBOL(secpath_set);

/* 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;
}
EXPORT_SYMBOL(xfrm_parse_spi);

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 = NULL;
	xfrm_address_t *daddr;
	struct xfrm_mode *inner_mode;
	u32 mark = skb->mark;
	unsigned int family = AF_UNSPEC;
	int decaps = 0;
	int async = 0;
	bool xfrm_gro = false;
	bool crypto_done = false;
	struct xfrm_offload *xo = xfrm_offload(skb);
	struct sec_path *sp;

	if (encap_type < 0) {
		x = xfrm_input_state(skb);

		if (unlikely(x->km.state != XFRM_STATE_VALID)) {
			if (x->km.state == XFRM_STATE_ACQ)
				XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
			else
				XFRM_INC_STATS(net,
					       LINUX_MIB_XFRMINSTATEINVALID);
			goto drop;
		}

		family = x->outer_mode->afinfo->family;

		/* An encap_type of -1 indicates async resumption. */
		if (encap_type == -1) {
			async = 1;
			seq = XFRM_SKB_CB(skb)->seq.input.low;
			goto resume;
		}

		/* encap_type < -1 indicates a GRO call. */
		encap_type = 0;
		seq = XFRM_SPI_SKB_CB(skb)->seq;

		if (xo && (xo->flags & CRYPTO_DONE)) {
			crypto_done = true;
			family = XFRM_SPI_SKB_CB(skb)->family;

			if (!(xo->status & CRYPTO_SUCCESS)) {
				if (xo->status &
				    (CRYPTO_TRANSPORT_AH_AUTH_FAILED |
				     CRYPTO_TRANSPORT_ESP_AUTH_FAILED |
				     CRYPTO_TUNNEL_AH_AUTH_FAILED |
				     CRYPTO_TUNNEL_ESP_AUTH_FAILED)) {

					xfrm_audit_state_icvfail(x, skb,
								 x->type->proto);
					x->stats.integrity_failed++;
					XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
					goto drop;
				}

				if (xo->status & CRYPTO_INVALID_PROTOCOL) {
					XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
					goto drop;
				}

				XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
				goto drop;
			}

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

		goto lock;
	}

	family = XFRM_SPI_SKB_CB(skb)->family;

	/* if tunnel is present override skb->mark value with tunnel i_key */
	switch (family) {
	case AF_INET:
		if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4)
			mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4->parms.i_key);
		break;
	case AF_INET6:
		if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6)
			mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6->parms.i_key);
		break;
	}

	sp = secpath_set(skb);
	if (!sp) {
		XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
		goto drop;
	}

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

	daddr = (xfrm_address_t *)(skb_network_header(skb) +
				   XFRM_SPI_SKB_CB(skb)->daddroff);
	do {
		sp = skb_sec_path(skb);

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

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

		skb->mark = xfrm_smark_get(skb->mark, x);

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

lock:
		spin_lock(&x->lock);

		if (unlikely(x->km.state != XFRM_STATE_VALID)) {
			if (x->km.state == XFRM_STATE_ACQ)
				XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
			else
				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);

		if (xfrm_tunnel_check(skb, x, family)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
			goto drop;
		}

		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);
		dev_hold(skb->dev);

		if (crypto_done)
			nexthdr = x->type_offload->input_tail(x, skb);
		else
			nexthdr = x->type->input(x, skb);

		if (nexthdr == -EINPROGRESS)
			return 0;
resume:
		dev_put(skb->dev);

		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) {
				XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
				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;
		}
		crypto_done = false;
	} while (!err);

	err = xfrm_rcv_cb(skb, family, x->type->proto, 0);
	if (err)
		goto drop;

	nf_reset(skb);

	if (decaps) {
		sp = skb_sec_path(skb);
		if (sp)
			sp->olen = 0;
		skb_dst_drop(skb);
		gro_cells_receive(&gro_cells, skb);
		return 0;
	} else {
		xo = xfrm_offload(skb);
		if (xo)
			xfrm_gro = xo->flags & XFRM_GRO;

		err = x->inner_mode->afinfo->transport_finish(skb, xfrm_gro || async);
		if (xfrm_gro) {
			sp = skb_sec_path(skb);
			if (sp)
				sp->olen = 0;
			skb_dst_drop(skb);
			gro_cells_receive(&gro_cells, skb);
			return err;
		}

		return err;
	}

drop_unlock:
	spin_unlock(&x->lock);
drop:
	xfrm_rcv_cb(skb, family, x && x->type ? x->type->proto : nexthdr, -1);
	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);

static void xfrm_trans_reinject(unsigned long data)
{
	struct xfrm_trans_tasklet *trans = (void *)data;
	struct sk_buff_head queue;
	struct sk_buff *skb;

	__skb_queue_head_init(&queue);
	skb_queue_splice_init(&trans->queue, &queue);

	while ((skb = __skb_dequeue(&queue)))
		XFRM_TRANS_SKB_CB(skb)->finish(dev_net(skb->dev), NULL, skb);
}

int xfrm_trans_queue(struct sk_buff *skb,
		     int (*finish)(struct net *, struct sock *,
				   struct sk_buff *))
{
	struct xfrm_trans_tasklet *trans;

	trans = this_cpu_ptr(&xfrm_trans_tasklet);

	if (skb_queue_len(&trans->queue) >= netdev_max_backlog)
		return -ENOBUFS;

	XFRM_TRANS_SKB_CB(skb)->finish = finish;
	__skb_queue_tail(&trans->queue, skb);
	tasklet_schedule(&trans->tasklet);
	return 0;
}
EXPORT_SYMBOL(xfrm_trans_queue);

void __init xfrm_input_init(void)
{
	int err;
	int i;

	init_dummy_netdev(&xfrm_napi_dev);
	err = gro_cells_init(&gro_cells, &xfrm_napi_dev);
	if (err)
		gro_cells.cells = NULL;

	secpath_cachep = kmem_cache_create("secpath_cache",
					   sizeof(struct sec_path),
					   0, SLAB_HWCACHE_ALIGN|SLAB_PANIC,
					   NULL);

	for_each_possible_cpu(i) {
		struct xfrm_trans_tasklet *trans;

		trans = &per_cpu(xfrm_trans_tasklet, i);
		__skb_queue_head_init(&trans->queue);
		tasklet_init(&trans->tasklet, xfrm_trans_reinject,
			     (unsigned long)trans);
	}
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* xfrm_input.c
	4	*
	5	* Changes:
	6	* YOSHIFUJI Hideaki @USAGI
	7	* Split up af-specific portion
a716c119	8	*
1da177e4 LT	9	*/
1da177e4 LT	10
acf568ee	11	#include <linux/bottom_half.h>
f8c3d0dd	12	#include <linux/cache.h>
acf568ee	13	#include <linux/interrupt.h>
1da177e4 LT	14	#include <linux/slab.h>
1da177e4 LT	15	#include <linux/module.h>
716062fd	16	#include <linux/netdevice.h>
acf568ee	17	#include <linux/percpu.h>
716062fd	18	#include <net/dst.h>
1da177e4 LT	19	#include <net/ip.h>
1da177e4 LT	20	#include <net/xfrm.h>
049f8e2e AD	21	#include <net/ip_tunnels.h>
049f8e2e AD	22	#include <net/ip6_tunnel.h>
1da177e4	23
acf568ee HX	24	struct xfrm_trans_tasklet {
	25	struct tasklet_struct tasklet;
	26	struct sk_buff_head queue;
	27	};
	28
	29	struct xfrm_trans_cb {
9a3fb9fb SK	30	union {
	31	struct inet_skb_parm h4;
	32	#if IS_ENABLED(CONFIG_IPV6)
	33	struct inet6_skb_parm h6;
	34	#endif
	35	} header;
acf568ee HX	36	int (finish)(struct net net, struct sock sk, struct sk_buff skb);
	37	};
	38
	39	#define XFRM_TRANS_SKB_CB(__skb) ((struct xfrm_trans_cb *)&((__skb)->cb[0]))
	40
f8c3d0dd	41	static struct kmem_cache *secpath_cachep __ro_after_init;
1da177e4	42
2f32b51b	43	static DEFINE_SPINLOCK(xfrm_input_afinfo_lock);
960fdfde	44	static struct xfrm_input_afinfo const __rcu *xfrm_input_afinfo[AF_INET6 + 1];
2f32b51b	45
1995876a SK	46	static struct gro_cells gro_cells;
	47	static struct net_device xfrm_napi_dev;
	48
acf568ee HX	49	static DEFINE_PER_CPU(struct xfrm_trans_tasklet, xfrm_trans_tasklet);
acf568ee HX	50
960fdfde	51	int xfrm_input_register_afinfo(const struct xfrm_input_afinfo *afinfo)
2f32b51b SK	52	{
	53	int err = 0;
	54
960fdfde	55	if (WARN_ON(afinfo->family >= ARRAY_SIZE(xfrm_input_afinfo)))
2f32b51b	56	return -EAFNOSUPPORT;
960fdfde	57
2f32b51b SK	58	spin_lock_bh(&xfrm_input_afinfo_lock);
2f32b51b SK	59	if (unlikely(xfrm_input_afinfo[afinfo->family] != NULL))
f31e8d4f	60	err = -EEXIST;
2f32b51b SK	61	else
	62	rcu_assign_pointer(xfrm_input_afinfo[afinfo->family], afinfo);
	63	spin_unlock_bh(&xfrm_input_afinfo_lock);
	64	return err;
	65	}
	66	EXPORT_SYMBOL(xfrm_input_register_afinfo);
	67
960fdfde	68	int xfrm_input_unregister_afinfo(const struct xfrm_input_afinfo *afinfo)
2f32b51b SK	69	{
	70	int err = 0;
	71
2f32b51b SK	72	spin_lock_bh(&xfrm_input_afinfo_lock);
	73	if (likely(xfrm_input_afinfo[afinfo->family] != NULL)) {
	74	if (unlikely(xfrm_input_afinfo[afinfo->family] != afinfo))
	75	err = -EINVAL;
	76	else
	77	RCU_INIT_POINTER(xfrm_input_afinfo[afinfo->family], NULL);
	78	}
	79	spin_unlock_bh(&xfrm_input_afinfo_lock);
	80	synchronize_rcu();
	81	return err;
	82	}
	83	EXPORT_SYMBOL(xfrm_input_unregister_afinfo);
	84
960fdfde	85	static const struct xfrm_input_afinfo *xfrm_input_get_afinfo(unsigned int family)
2f32b51b	86	{
960fdfde	87	const struct xfrm_input_afinfo *afinfo;
2f32b51b	88
960fdfde	89	if (WARN_ON_ONCE(family >= ARRAY_SIZE(xfrm_input_afinfo)))
2f32b51b	90	return NULL;
960fdfde	91
2f32b51b SK	92	rcu_read_lock();
	93	afinfo = rcu_dereference(xfrm_input_afinfo[family]);
	94	if (unlikely(!afinfo))
	95	rcu_read_unlock();
	96	return afinfo;
	97	}
	98
2f32b51b SK	99	static int xfrm_rcv_cb(struct sk_buff *skb, unsigned int family, u8 protocol,
	100	int err)
	101	{
	102	int ret;
960fdfde	103	const struct xfrm_input_afinfo *afinfo = xfrm_input_get_afinfo(family);
2f32b51b SK	104
	105	if (!afinfo)
	106	return -EAFNOSUPPORT;
	107
	108	ret = afinfo->callback(skb, protocol, err);
960fdfde	109	rcu_read_unlock();
2f32b51b SK	110
	111	return ret;
	112	}
	113
1da177e4 LT	114	void __secpath_destroy(struct sec_path *sp)
	115	{
	116	int i;
	117	for (i = 0; i < sp->len; i++)
dbe5b4aa	118	xfrm_state_put(sp->xvec[i]);
1da177e4 LT	119	kmem_cache_free(secpath_cachep, sp);
	120	}
	121	EXPORT_SYMBOL(__secpath_destroy);
	122
	123	struct sec_path secpath_dup(struct sec_path src)
	124	{
	125	struct sec_path *sp;
	126
54e6ecb2	127	sp = kmem_cache_alloc(secpath_cachep, GFP_ATOMIC);
1da177e4 LT	128	if (!sp)
	129	return NULL;
	130
	131	sp->len = 0;
54ef207a SK	132	sp->olen = 0;
54ef207a SK	133
f1193e91	134	memset(sp->ovec, 0, sizeof(sp->ovec));
d77e38e6	135
1da177e4 LT	136	if (src) {
	137	int i;
	138
	139	memcpy(sp, src, sizeof(*sp));
	140	for (i = 0; i < sp->len; i++)
dbe5b4aa	141	xfrm_state_hold(sp->xvec[i]);
1da177e4	142	}
55eabed6	143	refcount_set(&sp->refcnt, 1);
1da177e4 LT	144	return sp;
	145	}
	146	EXPORT_SYMBOL(secpath_dup);
	147
0ca64da1	148	struct sec_path secpath_set(struct sk_buff skb)
b0fcee82	149	{
0ca64da1	150	struct sec_path *sp = skb->sp;
b0fcee82 SK	151
b0fcee82 SK	152	/* Allocate new secpath or COW existing one. */
0ca64da1	153	if (!sp \|\| refcount_read(&sp->refcnt) != 1) {
b0fcee82 SK	154	sp = secpath_dup(skb->sp);
b0fcee82 SK	155	if (!sp)
0ca64da1	156	return NULL;
b0fcee82 SK	157
	158	if (skb->sp)
	159	secpath_put(skb->sp);
	160	skb->sp = sp;
	161	}
0ca64da1 FW	162
0ca64da1 FW	163	return sp;
b0fcee82 SK	164	}
	165	EXPORT_SYMBOL(secpath_set);
	166
1da177e4 LT	167	/* Fetch spi and seq from ipsec header */
1da177e4 LT	168
6067b2ba	169	int xfrm_parse_spi(struct sk_buff skb, u8 nexthdr, __be32 spi, __be32 *seq)
1da177e4 LT	170	{
1da177e4 LT	171	int offset, offset_seq;
44072500	172	int hlen;
1da177e4 LT	173
	174	switch (nexthdr) {
	175	case IPPROTO_AH:
44072500	176	hlen = sizeof(struct ip_auth_hdr);
1da177e4 LT	177	offset = offsetof(struct ip_auth_hdr, spi);
	178	offset_seq = offsetof(struct ip_auth_hdr, seq_no);
	179	break;
	180	case IPPROTO_ESP:
44072500	181	hlen = sizeof(struct ip_esp_hdr);
1da177e4 LT	182	offset = offsetof(struct ip_esp_hdr, spi);
	183	offset_seq = offsetof(struct ip_esp_hdr, seq_no);
	184	break;
	185	case IPPROTO_COMP:
	186	if (!pskb_may_pull(skb, sizeof(struct ip_comp_hdr)))
	187	return -EINVAL;
3e94c2dc	188	spi = htonl(ntohs((__be16 *)(skb_transport_header(skb) + 2)));
1da177e4 LT	189	*seq = 0;
	190	return 0;
	191	default:
	192	return 1;
	193	}
	194
44072500	195	if (!pskb_may_pull(skb, hlen))
1da177e4 LT	196	return -EINVAL;
1da177e4 LT	197
3e94c2dc WC	198	spi = (__be32 *)(skb_transport_header(skb) + offset);
3e94c2dc WC	199	seq = (__be32 *)(skb_transport_header(skb) + offset_seq);
1da177e4 LT	200	return 0;
1da177e4 LT	201	}
1e295370	202	EXPORT_SYMBOL(xfrm_parse_spi);
1da177e4	203
227620e2 HX	204	int xfrm_prepare_input(struct xfrm_state x, struct sk_buff skb)
227620e2 HX	205	{
df9dcb45	206	struct xfrm_mode *inner_mode = x->inner_mode;
227620e2 HX	207	int err;
	208
	209	err = x->outer_mode->afinfo->extract_input(x, skb);
	210	if (err)
	211	return err;
	212
df9dcb45 KM	213	if (x->sel.family == AF_UNSPEC) {
	214	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
	215	if (inner_mode == NULL)
	216	return -EAFNOSUPPORT;
	217	}
	218
	219	skb->protocol = inner_mode->afinfo->eth_proto;
	220	return inner_mode->input2(x, skb);
227620e2 HX	221	}
	222	EXPORT_SYMBOL(xfrm_prepare_input);
	223
716062fd HX	224	int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
716062fd HX	225	{
bd235e3c	226	struct net *net = dev_net(skb->dev);
716062fd HX	227	int err;
716062fd HX	228	__be32 seq;
2cd08467	229	__be32 seq_hi;
3328715e	230	struct xfrm_state *x = NULL;
1bf06cd2	231	xfrm_address_t *daddr;
df9dcb45	232	struct xfrm_mode *inner_mode;
049f8e2e	233	u32 mark = skb->mark;
4ce3dbe3	234	unsigned int family = AF_UNSPEC;
716062fd	235	int decaps = 0;
1bf06cd2	236	int async = 0;
7785bba2	237	bool xfrm_gro = false;
d77e38e6 SK	238	bool crypto_done = false;
d77e38e6 SK	239	struct xfrm_offload *xo = xfrm_offload(skb);
0ca64da1	240	struct sec_path *sp;
1bf06cd2	241
1bf06cd2	242	if (encap_type < 0) {
00501121	243	x = xfrm_input_state(skb);
4ce3dbe3 AH	244
	245	if (unlikely(x->km.state != XFRM_STATE_VALID)) {
	246	if (x->km.state == XFRM_STATE_ACQ)
	247	XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
	248	else
	249	XFRM_INC_STATS(net,
	250	LINUX_MIB_XFRMINSTATEINVALID);
	251	goto drop;
	252	}
	253
3328715e	254	family = x->outer_mode->afinfo->family;
7785bba2 SK	255
	256	/* An encap_type of -1 indicates async resumption. */
	257	if (encap_type == -1) {
	258	async = 1;
	259	seq = XFRM_SKB_CB(skb)->seq.input.low;
	260	goto resume;
	261	}
bcd1f8a4	262
7785bba2 SK	263	/* encap_type < -1 indicates a GRO call. */
	264	encap_type = 0;
	265	seq = XFRM_SPI_SKB_CB(skb)->seq;
716062fd	266
bcd1f8a4 SK	267	if (xo && (xo->flags & CRYPTO_DONE)) {
bcd1f8a4 SK	268	crypto_done = true;
bcd1f8a4 SK	269	family = XFRM_SPI_SKB_CB(skb)->family;
	270
	271	if (!(xo->status & CRYPTO_SUCCESS)) {
	272	if (xo->status &
	273	(CRYPTO_TRANSPORT_AH_AUTH_FAILED \|
	274	CRYPTO_TRANSPORT_ESP_AUTH_FAILED \|
	275	CRYPTO_TUNNEL_AH_AUTH_FAILED \|
	276	CRYPTO_TUNNEL_ESP_AUTH_FAILED)) {
	277
	278	xfrm_audit_state_icvfail(x, skb,
	279	x->type->proto);
	280	x->stats.integrity_failed++;
	281	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
	282	goto drop;
	283	}
	284
47ebcc0b YK	285	if (xo->status & CRYPTO_INVALID_PROTOCOL) {
	286	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
	287	goto drop;
	288	}
	289
bcd1f8a4	290	XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
d77e38e6 SK	291	goto drop;
	292	}
	293
bcd1f8a4 SK	294	if ((err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
	295	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
	296	goto drop;
	297	}
d77e38e6 SK	298	}
d77e38e6 SK	299
7785bba2	300	goto lock;
1bf06cd2	301	}
716062fd	302
3328715e SK	303	family = XFRM_SPI_SKB_CB(skb)->family;
3328715e SK	304
049f8e2e	305	/* if tunnel is present override skb->mark value with tunnel i_key */
1625f452 AK	306	switch (family) {
	307	case AF_INET:
	308	if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4)
049f8e2e	309	mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4->parms.i_key);
1625f452 AK	310	break;
	311	case AF_INET6:
	312	if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6)
049f8e2e	313	mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6->parms.i_key);
1625f452	314	break;
049f8e2e AD	315	}
049f8e2e AD	316
0ca64da1 FW	317	sp = secpath_set(skb);
0ca64da1 FW	318	if (!sp) {
b0fcee82 SK	319	XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
b0fcee82 SK	320	goto drop;
b2aa5e9d HX	321	}
b2aa5e9d HX	322
716062fd	323	seq = 0;
0aa64774	324	if (!spi && (err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
f203b76d	325	secpath_reset(skb);
59c9940e	326	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
716062fd	327	goto drop;
0aa64774	328	}
716062fd	329
cb79a180 FW	330	daddr = (xfrm_address_t *)(skb_network_header(skb) +
cb79a180 FW	331	XFRM_SPI_SKB_CB(skb)->daddroff);
716062fd	332	do {
2294be0f FW	333	sp = skb_sec_path(skb);
	334
	335	if (sp->len == XFRM_MAX_DEPTH) {
f203b76d	336	secpath_reset(skb);
59c9940e	337	XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
716062fd	338	goto drop;
0aa64774	339	}
716062fd	340
049f8e2e	341	x = xfrm_state_lookup(net, mark, daddr, spi, nexthdr, family);
0aa64774	342	if (x == NULL) {
f203b76d	343	secpath_reset(skb);
59c9940e	344	XFRM_INC_STATS(net, LINUX_MIB_XFRMINNOSTATES);
afeb14b4	345	xfrm_audit_state_notfound(skb, family, spi, seq);
716062fd	346	goto drop;
0aa64774	347	}
716062fd	348
9b42c1f1 SK	349	skb->mark = xfrm_smark_get(skb->mark, x);
9b42c1f1 SK	350
2294be0f	351	sp->xvec[sp->len++] = x;
b2aa5e9d	352
7785bba2	353	lock:
716062fd	354	spin_lock(&x->lock);
4c4d41f2	355
0aa64774	356	if (unlikely(x->km.state != XFRM_STATE_VALID)) {
dc0565ce LR	357	if (x->km.state == XFRM_STATE_ACQ)
	358	XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
	359	else
	360	XFRM_INC_STATS(net,
	361	LINUX_MIB_XFRMINSTATEINVALID);
716062fd	362	goto drop_unlock;
0aa64774	363	}
716062fd	364
3de77cf2 DM	365	if ((x->encap ? x->encap->encap_type : 0) != encap_type) {
	366	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMISMATCH);
	367	goto drop_unlock;
	368	}
	369
36ae0148	370	if (x->repl->check(x, skb, seq)) {
59c9940e	371	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
716062fd	372	goto drop_unlock;
0aa64774	373	}
716062fd	374
0aa64774	375	if (xfrm_state_check_expire(x)) {
59c9940e	376	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEEXPIRED);
716062fd	377	goto drop_unlock;
0aa64774	378	}
716062fd	379
0ebea8ef HX	380	spin_unlock(&x->lock);
0ebea8ef HX	381
68c11e98 AD	382	if (xfrm_tunnel_check(skb, x, family)) {
	383	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
	384	goto drop;
	385	}
	386
2cd08467 SK	387	seq_hi = htonl(xfrm_replay_seqhi(x, seq));
2cd08467 SK	388
1ce3644a	389	XFRM_SKB_CB(skb)->seq.input.low = seq;
2cd08467	390	XFRM_SKB_CB(skb)->seq.input.hi = seq_hi;
1bf06cd2	391
3bc07321	392	skb_dst_force(skb);
071d36bf	393	dev_hold(skb->dev);
3bc07321	394
d77e38e6 SK	395	if (crypto_done)
	396	nexthdr = x->type_offload->input_tail(x, skb);
	397	else
	398	nexthdr = x->type->input(x, skb);
0ebea8ef	399
1bf06cd2 HX	400	if (nexthdr == -EINPROGRESS)
1bf06cd2 HX	401	return 0;
1bf06cd2	402	resume:
071d36bf	403	dev_put(skb->dev);
071d36bf	404
0ebea8ef	405	spin_lock(&x->lock);
668dc8af	406	if (nexthdr <= 0) {
9dd3245a HX	407	if (nexthdr == -EBADMSG) {
	408	xfrm_audit_state_icvfail(x, skb,
	409	x->type->proto);
668dc8af	410	x->stats.integrity_failed++;
9dd3245a	411	}
59c9940e	412	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
716062fd	413	goto drop_unlock;
668dc8af	414	}
716062fd	415
716062fd HX	416	/* only the first xfrm gets the encap type */
	417	encap_type = 0;
	418
3b59df46	419	if (async && x->repl->recheck(x, skb, seq)) {
bcf66bf5 SK	420	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
	421	goto drop_unlock;
	422	}
	423
9fdc4883	424	x->repl->advance(x, seq);
716062fd HX	425
	426	x->curlft.bytes += skb->len;
	427	x->curlft.packets++;
	428
	429	spin_unlock(&x->lock);
	430
60d5fcfb HX	431	XFRM_MODE_SKB_CB(skb)->protocol = nexthdr;
60d5fcfb HX	432
df9dcb45 KM	433	inner_mode = x->inner_mode;
	434
	435	if (x->sel.family == AF_UNSPEC) {
	436	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
cb866e32 SK	437	if (inner_mode == NULL) {
cb866e32 SK	438	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
df9dcb45	439	goto drop;
cb866e32	440	}
df9dcb45 KM	441	}
	442
	443	if (inner_mode->input(x, skb)) {
59c9940e	444	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
716062fd	445	goto drop;
0aa64774	446	}
716062fd HX	447
	448	if (x->outer_mode->flags & XFRM_MODE_FLAG_TUNNEL) {
	449	decaps = 1;
	450	break;
	451	}
	452
1bf06cd2 HX	453	/*
	454	* We need the inner address. However, we only get here for
	455	* transport mode so the outer address is identical.
	456	*/
	457	daddr = &x->id.daddr;
2fcb45b6	458	family = x->outer_mode->afinfo->family;
1bf06cd2	459
716062fd	460	err = xfrm_parse_spi(skb, nexthdr, &spi, &seq);
0aa64774	461	if (err < 0) {
59c9940e	462	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
716062fd	463	goto drop;
0aa64774	464	}
782710e3	465	crypto_done = false;
716062fd HX	466	} while (!err);
716062fd HX	467
3328715e SK	468	err = xfrm_rcv_cb(skb, family, x->type->proto, 0);
	469	if (err)
	470	goto drop;
	471
716062fd HX	472	nf_reset(skb);
	473
	474	if (decaps) {
2294be0f FW	475	sp = skb_sec_path(skb);
	476	if (sp)
	477	sp->olen = 0;
adf30907	478	skb_dst_drop(skb);
1995876a	479	gro_cells_receive(&gro_cells, skb);
716062fd HX	480	return 0;
716062fd HX	481	} else {
7785bba2 SK	482	xo = xfrm_offload(skb);
	483	if (xo)
	484	xfrm_gro = xo->flags & XFRM_GRO;
	485
cfcf99f9	486	err = x->inner_mode->afinfo->transport_finish(skb, xfrm_gro \|\| async);
7785bba2	487	if (xfrm_gro) {
2294be0f FW	488	sp = skb_sec_path(skb);
	489	if (sp)
	490	sp->olen = 0;
7785bba2 SK	491	skb_dst_drop(skb);
	492	gro_cells_receive(&gro_cells, skb);
	493	return err;
	494	}
	495
	496	return err;
716062fd HX	497	}
	498
	499	drop_unlock:
	500	spin_unlock(&x->lock);
716062fd	501	drop:
3328715e	502	xfrm_rcv_cb(skb, family, x && x->type ? x->type->proto : nexthdr, -1);
716062fd HX	503	kfree_skb(skb);
	504	return 0;
	505	}
	506	EXPORT_SYMBOL(xfrm_input);
	507
1bf06cd2 HX	508	int xfrm_input_resume(struct sk_buff *skb, int nexthdr)
	509	{
	510	return xfrm_input(skb, nexthdr, 0, -1);
	511	}
	512	EXPORT_SYMBOL(xfrm_input_resume);
	513
acf568ee HX	514	static void xfrm_trans_reinject(unsigned long data)
	515	{
	516	struct xfrm_trans_tasklet trans = (void )data;
	517	struct sk_buff_head queue;
	518	struct sk_buff *skb;
	519
	520	__skb_queue_head_init(&queue);
	521	skb_queue_splice_init(&trans->queue, &queue);
	522
	523	while ((skb = __skb_dequeue(&queue)))
	524	XFRM_TRANS_SKB_CB(skb)->finish(dev_net(skb->dev), NULL, skb);
	525	}
	526
	527	int xfrm_trans_queue(struct sk_buff *skb,
	528	int (finish)(struct net , struct sock *,
	529	struct sk_buff *))
	530	{
	531	struct xfrm_trans_tasklet *trans;
	532
	533	trans = this_cpu_ptr(&xfrm_trans_tasklet);
	534
	535	if (skb_queue_len(&trans->queue) >= netdev_max_backlog)
	536	return -ENOBUFS;
	537
	538	XFRM_TRANS_SKB_CB(skb)->finish = finish;
d16b46e4	539	__skb_queue_tail(&trans->queue, skb);
acf568ee HX	540	tasklet_schedule(&trans->tasklet);
	541	return 0;
	542	}
	543	EXPORT_SYMBOL(xfrm_trans_queue);
	544
1da177e4 LT	545	void __init xfrm_input_init(void)
1da177e4 LT	546	{
1995876a	547	int err;
acf568ee	548	int i;
1995876a SK	549
	550	init_dummy_netdev(&xfrm_napi_dev);
	551	err = gro_cells_init(&gro_cells, &xfrm_napi_dev);
	552	if (err)
	553	gro_cells.cells = NULL;
	554
1da177e4 LT	555	secpath_cachep = kmem_cache_create("secpath_cache",
1da177e4 LT	556	sizeof(struct sec_path),
e5d679f3	557	0, SLAB_HWCACHE_ALIGN\|SLAB_PANIC,
20c2df83	558	NULL);
acf568ee HX	559
	560	for_each_possible_cpu(i) {
	561	struct xfrm_trans_tasklet *trans;
	562
	563	trans = &per_cpu(xfrm_trans_tasklet, i);
	564	__skb_queue_head_init(&trans->queue);
	565	tasklet_init(&trans->tasklet, xfrm_trans_reinject,
	566	(unsigned long)trans);
	567	}
1da177e4	568	}