[linux-block.git] / net / netfilter / nf_flow_table_core.c

// SPDX-License-Identifier: GPL-2.0-only
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/netfilter.h>
#include <linux/rhashtable.h>
#include <linux/netdevice.h>
#include <net/ip.h>
#include <net/ip6_route.h>
#include <net/netfilter/nf_tables.h>
#include <net/netfilter/nf_flow_table.h>
#include <net/netfilter/nf_conntrack.h>
#include <net/netfilter/nf_conntrack_core.h>
#include <net/netfilter/nf_conntrack_tuple.h>

struct flow_offload_entry {
	struct flow_offload	flow;
	struct nf_conn		*ct;
	struct rcu_head		rcu_head;
};

static DEFINE_MUTEX(flowtable_lock);
static LIST_HEAD(flowtables);

static void
flow_offload_fill_dir(struct flow_offload *flow, struct nf_conn *ct,
		      struct nf_flow_route *route,
		      enum flow_offload_tuple_dir dir)
{
	struct flow_offload_tuple *ft = &flow->tuplehash[dir].tuple;
	struct nf_conntrack_tuple *ctt = &ct->tuplehash[dir].tuple;
	struct dst_entry *other_dst = route->tuple[!dir].dst;
	struct dst_entry *dst = route->tuple[dir].dst;

	ft->dir = dir;

	switch (ctt->src.l3num) {
	case NFPROTO_IPV4:
		ft->src_v4 = ctt->src.u3.in;
		ft->dst_v4 = ctt->dst.u3.in;
		ft->mtu = ip_dst_mtu_maybe_forward(dst, true);
		break;
	case NFPROTO_IPV6:
		ft->src_v6 = ctt->src.u3.in6;
		ft->dst_v6 = ctt->dst.u3.in6;
		ft->mtu = ip6_dst_mtu_forward(dst);
		break;
	}

	ft->l3proto = ctt->src.l3num;
	ft->l4proto = ctt->dst.protonum;
	ft->src_port = ctt->src.u.tcp.port;
	ft->dst_port = ctt->dst.u.tcp.port;

	ft->iifidx = other_dst->dev->ifindex;
	ft->oifidx = dst->dev->ifindex;
	ft->dst_cache = dst;
}

struct flow_offload *
flow_offload_alloc(struct nf_conn *ct, struct nf_flow_route *route)
{
	struct flow_offload_entry *entry;
	struct flow_offload *flow;

	if (unlikely(nf_ct_is_dying(ct) ||
	    !atomic_inc_not_zero(&ct->ct_general.use)))
		return NULL;

	entry = kzalloc(sizeof(*entry), GFP_ATOMIC);
	if (!entry)
		goto err_ct_refcnt;

	flow = &entry->flow;

	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst))
		goto err_dst_cache_original;

	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_REPLY].dst))
		goto err_dst_cache_reply;

	entry->ct = ct;

	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_ORIGINAL);
	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_REPLY);

	if (ct->status & IPS_SRC_NAT)
		flow->flags |= FLOW_OFFLOAD_SNAT;
	if (ct->status & IPS_DST_NAT)
		flow->flags |= FLOW_OFFLOAD_DNAT;

	return flow;

err_dst_cache_reply:
	dst_release(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst);
err_dst_cache_original:
	kfree(entry);
err_ct_refcnt:
	nf_ct_put(ct);

	return NULL;
}
EXPORT_SYMBOL_GPL(flow_offload_alloc);

static void flow_offload_fixup_tcp(struct ip_ct_tcp *tcp)
{
	tcp->state = TCP_CONNTRACK_ESTABLISHED;
	tcp->seen[0].td_maxwin = 0;
	tcp->seen[1].td_maxwin = 0;
}

#define NF_FLOWTABLE_TCP_PICKUP_TIMEOUT	(120 * HZ)
#define NF_FLOWTABLE_UDP_PICKUP_TIMEOUT	(30 * HZ)

static void flow_offload_fixup_ct_state(struct nf_conn *ct)
{
	const struct nf_conntrack_l4proto *l4proto;
	unsigned int timeout;
	int l4num;

	l4num = nf_ct_protonum(ct);
	if (l4num == IPPROTO_TCP)
		flow_offload_fixup_tcp(&ct->proto.tcp);

	l4proto = nf_ct_l4proto_find(l4num);
	if (!l4proto)
		return;

	if (l4num == IPPROTO_TCP)
		timeout = NF_FLOWTABLE_TCP_PICKUP_TIMEOUT;
	else if (l4num == IPPROTO_UDP)
		timeout = NF_FLOWTABLE_UDP_PICKUP_TIMEOUT;
	else
		return;

	ct->timeout = nfct_time_stamp + timeout;
}

void flow_offload_free(struct flow_offload *flow)
{
	struct flow_offload_entry *e;

	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_cache);
	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_cache);
	e = container_of(flow, struct flow_offload_entry, flow);
	if (flow->flags & FLOW_OFFLOAD_DYING)
		nf_ct_delete(e->ct, 0, 0);
	nf_ct_put(e->ct);
	kfree_rcu(e, rcu_head);
}
EXPORT_SYMBOL_GPL(flow_offload_free);

static u32 flow_offload_hash(const void *data, u32 len, u32 seed)
{
	const struct flow_offload_tuple *tuple = data;

	return jhash(tuple, offsetof(struct flow_offload_tuple, dir), seed);
}

static u32 flow_offload_hash_obj(const void *data, u32 len, u32 seed)
{
	const struct flow_offload_tuple_rhash *tuplehash = data;

	return jhash(&tuplehash->tuple, offsetof(struct flow_offload_tuple, dir), seed);
}

static int flow_offload_hash_cmp(struct rhashtable_compare_arg *arg,
					const void *ptr)
{
	const struct flow_offload_tuple *tuple = arg->key;
	const struct flow_offload_tuple_rhash *x = ptr;

	if (memcmp(&x->tuple, tuple, offsetof(struct flow_offload_tuple, dir)))
		return 1;

	return 0;
}

static const struct rhashtable_params nf_flow_offload_rhash_params = {
	.head_offset		= offsetof(struct flow_offload_tuple_rhash, node),
	.hashfn			= flow_offload_hash,
	.obj_hashfn		= flow_offload_hash_obj,
	.obj_cmpfn		= flow_offload_hash_cmp,
	.automatic_shrinking	= true,
};

int flow_offload_add(struct nf_flowtable *flow_table, struct flow_offload *flow)
{
	int err;

	err = rhashtable_insert_fast(&flow_table->rhashtable,
				     &flow->tuplehash[0].node,
				     nf_flow_offload_rhash_params);
	if (err < 0)
		return err;

	err = rhashtable_insert_fast(&flow_table->rhashtable,
				     &flow->tuplehash[1].node,
				     nf_flow_offload_rhash_params);
	if (err < 0) {
		rhashtable_remove_fast(&flow_table->rhashtable,
				       &flow->tuplehash[0].node,
				       nf_flow_offload_rhash_params);
		return err;
	}

	flow->timeout = (u32)jiffies;
	return 0;
}
EXPORT_SYMBOL_GPL(flow_offload_add);

static void flow_offload_del(struct nf_flowtable *flow_table,
			     struct flow_offload *flow)
{
	struct flow_offload_entry *e;

	rhashtable_remove_fast(&flow_table->rhashtable,
			       &flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].node,
			       nf_flow_offload_rhash_params);
	rhashtable_remove_fast(&flow_table->rhashtable,
			       &flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].node,
			       nf_flow_offload_rhash_params);

	e = container_of(flow, struct flow_offload_entry, flow);
	clear_bit(IPS_OFFLOAD_BIT, &e->ct->status);

	flow_offload_free(flow);
}

void flow_offload_teardown(struct flow_offload *flow)
{
	struct flow_offload_entry *e;

	flow->flags |= FLOW_OFFLOAD_TEARDOWN;

	e = container_of(flow, struct flow_offload_entry, flow);
	flow_offload_fixup_ct_state(e->ct);
}
EXPORT_SYMBOL_GPL(flow_offload_teardown);

struct flow_offload_tuple_rhash *
flow_offload_lookup(struct nf_flowtable *flow_table,
		    struct flow_offload_tuple *tuple)
{
	struct flow_offload_tuple_rhash *tuplehash;
	struct flow_offload *flow;
	struct flow_offload_entry *e;
	int dir;

	tuplehash = rhashtable_lookup(&flow_table->rhashtable, tuple,
				      nf_flow_offload_rhash_params);
	if (!tuplehash)
		return NULL;

	dir = tuplehash->tuple.dir;
	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	if (flow->flags & (FLOW_OFFLOAD_DYING | FLOW_OFFLOAD_TEARDOWN))
		return NULL;

	e = container_of(flow, struct flow_offload_entry, flow);
	if (unlikely(nf_ct_is_dying(e->ct)))
		return NULL;

	return tuplehash;
}
EXPORT_SYMBOL_GPL(flow_offload_lookup);

static int
nf_flow_table_iterate(struct nf_flowtable *flow_table,
		      void (*iter)(struct flow_offload *flow, void *data),
		      void *data)
{
	struct flow_offload_tuple_rhash *tuplehash;
	struct rhashtable_iter hti;
	struct flow_offload *flow;
	int err = 0;

	rhashtable_walk_enter(&flow_table->rhashtable, &hti);
	rhashtable_walk_start(&hti);

	while ((tuplehash = rhashtable_walk_next(&hti))) {
		if (IS_ERR(tuplehash)) {
			if (PTR_ERR(tuplehash) != -EAGAIN) {
				err = PTR_ERR(tuplehash);
				break;
			}
			continue;
		}
		if (tuplehash->tuple.dir)
			continue;

		flow = container_of(tuplehash, struct flow_offload, tuplehash[0]);

		iter(flow, data);
	}
	rhashtable_walk_stop(&hti);
	rhashtable_walk_exit(&hti);

	return err;
}

static inline bool nf_flow_has_expired(const struct flow_offload *flow)
{
	return (__s32)(flow->timeout - (u32)jiffies) <= 0;
}

static void nf_flow_offload_gc_step(struct flow_offload *flow, void *data)
{
	struct nf_flowtable *flow_table = data;
	struct flow_offload_entry *e;

	e = container_of(flow, struct flow_offload_entry, flow);
	if (nf_flow_has_expired(flow) || nf_ct_is_dying(e->ct) ||
	    (flow->flags & (FLOW_OFFLOAD_DYING | FLOW_OFFLOAD_TEARDOWN)))
		flow_offload_del(flow_table, flow);
}

static void nf_flow_offload_work_gc(struct work_struct *work)
{
	struct nf_flowtable *flow_table;

	flow_table = container_of(work, struct nf_flowtable, gc_work.work);
	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
	queue_delayed_work(system_power_efficient_wq, &flow_table->gc_work, HZ);
}

static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
				__be16 port, __be16 new_port)
{
	struct tcphdr *tcph;

	if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) ||
	    skb_try_make_writable(skb, thoff + sizeof(*tcph)))
		return -1;

	tcph = (void *)(skb_network_header(skb) + thoff);
	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, true);

	return 0;
}

static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
				__be16 port, __be16 new_port)
{
	struct udphdr *udph;

	if (!pskb_may_pull(skb, thoff + sizeof(*udph)) ||
	    skb_try_make_writable(skb, thoff + sizeof(*udph)))
		return -1;

	udph = (void *)(skb_network_header(skb) + thoff);
	if (udph->check || skb->ip_summed == CHECKSUM_PARTIAL) {
		inet_proto_csum_replace2(&udph->check, skb, port,
					 new_port, true);
		if (!udph->check)
			udph->check = CSUM_MANGLED_0;
	}

	return 0;
}

static int nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
			    u8 protocol, __be16 port, __be16 new_port)
{
	switch (protocol) {
	case IPPROTO_TCP:
		if (nf_flow_nat_port_tcp(skb, thoff, port, new_port) < 0)
			return NF_DROP;
		break;
	case IPPROTO_UDP:
		if (nf_flow_nat_port_udp(skb, thoff, port, new_port) < 0)
			return NF_DROP;
		break;
	}

	return 0;
}

int nf_flow_snat_port(const struct flow_offload *flow,
		      struct sk_buff *skb, unsigned int thoff,
		      u8 protocol, enum flow_offload_tuple_dir dir)
{
	struct flow_ports *hdr;
	__be16 port, new_port;

	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) ||
	    skb_try_make_writable(skb, thoff + sizeof(*hdr)))
		return -1;

	hdr = (void *)(skb_network_header(skb) + thoff);

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		port = hdr->source;
		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_port;
		hdr->source = new_port;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		port = hdr->dest;
		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_port;
		hdr->dest = new_port;
		break;
	default:
		return -1;
	}

	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
}
EXPORT_SYMBOL_GPL(nf_flow_snat_port);

int nf_flow_dnat_port(const struct flow_offload *flow,
		      struct sk_buff *skb, unsigned int thoff,
		      u8 protocol, enum flow_offload_tuple_dir dir)
{
	struct flow_ports *hdr;
	__be16 port, new_port;

	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) ||
	    skb_try_make_writable(skb, thoff + sizeof(*hdr)))
		return -1;

	hdr = (void *)(skb_network_header(skb) + thoff);

	switch (dir) {
	case FLOW_OFFLOAD_DIR_ORIGINAL:
		port = hdr->dest;
		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_port;
		hdr->dest = new_port;
		break;
	case FLOW_OFFLOAD_DIR_REPLY:
		port = hdr->source;
		new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_port;
		hdr->source = new_port;
		break;
	default:
		return -1;
	}

	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
}
EXPORT_SYMBOL_GPL(nf_flow_dnat_port);

int nf_flow_table_init(struct nf_flowtable *flowtable)
{
	int err;

	INIT_DEFERRABLE_WORK(&flowtable->gc_work, nf_flow_offload_work_gc);

	err = rhashtable_init(&flowtable->rhashtable,
			      &nf_flow_offload_rhash_params);
	if (err < 0)
		return err;

	queue_delayed_work(system_power_efficient_wq,
			   &flowtable->gc_work, HZ);

	mutex_lock(&flowtable_lock);
	list_add(&flowtable->list, &flowtables);
	mutex_unlock(&flowtable_lock);

	return 0;
}
EXPORT_SYMBOL_GPL(nf_flow_table_init);

static void nf_flow_table_do_cleanup(struct flow_offload *flow, void *data)
{
	struct net_device *dev = data;
	struct flow_offload_entry *e;

	e = container_of(flow, struct flow_offload_entry, flow);

	if (!dev) {
		flow_offload_teardown(flow);
		return;
	}
	if (net_eq(nf_ct_net(e->ct), dev_net(dev)) &&
	    (flow->tuplehash[0].tuple.iifidx == dev->ifindex ||
	     flow->tuplehash[1].tuple.iifidx == dev->ifindex))
		flow_offload_dead(flow);
}

static void nf_flow_table_iterate_cleanup(struct nf_flowtable *flowtable,
					  struct net_device *dev)
{
	nf_flow_table_iterate(flowtable, nf_flow_table_do_cleanup, dev);
	flush_delayed_work(&flowtable->gc_work);
}

void nf_flow_table_cleanup(struct net_device *dev)
{
	struct nf_flowtable *flowtable;

	mutex_lock(&flowtable_lock);
	list_for_each_entry(flowtable, &flowtables, list)
		nf_flow_table_iterate_cleanup(flowtable, dev);
	mutex_unlock(&flowtable_lock);
}
EXPORT_SYMBOL_GPL(nf_flow_table_cleanup);

void nf_flow_table_free(struct nf_flowtable *flow_table)
{
	mutex_lock(&flowtable_lock);
	list_del(&flow_table->list);
	mutex_unlock(&flowtable_lock);
	cancel_delayed_work_sync(&flow_table->gc_work);
	nf_flow_table_iterate(flow_table, nf_flow_table_do_cleanup, NULL);
	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
	rhashtable_destroy(&flow_table->rhashtable);
}
EXPORT_SYMBOL_GPL(nf_flow_table_free);

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Pablo Neira Ayuso <pablo@netfilter.org>");
Commit	Line	Data
09c434b8	1	// SPDX-License-Identifier: GPL-2.0-only
ac2a6666 PNA	2	#include <linux/kernel.h>
	3	#include <linux/init.h>
	4	#include <linux/module.h>
	5	#include <linux/netfilter.h>
	6	#include <linux/rhashtable.h>
	7	#include <linux/netdevice.h>
4f3780c0 FF	8	#include <net/ip.h>
4f3780c0 FF	9	#include <net/ip6_route.h>
c0ea1bcb	10	#include <net/netfilter/nf_tables.h>
ac2a6666 PNA	11	#include <net/netfilter/nf_flow_table.h>
	12	#include <net/netfilter/nf_conntrack.h>
	13	#include <net/netfilter/nf_conntrack_core.h>
	14	#include <net/netfilter/nf_conntrack_tuple.h>
	15
	16	struct flow_offload_entry {
	17	struct flow_offload flow;
	18	struct nf_conn *ct;
	19	struct rcu_head rcu_head;
	20	};
	21
84453a90 FF	22	static DEFINE_MUTEX(flowtable_lock);
	23	static LIST_HEAD(flowtables);
	24
047b300e FF	25	static void
	26	flow_offload_fill_dir(struct flow_offload flow, struct nf_conn ct,
	27	struct nf_flow_route *route,
	28	enum flow_offload_tuple_dir dir)
	29	{
	30	struct flow_offload_tuple *ft = &flow->tuplehash[dir].tuple;
	31	struct nf_conntrack_tuple *ctt = &ct->tuplehash[dir].tuple;
10f4e765	32	struct dst_entry *other_dst = route->tuple[!dir].dst;
4f3780c0	33	struct dst_entry *dst = route->tuple[dir].dst;
047b300e FF	34
	35	ft->dir = dir;
	36
	37	switch (ctt->src.l3num) {
	38	case NFPROTO_IPV4:
	39	ft->src_v4 = ctt->src.u3.in;
	40	ft->dst_v4 = ctt->dst.u3.in;
4f3780c0	41	ft->mtu = ip_dst_mtu_maybe_forward(dst, true);
047b300e FF	42	break;
	43	case NFPROTO_IPV6:
	44	ft->src_v6 = ctt->src.u3.in6;
	45	ft->dst_v6 = ctt->dst.u3.in6;
4f3780c0	46	ft->mtu = ip6_dst_mtu_forward(dst);
047b300e FF	47	break;
	48	}
	49
	50	ft->l3proto = ctt->src.l3num;
	51	ft->l4proto = ctt->dst.protonum;
	52	ft->src_port = ctt->src.u.tcp.port;
	53	ft->dst_port = ctt->dst.u.tcp.port;
	54
10f4e765	55	ft->iifidx = other_dst->dev->ifindex;
10f4e765	56	ft->oifidx = dst->dev->ifindex;
4f3780c0	57	ft->dst_cache = dst;
047b300e FF	58	}
047b300e FF	59
ac2a6666 PNA	60	struct flow_offload *
	61	flow_offload_alloc(struct nf_conn ct, struct nf_flow_route route)
	62	{
	63	struct flow_offload_entry *entry;
	64	struct flow_offload *flow;
	65
	66	if (unlikely(nf_ct_is_dying(ct) \|\|
	67	!atomic_inc_not_zero(&ct->ct_general.use)))
	68	return NULL;
	69
	70	entry = kzalloc(sizeof(*entry), GFP_ATOMIC);
	71	if (!entry)
	72	goto err_ct_refcnt;
	73
	74	flow = &entry->flow;
	75
	76	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst))
	77	goto err_dst_cache_original;
	78
	79	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_REPLY].dst))
	80	goto err_dst_cache_reply;
	81
	82	entry->ct = ct;
	83
047b300e FF	84	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_ORIGINAL);
047b300e FF	85	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_REPLY);
ac2a6666 PNA	86
	87	if (ct->status & IPS_SRC_NAT)
	88	flow->flags \|= FLOW_OFFLOAD_SNAT;
df1e2025	89	if (ct->status & IPS_DST_NAT)
ac2a6666 PNA	90	flow->flags \|= FLOW_OFFLOAD_DNAT;
	91
	92	return flow;
	93
	94	err_dst_cache_reply:
	95	dst_release(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst);
	96	err_dst_cache_original:
	97	kfree(entry);
	98	err_ct_refcnt:
	99	nf_ct_put(ct);
	100
	101	return NULL;
	102	}
	103	EXPORT_SYMBOL_GPL(flow_offload_alloc);
	104
da5984e5 FF	105	static void flow_offload_fixup_tcp(struct ip_ct_tcp *tcp)
	106	{
	107	tcp->state = TCP_CONNTRACK_ESTABLISHED;
	108	tcp->seen[0].td_maxwin = 0;
	109	tcp->seen[1].td_maxwin = 0;
	110	}
	111
e97d9404 FW	112	#define NF_FLOWTABLE_TCP_PICKUP_TIMEOUT (120 * HZ)
	113	#define NF_FLOWTABLE_UDP_PICKUP_TIMEOUT (30 * HZ)
	114
da5984e5 FF	115	static void flow_offload_fixup_ct_state(struct nf_conn *ct)
	116	{
	117	const struct nf_conntrack_l4proto *l4proto;
da5984e5 FF	118	unsigned int timeout;
	119	int l4num;
	120
	121	l4num = nf_ct_protonum(ct);
	122	if (l4num == IPPROTO_TCP)
	123	flow_offload_fixup_tcp(&ct->proto.tcp);
	124
4a60dc74	125	l4proto = nf_ct_l4proto_find(l4num);
da5984e5 FF	126	if (!l4proto)
	127	return;
	128
da5984e5	129	if (l4num == IPPROTO_TCP)
e97d9404	130	timeout = NF_FLOWTABLE_TCP_PICKUP_TIMEOUT;
da5984e5	131	else if (l4num == IPPROTO_UDP)
e97d9404	132	timeout = NF_FLOWTABLE_UDP_PICKUP_TIMEOUT;
da5984e5 FF	133	else
	134	return;
	135
	136	ct->timeout = nfct_time_stamp + timeout;
	137	}
	138
ac2a6666 PNA	139	void flow_offload_free(struct flow_offload *flow)
	140	{
	141	struct flow_offload_entry *e;
	142
	143	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_cache);
	144	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_cache);
	145	e = container_of(flow, struct flow_offload_entry, flow);
da5984e5 FF	146	if (flow->flags & FLOW_OFFLOAD_DYING)
da5984e5 FF	147	nf_ct_delete(e->ct, 0, 0);
0ff90b6c FF	148	nf_ct_put(e->ct);
0ff90b6c FF	149	kfree_rcu(e, rcu_head);
ac2a6666 PNA	150	}
	151	EXPORT_SYMBOL_GPL(flow_offload_free);
	152
a268de77 FF	153	static u32 flow_offload_hash(const void *data, u32 len, u32 seed)
	154	{
	155	const struct flow_offload_tuple *tuple = data;
	156
	157	return jhash(tuple, offsetof(struct flow_offload_tuple, dir), seed);
	158	}
	159
	160	static u32 flow_offload_hash_obj(const void *data, u32 len, u32 seed)
	161	{
	162	const struct flow_offload_tuple_rhash *tuplehash = data;
	163
	164	return jhash(&tuplehash->tuple, offsetof(struct flow_offload_tuple, dir), seed);
	165	}
	166
	167	static int flow_offload_hash_cmp(struct rhashtable_compare_arg *arg,
	168	const void *ptr)
	169	{
	170	const struct flow_offload_tuple *tuple = arg->key;
	171	const struct flow_offload_tuple_rhash *x = ptr;
	172
	173	if (memcmp(&x->tuple, tuple, offsetof(struct flow_offload_tuple, dir)))
	174	return 1;
	175
	176	return 0;
	177	}
	178
	179	static const struct rhashtable_params nf_flow_offload_rhash_params = {
	180	.head_offset = offsetof(struct flow_offload_tuple_rhash, node),
	181	.hashfn = flow_offload_hash,
	182	.obj_hashfn = flow_offload_hash_obj,
	183	.obj_cmpfn = flow_offload_hash_cmp,
	184	.automatic_shrinking = true,
	185	};
	186
ac2a6666 PNA	187	int flow_offload_add(struct nf_flowtable flow_table, struct flow_offload flow)
ac2a6666 PNA	188	{
43c8f131 TY	189	int err;
	190
	191	err = rhashtable_insert_fast(&flow_table->rhashtable,
	192	&flow->tuplehash[0].node,
	193	nf_flow_offload_rhash_params);
	194	if (err < 0)
	195	return err;
	196
	197	err = rhashtable_insert_fast(&flow_table->rhashtable,
	198	&flow->tuplehash[1].node,
	199	nf_flow_offload_rhash_params);
	200	if (err < 0) {
	201	rhashtable_remove_fast(&flow_table->rhashtable,
	202	&flow->tuplehash[0].node,
	203	nf_flow_offload_rhash_params);
	204	return err;
	205	}
ac2a6666	206
43c8f131	207	flow->timeout = (u32)jiffies;
ac2a6666 PNA	208	return 0;
	209	}
	210	EXPORT_SYMBOL_GPL(flow_offload_add);
	211
0ff90b6c FF	212	static void flow_offload_del(struct nf_flowtable *flow_table,
0ff90b6c FF	213	struct flow_offload *flow)
ac2a6666	214	{
da5984e5 FF	215	struct flow_offload_entry *e;
da5984e5 FF	216
ac2a6666 PNA	217	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	218	&flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].node,
a268de77	219	nf_flow_offload_rhash_params);
ac2a6666 PNA	220	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	221	&flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].node,
a268de77	222	nf_flow_offload_rhash_params);
ac2a6666	223
da5984e5 FF	224	e = container_of(flow, struct flow_offload_entry, flow);
	225	clear_bit(IPS_OFFLOAD_BIT, &e->ct->status);
	226
0ff90b6c	227	flow_offload_free(flow);
ac2a6666	228	}
ac2a6666	229
59c466dd FF	230	void flow_offload_teardown(struct flow_offload *flow)
59c466dd FF	231	{
da5984e5 FF	232	struct flow_offload_entry *e;
da5984e5 FF	233
59c466dd	234	flow->flags \|= FLOW_OFFLOAD_TEARDOWN;
da5984e5 FF	235
	236	e = container_of(flow, struct flow_offload_entry, flow);
	237	flow_offload_fixup_ct_state(e->ct);
59c466dd FF	238	}
	239	EXPORT_SYMBOL_GPL(flow_offload_teardown);
	240
ac2a6666 PNA	241	struct flow_offload_tuple_rhash *
	242	flow_offload_lookup(struct nf_flowtable *flow_table,
	243	struct flow_offload_tuple *tuple)
	244	{
ba03137f FF	245	struct flow_offload_tuple_rhash *tuplehash;
ba03137f FF	246	struct flow_offload *flow;
8cd2bc98	247	struct flow_offload_entry *e;
ba03137f FF	248	int dir;
ba03137f FF	249
a2d88182 TY	250	tuplehash = rhashtable_lookup(&flow_table->rhashtable, tuple,
a2d88182 TY	251	nf_flow_offload_rhash_params);
ba03137f FF	252	if (!tuplehash)
	253	return NULL;
	254
	255	dir = tuplehash->tuple.dir;
	256	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	257	if (flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN))
	258	return NULL;
	259
8cd2bc98 TY	260	e = container_of(flow, struct flow_offload_entry, flow);
	261	if (unlikely(nf_ct_is_dying(e->ct)))
	262	return NULL;
	263
ba03137f	264	return tuplehash;
ac2a6666 PNA	265	}
	266	EXPORT_SYMBOL_GPL(flow_offload_lookup);
	267
49de9c09 TY	268	static int
	269	nf_flow_table_iterate(struct nf_flowtable *flow_table,
	270	void (iter)(struct flow_offload flow, void *data),
	271	void *data)
ac2a6666 PNA	272	{
	273	struct flow_offload_tuple_rhash *tuplehash;
	274	struct rhashtable_iter hti;
	275	struct flow_offload *flow;
0de22baa	276	int err = 0;
ac2a6666	277
0de22baa	278	rhashtable_walk_enter(&flow_table->rhashtable, &hti);
ac2a6666 PNA	279	rhashtable_walk_start(&hti);
	280
	281	while ((tuplehash = rhashtable_walk_next(&hti))) {
	282	if (IS_ERR(tuplehash)) {
0de22baa TY	283	if (PTR_ERR(tuplehash) != -EAGAIN) {
	284	err = PTR_ERR(tuplehash);
	285	break;
	286	}
ac2a6666 PNA	287	continue;
	288	}
	289	if (tuplehash->tuple.dir)
	290	continue;
	291
	292	flow = container_of(tuplehash, struct flow_offload, tuplehash[0]);
	293
	294	iter(flow, data);
	295	}
ac2a6666 PNA	296	rhashtable_walk_stop(&hti);
	297	rhashtable_walk_exit(&hti);
	298
	299	return err;
	300	}
ac2a6666 PNA	301
	302	static inline bool nf_flow_has_expired(const struct flow_offload *flow)
	303	{
	304	return (__s32)(flow->timeout - (u32)jiffies) <= 0;
	305	}
	306
b9660987	307	static void nf_flow_offload_gc_step(struct flow_offload flow, void data)
ac2a6666	308	{
b9660987	309	struct nf_flowtable *flow_table = data;
8cd2bc98	310	struct flow_offload_entry *e;
ac2a6666	311
8cd2bc98 TY	312	e = container_of(flow, struct flow_offload_entry, flow);
8cd2bc98 TY	313	if (nf_flow_has_expired(flow) \|\| nf_ct_is_dying(e->ct) \|\|
b9660987 TY	314	(flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN)))
b9660987 TY	315	flow_offload_del(flow_table, flow);
b408c5b0 PNA	316	}
b408c5b0 PNA	317
a268de77	318	static void nf_flow_offload_work_gc(struct work_struct *work)
b408c5b0 PNA	319	{
	320	struct nf_flowtable *flow_table;
	321
	322	flow_table = container_of(work, struct nf_flowtable, gc_work.work);
b9660987	323	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
ac2a6666 PNA	324	queue_delayed_work(system_power_efficient_wq, &flow_table->gc_work, HZ);
ac2a6666 PNA	325	}
ac2a6666 PNA	326
	327	static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
	328	__be16 port, __be16 new_port)
	329	{
	330	struct tcphdr *tcph;
	331
	332	if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) \|\|
	333	skb_try_make_writable(skb, thoff + sizeof(*tcph)))
	334	return -1;
	335
	336	tcph = (void *)(skb_network_header(skb) + thoff);
	337	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, true);
	338
	339	return 0;
	340	}
	341
	342	static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
	343	__be16 port, __be16 new_port)
	344	{
	345	struct udphdr *udph;
	346
	347	if (!pskb_may_pull(skb, thoff + sizeof(*udph)) \|\|
	348	skb_try_make_writable(skb, thoff + sizeof(*udph)))
	349	return -1;
	350
	351	udph = (void *)(skb_network_header(skb) + thoff);
	352	if (udph->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	353	inet_proto_csum_replace2(&udph->check, skb, port,
	354	new_port, true);
	355	if (!udph->check)
	356	udph->check = CSUM_MANGLED_0;
	357	}
	358
	359	return 0;
	360	}
	361
	362	static int nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
	363	u8 protocol, __be16 port, __be16 new_port)
	364	{
	365	switch (protocol) {
	366	case IPPROTO_TCP:
	367	if (nf_flow_nat_port_tcp(skb, thoff, port, new_port) < 0)
	368	return NF_DROP;
	369	break;
	370	case IPPROTO_UDP:
	371	if (nf_flow_nat_port_udp(skb, thoff, port, new_port) < 0)
	372	return NF_DROP;
	373	break;
	374	}
	375
	376	return 0;
	377	}
	378
	379	int nf_flow_snat_port(const struct flow_offload *flow,
	380	struct sk_buff *skb, unsigned int thoff,
	381	u8 protocol, enum flow_offload_tuple_dir dir)
	382	{
	383	struct flow_ports *hdr;
	384	__be16 port, new_port;
	385
	386	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
	387	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
	388	return -1;
	389
390	hdr = (void *)(skb_network_header(skb) + thoff);
391
392	switch (dir) {
393	case FLOW_OFFLOAD_DIR_ORIGINAL:
394	port = hdr->source;
395	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_port;
396	hdr->source = new_port;
397	break;
398	case FLOW_OFFLOAD_DIR_REPLY:
399	port = hdr->dest;
400	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_port;
401	hdr->dest = new_port;
402	break;
403	default:
404	return -1;
405	}
406
407	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
408	}
409	EXPORT_SYMBOL_GPL(nf_flow_snat_port);
410
411	int nf_flow_dnat_port(const struct flow_offload *flow,
412	struct sk_buff *skb, unsigned int thoff,
413	u8 protocol, enum flow_offload_tuple_dir dir)
414	{
415	struct flow_ports *hdr;
416	__be16 port, new_port;
417
418	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
419	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
420	return -1;
421
422	hdr = (void *)(skb_network_header(skb) + thoff);
423
424	switch (dir) {
425	case FLOW_OFFLOAD_DIR_ORIGINAL:
426	port = hdr->dest;
427	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_port;
428	hdr->dest = new_port;
429	break;
430	case FLOW_OFFLOAD_DIR_REPLY:
431	port = hdr->source;
432	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_port;
433	hdr->source = new_port;
434	break;
435	default:
436	return -1;
437	}
438
439	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
440	}
441	EXPORT_SYMBOL_GPL(nf_flow_dnat_port);
442
a268de77 FF	443	int nf_flow_table_init(struct nf_flowtable *flowtable)
	444	{
	445	int err;
	446
	447	INIT_DEFERRABLE_WORK(&flowtable->gc_work, nf_flow_offload_work_gc);
	448
	449	err = rhashtable_init(&flowtable->rhashtable,
	450	&nf_flow_offload_rhash_params);
	451	if (err < 0)
	452	return err;
	453
	454	queue_delayed_work(system_power_efficient_wq,
	455	&flowtable->gc_work, HZ);
	456
84453a90 FF	457	mutex_lock(&flowtable_lock);
	458	list_add(&flowtable->list, &flowtables);
	459	mutex_unlock(&flowtable_lock);
	460
a268de77 FF	461	return 0;
	462	}
	463	EXPORT_SYMBOL_GPL(nf_flow_table_init);
	464
c0ea1bcb PNA	465	static void nf_flow_table_do_cleanup(struct flow_offload flow, void data)
	466	{
	467	struct net_device *dev = data;
a3fb3698 TY	468	struct flow_offload_entry *e;
	469
	470	e = container_of(flow, struct flow_offload_entry, flow);
c0ea1bcb	471
59c466dd FF	472	if (!dev) {
59c466dd FF	473	flow_offload_teardown(flow);
c0ea1bcb	474	return;
59c466dd	475	}
a3fb3698 TY	476	if (net_eq(nf_ct_net(e->ct), dev_net(dev)) &&
	477	(flow->tuplehash[0].tuple.iifidx == dev->ifindex \|\|
	478	flow->tuplehash[1].tuple.iifidx == dev->ifindex))
59c466dd	479	flow_offload_dead(flow);
c0ea1bcb PNA	480	}
	481
	482	static void nf_flow_table_iterate_cleanup(struct nf_flowtable *flowtable,
84453a90	483	struct net_device *dev)
c0ea1bcb	484	{
84453a90	485	nf_flow_table_iterate(flowtable, nf_flow_table_do_cleanup, dev);
c0ea1bcb PNA	486	flush_delayed_work(&flowtable->gc_work);
	487	}
	488
5f1be84a	489	void nf_flow_table_cleanup(struct net_device *dev)
c0ea1bcb	490	{
84453a90 FF	491	struct nf_flowtable *flowtable;
	492
	493	mutex_lock(&flowtable_lock);
	494	list_for_each_entry(flowtable, &flowtables, list)
	495	nf_flow_table_iterate_cleanup(flowtable, dev);
	496	mutex_unlock(&flowtable_lock);
c0ea1bcb PNA	497	}
	498	EXPORT_SYMBOL_GPL(nf_flow_table_cleanup);
	499
b408c5b0 PNA	500	void nf_flow_table_free(struct nf_flowtable *flow_table)
b408c5b0 PNA	501	{
84453a90 FF	502	mutex_lock(&flowtable_lock);
	503	list_del(&flow_table->list);
	504	mutex_unlock(&flowtable_lock);
a268de77	505	cancel_delayed_work_sync(&flow_table->gc_work);
b408c5b0	506	nf_flow_table_iterate(flow_table, nf_flow_table_do_cleanup, NULL);
b9660987	507	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
a268de77	508	rhashtable_destroy(&flow_table->rhashtable);
b408c5b0 PNA	509	}
	510	EXPORT_SYMBOL_GPL(nf_flow_table_free);
	511
ac2a6666 PNA	512	MODULE_LICENSE("GPL");
ac2a6666 PNA	513	MODULE_AUTHOR("Pablo Neira Ayuso <pablo@netfilter.org>");