[people/arne_f/kernel.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/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 __read_mostly;

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);

int secpath_set(struct sk_buff *skb)
{
	struct sec_path *sp;

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

		if (skb->sp)
			secpath_put(skb->sp);
		skb->sp = sp;
	}
	return 0;
}
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);

	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);

			if (encap_type == -1)
				dev_put(skb->dev);
			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;
			x = xfrm_input_state(skb);
			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;
	}

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

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

	daddr = (xfrm_address_t *)(skb_network_header(skb) +
				   XFRM_SPI_SKB_CB(skb)->daddroff);
	do {
		if (skb->sp->len == XFRM_MAX_DEPTH) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
			goto drop;
		}

		x = xfrm_state_lookup(net, 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;

		skb_dst_force(skb);
		if (!skb_dst(skb)) {
			XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
			goto drop;
		}

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;

		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) {
		if (skb->sp)
			skb->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) {
			if (skb->sp)
				skb->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
e095ecae HX	11	#include <linux/bottom_half.h>
e095ecae HX	12	#include <linux/interrupt.h>
1da177e4 LT	13	#include <linux/slab.h>
1da177e4 LT	14	#include <linux/module.h>
716062fd	15	#include <linux/netdevice.h>
e095ecae	16	#include <linux/percpu.h>
716062fd	17	#include <net/dst.h>
1da177e4 LT	18	#include <net/ip.h>
1da177e4 LT	19	#include <net/xfrm.h>
049f8e2e AD	20	#include <net/ip_tunnels.h>
049f8e2e AD	21	#include <net/ip6_tunnel.h>
1da177e4	22
e095ecae HX	23	struct xfrm_trans_tasklet {
	24	struct tasklet_struct tasklet;
	25	struct sk_buff_head queue;
	26	};
	27
	28	struct xfrm_trans_cb {
58be6253 SK	29	union {
	30	struct inet_skb_parm h4;
	31	#if IS_ENABLED(CONFIG_IPV6)
	32	struct inet6_skb_parm h6;
	33	#endif
	34	} header;
e095ecae HX	35	int (finish)(struct net net, struct sock sk, struct sk_buff skb);
	36	};
	37
	38	#define XFRM_TRANS_SKB_CB(__skb) ((struct xfrm_trans_cb *)&((__skb)->cb[0]))
	39
e18b890b	40	static struct kmem_cache *secpath_cachep __read_mostly;
1da177e4	41
2f32b51b	42	static DEFINE_SPINLOCK(xfrm_input_afinfo_lock);
960fdfde	43	static struct xfrm_input_afinfo const __rcu *xfrm_input_afinfo[AF_INET6 + 1];
2f32b51b	44
1995876a SK	45	static struct gro_cells gro_cells;
	46	static struct net_device xfrm_napi_dev;
	47
e095ecae HX	48	static DEFINE_PER_CPU(struct xfrm_trans_tasklet, xfrm_trans_tasklet);
e095ecae HX	49
960fdfde	50	int xfrm_input_register_afinfo(const struct xfrm_input_afinfo *afinfo)
2f32b51b SK	51	{
	52	int err = 0;
	53
960fdfde	54	if (WARN_ON(afinfo->family >= ARRAY_SIZE(xfrm_input_afinfo)))
2f32b51b	55	return -EAFNOSUPPORT;
960fdfde	56
2f32b51b SK	57	spin_lock_bh(&xfrm_input_afinfo_lock);
2f32b51b SK	58	if (unlikely(xfrm_input_afinfo[afinfo->family] != NULL))
f31e8d4f	59	err = -EEXIST;
2f32b51b SK	60	else
	61	rcu_assign_pointer(xfrm_input_afinfo[afinfo->family], afinfo);
	62	spin_unlock_bh(&xfrm_input_afinfo_lock);
	63	return err;
	64	}
	65	EXPORT_SYMBOL(xfrm_input_register_afinfo);
	66
960fdfde	67	int xfrm_input_unregister_afinfo(const struct xfrm_input_afinfo *afinfo)
2f32b51b SK	68	{
	69	int err = 0;
	70
2f32b51b SK	71	spin_lock_bh(&xfrm_input_afinfo_lock);
	72	if (likely(xfrm_input_afinfo[afinfo->family] != NULL)) {
	73	if (unlikely(xfrm_input_afinfo[afinfo->family] != afinfo))
	74	err = -EINVAL;
	75	else
	76	RCU_INIT_POINTER(xfrm_input_afinfo[afinfo->family], NULL);
	77	}
	78	spin_unlock_bh(&xfrm_input_afinfo_lock);
	79	synchronize_rcu();
	80	return err;
	81	}
	82	EXPORT_SYMBOL(xfrm_input_unregister_afinfo);
	83
960fdfde	84	static const struct xfrm_input_afinfo *xfrm_input_get_afinfo(unsigned int family)
2f32b51b	85	{
960fdfde	86	const struct xfrm_input_afinfo *afinfo;
2f32b51b	87
960fdfde	88	if (WARN_ON_ONCE(family >= ARRAY_SIZE(xfrm_input_afinfo)))
2f32b51b	89	return NULL;
960fdfde	90
2f32b51b SK	91	rcu_read_lock();
	92	afinfo = rcu_dereference(xfrm_input_afinfo[family]);
	93	if (unlikely(!afinfo))
	94	rcu_read_unlock();
	95	return afinfo;
	96	}
	97
2f32b51b SK	98	static int xfrm_rcv_cb(struct sk_buff *skb, unsigned int family, u8 protocol,
	99	int err)
	100	{
	101	int ret;
960fdfde	102	const struct xfrm_input_afinfo *afinfo = xfrm_input_get_afinfo(family);
2f32b51b SK	103
	104	if (!afinfo)
	105	return -EAFNOSUPPORT;
	106
	107	ret = afinfo->callback(skb, protocol, err);
960fdfde	108	rcu_read_unlock();
2f32b51b SK	109
	110	return ret;
	111	}
	112
1da177e4 LT	113	void __secpath_destroy(struct sec_path *sp)
	114	{
	115	int i;
	116	for (i = 0; i < sp->len; i++)
dbe5b4aa	117	xfrm_state_put(sp->xvec[i]);
1da177e4 LT	118	kmem_cache_free(secpath_cachep, sp);
	119	}
	120	EXPORT_SYMBOL(__secpath_destroy);
	121
	122	struct sec_path secpath_dup(struct sec_path src)
	123	{
	124	struct sec_path *sp;
	125
54e6ecb2	126	sp = kmem_cache_alloc(secpath_cachep, GFP_ATOMIC);
1da177e4 LT	127	if (!sp)
	128	return NULL;
	129
	130	sp->len = 0;
54ef207a SK	131	sp->olen = 0;
54ef207a SK	132
3dc92564	133	memset(sp->ovec, 0, sizeof(sp->ovec));
d77e38e6	134
1da177e4 LT	135	if (src) {
	136	int i;
	137
	138	memcpy(sp, src, sizeof(*sp));
	139	for (i = 0; i < sp->len; i++)
dbe5b4aa	140	xfrm_state_hold(sp->xvec[i]);
1da177e4	141	}
55eabed6	142	refcount_set(&sp->refcnt, 1);
1da177e4 LT	143	return sp;
	144	}
	145	EXPORT_SYMBOL(secpath_dup);
	146
b0fcee82 SK	147	int secpath_set(struct sk_buff *skb)
	148	{
	149	struct sec_path *sp;
	150
	151	/* Allocate new secpath or COW existing one. */
55eabed6	152	if (!skb->sp \|\| refcount_read(&skb->sp->refcnt) != 1) {
b0fcee82 SK	153	sp = secpath_dup(skb->sp);
	154	if (!sp)
	155	return -ENOMEM;
	156
	157	if (skb->sp)
	158	secpath_put(skb->sp);
	159	skb->sp = sp;
	160	}
	161	return 0;
	162	}
	163	EXPORT_SYMBOL(secpath_set);
	164
1da177e4 LT	165	/* Fetch spi and seq from ipsec header */
1da177e4 LT	166
6067b2ba	167	int xfrm_parse_spi(struct sk_buff skb, u8 nexthdr, __be32 spi, __be32 *seq)
1da177e4 LT	168	{
1da177e4 LT	169	int offset, offset_seq;
44072500	170	int hlen;
1da177e4 LT	171
	172	switch (nexthdr) {
	173	case IPPROTO_AH:
44072500	174	hlen = sizeof(struct ip_auth_hdr);
1da177e4 LT	175	offset = offsetof(struct ip_auth_hdr, spi);
	176	offset_seq = offsetof(struct ip_auth_hdr, seq_no);
	177	break;
	178	case IPPROTO_ESP:
44072500	179	hlen = sizeof(struct ip_esp_hdr);
1da177e4 LT	180	offset = offsetof(struct ip_esp_hdr, spi);
	181	offset_seq = offsetof(struct ip_esp_hdr, seq_no);
	182	break;
	183	case IPPROTO_COMP:
	184	if (!pskb_may_pull(skb, sizeof(struct ip_comp_hdr)))
	185	return -EINVAL;
3e94c2dc	186	spi = htonl(ntohs((__be16 *)(skb_transport_header(skb) + 2)));
1da177e4 LT	187	*seq = 0;
	188	return 0;
	189	default:
	190	return 1;
	191	}
	192
44072500	193	if (!pskb_may_pull(skb, hlen))
1da177e4 LT	194	return -EINVAL;
1da177e4 LT	195
3e94c2dc WC	196	spi = (__be32 *)(skb_transport_header(skb) + offset);
3e94c2dc WC	197	seq = (__be32 *)(skb_transport_header(skb) + offset_seq);
1da177e4 LT	198	return 0;
1da177e4 LT	199	}
1e295370	200	EXPORT_SYMBOL(xfrm_parse_spi);
1da177e4	201
227620e2 HX	202	int xfrm_prepare_input(struct xfrm_state x, struct sk_buff skb)
227620e2 HX	203	{
df9dcb45	204	struct xfrm_mode *inner_mode = x->inner_mode;
227620e2 HX	205	int err;
	206
	207	err = x->outer_mode->afinfo->extract_input(x, skb);
	208	if (err)
	209	return err;
	210
df9dcb45 KM	211	if (x->sel.family == AF_UNSPEC) {
	212	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
	213	if (inner_mode == NULL)
	214	return -EAFNOSUPPORT;
	215	}
	216
	217	skb->protocol = inner_mode->afinfo->eth_proto;
	218	return inner_mode->input2(x, skb);
227620e2 HX	219	}
	220	EXPORT_SYMBOL(xfrm_prepare_input);
	221
716062fd HX	222	int xfrm_input(struct sk_buff *skb, int nexthdr, __be32 spi, int encap_type)
716062fd HX	223	{
bd235e3c	224	struct net *net = dev_net(skb->dev);
716062fd HX	225	int err;
716062fd HX	226	__be32 seq;
2cd08467	227	__be32 seq_hi;
3328715e	228	struct xfrm_state *x = NULL;
1bf06cd2	229	xfrm_address_t *daddr;
df9dcb45	230	struct xfrm_mode *inner_mode;
049f8e2e	231	u32 mark = skb->mark;
c232fd3d	232	unsigned int family = AF_UNSPEC;
716062fd	233	int decaps = 0;
1bf06cd2	234	int async = 0;
7785bba2	235	bool xfrm_gro = false;
d77e38e6 SK	236	bool crypto_done = false;
d77e38e6 SK	237	struct xfrm_offload *xo = xfrm_offload(skb);
1bf06cd2	238
1bf06cd2	239	if (encap_type < 0) {
00501121	240	x = xfrm_input_state(skb);
c232fd3d AH	241
	242	if (unlikely(x->km.state != XFRM_STATE_VALID)) {
	243	if (x->km.state == XFRM_STATE_ACQ)
	244	XFRM_INC_STATS(net, LINUX_MIB_XFRMACQUIREERROR);
	245	else
	246	XFRM_INC_STATS(net,
	247	LINUX_MIB_XFRMINSTATEINVALID);
3bbb8eec XX	248
	249	if (encap_type == -1)
	250	dev_put(skb->dev);
c232fd3d AH	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)) {
	268	crypto_done = true;
	269	x = xfrm_input_state(skb);
	270	family = XFRM_SPI_SKB_CB(skb)->family;
	271
	272	if (!(xo->status & CRYPTO_SUCCESS)) {
	273	if (xo->status &
	274	(CRYPTO_TRANSPORT_AH_AUTH_FAILED \|
	275	CRYPTO_TRANSPORT_ESP_AUTH_FAILED \|
	276	CRYPTO_TUNNEL_AH_AUTH_FAILED \|
	277	CRYPTO_TUNNEL_ESP_AUTH_FAILED)) {
	278
	279	xfrm_audit_state_icvfail(x, skb,
	280	x->type->proto);
	281	x->stats.integrity_failed++;
	282	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
	283	goto drop;
	284	}
	285
47ebcc0b YK	286	if (xo->status & CRYPTO_INVALID_PROTOCOL) {
	287	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
	288	goto drop;
	289	}
	290
bcd1f8a4	291	XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
d77e38e6 SK	292	goto drop;
	293	}
	294
bcd1f8a4 SK	295	if ((err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
	296	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
	297	goto drop;
	298	}
d77e38e6 SK	299	}
d77e38e6 SK	300
7785bba2	301	goto lock;
1bf06cd2	302	}
716062fd	303
3328715e SK	304	family = XFRM_SPI_SKB_CB(skb)->family;
3328715e SK	305
049f8e2e	306	/* if tunnel is present override skb->mark value with tunnel i_key */
1625f452 AK	307	switch (family) {
	308	case AF_INET:
	309	if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4)
049f8e2e	310	mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip4->parms.i_key);
1625f452 AK	311	break;
	312	case AF_INET6:
	313	if (XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6)
049f8e2e	314	mark = be32_to_cpu(XFRM_TUNNEL_SKB_CB(skb)->tunnel.ip6->parms.i_key);
1625f452	315	break;
049f8e2e AD	316	}
049f8e2e AD	317
b0fcee82 SK	318	err = secpath_set(skb);
	319	if (err) {
	320	XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
	321	goto drop;
b2aa5e9d HX	322	}
b2aa5e9d HX	323
716062fd	324	seq = 0;
0aa64774	325	if (!spi && (err = xfrm_parse_spi(skb, nexthdr, &spi, &seq)) != 0) {
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 {
0aa64774	333	if (skb->sp->len == XFRM_MAX_DEPTH) {
59c9940e	334	XFRM_INC_STATS(net, LINUX_MIB_XFRMINBUFFERERROR);
716062fd	335	goto drop;
0aa64774	336	}
716062fd	337
049f8e2e	338	x = xfrm_state_lookup(net, mark, daddr, spi, nexthdr, family);
0aa64774	339	if (x == NULL) {
59c9940e	340	XFRM_INC_STATS(net, LINUX_MIB_XFRMINNOSTATES);
afeb14b4	341	xfrm_audit_state_notfound(skb, family, spi, seq);
716062fd	342	goto drop;
0aa64774	343	}
716062fd	344
b2aa5e9d HX	345	skb->sp->xvec[skb->sp->len++] = x;
b2aa5e9d HX	346
4ddb80f4 SK	347	skb_dst_force(skb);
	348	if (!skb_dst(skb)) {
	349	XFRM_INC_STATS(net, LINUX_MIB_XFRMINERROR);
	350	goto drop;
	351	}
	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
071d36bf	392	dev_hold(skb->dev);
3bc07321	393
d77e38e6 SK	394	if (crypto_done)
	395	nexthdr = x->type_offload->input_tail(x, skb);
	396	else
	397	nexthdr = x->type->input(x, skb);
0ebea8ef	398
1bf06cd2 HX	399	if (nexthdr == -EINPROGRESS)
1bf06cd2 HX	400	return 0;
1bf06cd2	401	resume:
071d36bf	402	dev_put(skb->dev);
071d36bf	403
0ebea8ef	404	spin_lock(&x->lock);
06d984ae	405	if (nexthdr < 0) {
9dd3245a HX	406	if (nexthdr == -EBADMSG) {
	407	xfrm_audit_state_icvfail(x, skb,
	408	x->type->proto);
668dc8af	409	x->stats.integrity_failed++;
9dd3245a	410	}
59c9940e	411	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEPROTOERROR);
716062fd	412	goto drop_unlock;
668dc8af	413	}
716062fd	414
716062fd HX	415	/* only the first xfrm gets the encap type */
	416	encap_type = 0;
	417
3b59df46	418	if (async && x->repl->recheck(x, skb, seq)) {
bcf66bf5 SK	419	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATESEQERROR);
	420	goto drop_unlock;
	421	}
	422
9fdc4883	423	x->repl->advance(x, seq);
716062fd HX	424
	425	x->curlft.bytes += skb->len;
	426	x->curlft.packets++;
	427
	428	spin_unlock(&x->lock);
	429
60d5fcfb HX	430	XFRM_MODE_SKB_CB(skb)->protocol = nexthdr;
60d5fcfb HX	431
df9dcb45 KM	432	inner_mode = x->inner_mode;
	433
	434	if (x->sel.family == AF_UNSPEC) {
	435	inner_mode = xfrm_ip2inner_mode(x, XFRM_MODE_SKB_CB(skb)->protocol);
cb866e32 SK	436	if (inner_mode == NULL) {
cb866e32 SK	437	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
df9dcb45	438	goto drop;
cb866e32	439	}
df9dcb45 KM	440	}
	441
	442	if (inner_mode->input(x, skb)) {
59c9940e	443	XFRM_INC_STATS(net, LINUX_MIB_XFRMINSTATEMODEERROR);
716062fd	444	goto drop;
0aa64774	445	}
716062fd HX	446
	447	if (x->outer_mode->flags & XFRM_MODE_FLAG_TUNNEL) {
	448	decaps = 1;
	449	break;
	450	}
	451
1bf06cd2 HX	452	/*
	453	* We need the inner address. However, we only get here for
	454	* transport mode so the outer address is identical.
	455	*/
	456	daddr = &x->id.daddr;
2fcb45b6	457	family = x->outer_mode->afinfo->family;
1bf06cd2	458
716062fd	459	err = xfrm_parse_spi(skb, nexthdr, &spi, &seq);
0aa64774	460	if (err < 0) {
59c9940e	461	XFRM_INC_STATS(net, LINUX_MIB_XFRMINHDRERROR);
716062fd	462	goto drop;
0aa64774	463	}
64f38286	464	crypto_done = false;
716062fd HX	465	} while (!err);
716062fd HX	466
3328715e SK	467	err = xfrm_rcv_cb(skb, family, x->type->proto, 0);
	468	if (err)
	469	goto drop;
	470
716062fd HX	471	nf_reset(skb);
	472
	473	if (decaps) {
23e9fcfe AK	474	if (skb->sp)
23e9fcfe AK	475	skb->sp->olen = 0;
adf30907	476	skb_dst_drop(skb);
1995876a	477	gro_cells_receive(&gro_cells, skb);
716062fd HX	478	return 0;
716062fd HX	479	} else {
7785bba2 SK	480	xo = xfrm_offload(skb);
	481	if (xo)
	482	xfrm_gro = xo->flags & XFRM_GRO;
	483
cfcf99f9	484	err = x->inner_mode->afinfo->transport_finish(skb, xfrm_gro \|\| async);
7785bba2	485	if (xfrm_gro) {
23e9fcfe AK	486	if (skb->sp)
23e9fcfe AK	487	skb->sp->olen = 0;
7785bba2 SK	488	skb_dst_drop(skb);
	489	gro_cells_receive(&gro_cells, skb);
	490	return err;
	491	}
	492
	493	return err;
716062fd HX	494	}
	495
	496	drop_unlock:
	497	spin_unlock(&x->lock);
716062fd	498	drop:
3328715e	499	xfrm_rcv_cb(skb, family, x && x->type ? x->type->proto : nexthdr, -1);
716062fd HX	500	kfree_skb(skb);
	501	return 0;
	502	}
	503	EXPORT_SYMBOL(xfrm_input);
	504
1bf06cd2 HX	505	int xfrm_input_resume(struct sk_buff *skb, int nexthdr)
	506	{
	507	return xfrm_input(skb, nexthdr, 0, -1);
	508	}
	509	EXPORT_SYMBOL(xfrm_input_resume);
	510
e095ecae HX	511	static void xfrm_trans_reinject(unsigned long data)
	512	{
	513	struct xfrm_trans_tasklet trans = (void )data;
	514	struct sk_buff_head queue;
	515	struct sk_buff *skb;
	516
	517	__skb_queue_head_init(&queue);
	518	skb_queue_splice_init(&trans->queue, &queue);
	519
	520	while ((skb = __skb_dequeue(&queue)))
	521	XFRM_TRANS_SKB_CB(skb)->finish(dev_net(skb->dev), NULL, skb);
	522	}
	523
	524	int xfrm_trans_queue(struct sk_buff *skb,
	525	int (finish)(struct net , struct sock *,
	526	struct sk_buff *))
	527	{
	528	struct xfrm_trans_tasklet *trans;
	529
	530	trans = this_cpu_ptr(&xfrm_trans_tasklet);
	531
	532	if (skb_queue_len(&trans->queue) >= netdev_max_backlog)
	533	return -ENOBUFS;
	534
	535	XFRM_TRANS_SKB_CB(skb)->finish = finish;
d2d85f8d	536	__skb_queue_tail(&trans->queue, skb);
e095ecae HX	537	tasklet_schedule(&trans->tasklet);
	538	return 0;
	539	}
	540	EXPORT_SYMBOL(xfrm_trans_queue);
	541
1da177e4 LT	542	void __init xfrm_input_init(void)
1da177e4 LT	543	{
1995876a	544	int err;
e095ecae	545	int i;
1995876a SK	546
	547	init_dummy_netdev(&xfrm_napi_dev);
	548	err = gro_cells_init(&gro_cells, &xfrm_napi_dev);
	549	if (err)
	550	gro_cells.cells = NULL;
	551
1da177e4 LT	552	secpath_cachep = kmem_cache_create("secpath_cache",
1da177e4 LT	553	sizeof(struct sec_path),
e5d679f3	554	0, SLAB_HWCACHE_ALIGN\|SLAB_PANIC,
20c2df83	555	NULL);
e095ecae HX	556
	557	for_each_possible_cpu(i) {
	558	struct xfrm_trans_tasklet *trans;
	559
	560	trans = &per_cpu(xfrm_trans_tasklet, i);
	561	__skb_queue_head_init(&trans->queue);
	562	tasklet_init(&trans->tasklet, xfrm_trans_reinject,
	563	(unsigned long)trans);
	564	}
1da177e4	565	}