[linux-2.6-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->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 inline __s32 nf_flow_timeout_delta(unsigned int timeout)
{
	return (__s32)(timeout - (u32)jiffies);
}

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

	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;

	if (nf_flow_timeout_delta(ct->timeout) > (__s32)timeout)
		ct->timeout = nfct_time_stamp + timeout;
}

static void flow_offload_fixup_ct_state(struct nf_conn *ct)
{
	if (nf_ct_protonum(ct) == IPPROTO_TCP)
		flow_offload_fixup_tcp(&ct->proto.tcp);
}

static void flow_offload_fixup_ct(struct nf_conn *ct)
{
	flow_offload_fixup_ct_state(ct);
	flow_offload_fixup_ct_timeout(ct);
}

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

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);

	if (nf_flow_has_expired(flow))
		flow_offload_fixup_ct(e->ct);
	else if (flow->flags & FLOW_OFFLOAD_TEARDOWN)
		flow_offload_fixup_ct_timeout(e->ct);

	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 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;
4f3780c0	56	ft->dst_cache = dst;
047b300e FF	57	}
047b300e FF	58
ac2a6666 PNA	59	struct flow_offload *
	60	flow_offload_alloc(struct nf_conn ct, struct nf_flow_route route)
	61	{
	62	struct flow_offload_entry *entry;
	63	struct flow_offload *flow;
	64
	65	if (unlikely(nf_ct_is_dying(ct) \|\|
	66	!atomic_inc_not_zero(&ct->ct_general.use)))
	67	return NULL;
	68
	69	entry = kzalloc(sizeof(*entry), GFP_ATOMIC);
	70	if (!entry)
	71	goto err_ct_refcnt;
	72
	73	flow = &entry->flow;
	74
	75	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst))
	76	goto err_dst_cache_original;
	77
	78	if (!dst_hold_safe(route->tuple[FLOW_OFFLOAD_DIR_REPLY].dst))
	79	goto err_dst_cache_reply;
	80
	81	entry->ct = ct;
	82
047b300e FF	83	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_ORIGINAL);
047b300e FF	84	flow_offload_fill_dir(flow, ct, route, FLOW_OFFLOAD_DIR_REPLY);
ac2a6666 PNA	85
	86	if (ct->status & IPS_SRC_NAT)
	87	flow->flags \|= FLOW_OFFLOAD_SNAT;
df1e2025	88	if (ct->status & IPS_DST_NAT)
ac2a6666 PNA	89	flow->flags \|= FLOW_OFFLOAD_DNAT;
	90
	91	return flow;
	92
	93	err_dst_cache_reply:
	94	dst_release(route->tuple[FLOW_OFFLOAD_DIR_ORIGINAL].dst);
	95	err_dst_cache_original:
	96	kfree(entry);
	97	err_ct_refcnt:
	98	nf_ct_put(ct);
	99
	100	return NULL;
	101	}
	102	EXPORT_SYMBOL_GPL(flow_offload_alloc);
	103
da5984e5 FF	104	static void flow_offload_fixup_tcp(struct ip_ct_tcp *tcp)
	105	{
	106	tcp->state = TCP_CONNTRACK_ESTABLISHED;
	107	tcp->seen[0].td_maxwin = 0;
	108	tcp->seen[1].td_maxwin = 0;
	109	}
	110
e97d9404 FW	111	#define NF_FLOWTABLE_TCP_PICKUP_TIMEOUT (120 * HZ)
	112	#define NF_FLOWTABLE_UDP_PICKUP_TIMEOUT (30 * HZ)
	113
1e5b2471 PNA	114	static inline __s32 nf_flow_timeout_delta(unsigned int timeout)
	115	{
	116	return (__s32)(timeout - (u32)jiffies);
	117	}
	118
	119	static void flow_offload_fixup_ct_timeout(struct nf_conn *ct)
da5984e5 FF	120	{
da5984e5 FF	121	const struct nf_conntrack_l4proto *l4proto;
1e5b2471	122	int l4num = nf_ct_protonum(ct);
da5984e5	123	unsigned int timeout;
da5984e5	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
1e5b2471 PNA	136	if (nf_flow_timeout_delta(ct->timeout) > (__s32)timeout)
	137	ct->timeout = nfct_time_stamp + timeout;
	138	}
	139
	140	static void flow_offload_fixup_ct_state(struct nf_conn *ct)
	141	{
	142	if (nf_ct_protonum(ct) == IPPROTO_TCP)
	143	flow_offload_fixup_tcp(&ct->proto.tcp);
	144	}
	145
	146	static void flow_offload_fixup_ct(struct nf_conn *ct)
	147	{
	148	flow_offload_fixup_ct_state(ct);
	149	flow_offload_fixup_ct_timeout(ct);
da5984e5 FF	150	}
da5984e5 FF	151
ac2a6666 PNA	152	void flow_offload_free(struct flow_offload *flow)
	153	{
	154	struct flow_offload_entry *e;
	155
	156	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_cache);
	157	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_cache);
	158	e = container_of(flow, struct flow_offload_entry, flow);
da5984e5 FF	159	if (flow->flags & FLOW_OFFLOAD_DYING)
da5984e5 FF	160	nf_ct_delete(e->ct, 0, 0);
0ff90b6c FF	161	nf_ct_put(e->ct);
0ff90b6c FF	162	kfree_rcu(e, rcu_head);
ac2a6666 PNA	163	}
	164	EXPORT_SYMBOL_GPL(flow_offload_free);
	165
a268de77 FF	166	static u32 flow_offload_hash(const void *data, u32 len, u32 seed)
	167	{
	168	const struct flow_offload_tuple *tuple = data;
	169
	170	return jhash(tuple, offsetof(struct flow_offload_tuple, dir), seed);
	171	}
	172
	173	static u32 flow_offload_hash_obj(const void *data, u32 len, u32 seed)
	174	{
	175	const struct flow_offload_tuple_rhash *tuplehash = data;
	176
	177	return jhash(&tuplehash->tuple, offsetof(struct flow_offload_tuple, dir), seed);
	178	}
	179
	180	static int flow_offload_hash_cmp(struct rhashtable_compare_arg *arg,
	181	const void *ptr)
	182	{
	183	const struct flow_offload_tuple *tuple = arg->key;
	184	const struct flow_offload_tuple_rhash *x = ptr;
	185
	186	if (memcmp(&x->tuple, tuple, offsetof(struct flow_offload_tuple, dir)))
	187	return 1;
	188
	189	return 0;
	190	}
	191
	192	static const struct rhashtable_params nf_flow_offload_rhash_params = {
	193	.head_offset = offsetof(struct flow_offload_tuple_rhash, node),
	194	.hashfn = flow_offload_hash,
	195	.obj_hashfn = flow_offload_hash_obj,
	196	.obj_cmpfn = flow_offload_hash_cmp,
	197	.automatic_shrinking = true,
	198	};
	199
ac2a6666 PNA	200	int flow_offload_add(struct nf_flowtable flow_table, struct flow_offload flow)
ac2a6666 PNA	201	{
43c8f131 TY	202	int err;
	203
	204	err = rhashtable_insert_fast(&flow_table->rhashtable,
	205	&flow->tuplehash[0].node,
	206	nf_flow_offload_rhash_params);
	207	if (err < 0)
	208	return err;
	209
	210	err = rhashtable_insert_fast(&flow_table->rhashtable,
	211	&flow->tuplehash[1].node,
	212	nf_flow_offload_rhash_params);
	213	if (err < 0) {
	214	rhashtable_remove_fast(&flow_table->rhashtable,
	215	&flow->tuplehash[0].node,
	216	nf_flow_offload_rhash_params);
	217	return err;
	218	}
ac2a6666	219
43c8f131	220	flow->timeout = (u32)jiffies;
ac2a6666 PNA	221	return 0;
	222	}
	223	EXPORT_SYMBOL_GPL(flow_offload_add);
	224
3e68db2f PNA	225	static inline bool nf_flow_has_expired(const struct flow_offload *flow)
3e68db2f PNA	226	{
1e5b2471	227	return nf_flow_timeout_delta(flow->timeout) <= 0;
3e68db2f PNA	228	}
3e68db2f PNA	229
0ff90b6c FF	230	static void flow_offload_del(struct nf_flowtable *flow_table,
0ff90b6c FF	231	struct flow_offload *flow)
ac2a6666	232	{
da5984e5 FF	233	struct flow_offload_entry *e;
da5984e5 FF	234
ac2a6666 PNA	235	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	236	&flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].node,
a268de77	237	nf_flow_offload_rhash_params);
ac2a6666 PNA	238	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	239	&flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].node,
a268de77	240	nf_flow_offload_rhash_params);
ac2a6666	241
da5984e5 FF	242	e = container_of(flow, struct flow_offload_entry, flow);
	243	clear_bit(IPS_OFFLOAD_BIT, &e->ct->status);
	244
3e68db2f PNA	245	if (nf_flow_has_expired(flow))
3e68db2f PNA	246	flow_offload_fixup_ct(e->ct);
1e5b2471 PNA	247	else if (flow->flags & FLOW_OFFLOAD_TEARDOWN)
1e5b2471 PNA	248	flow_offload_fixup_ct_timeout(e->ct);
3e68db2f	249
0ff90b6c	250	flow_offload_free(flow);
ac2a6666	251	}
ac2a6666	252
59c466dd FF	253	void flow_offload_teardown(struct flow_offload *flow)
59c466dd FF	254	{
da5984e5 FF	255	struct flow_offload_entry *e;
da5984e5 FF	256
59c466dd	257	flow->flags \|= FLOW_OFFLOAD_TEARDOWN;
da5984e5 FF	258
da5984e5 FF	259	e = container_of(flow, struct flow_offload_entry, flow);
1e5b2471	260	flow_offload_fixup_ct_state(e->ct);
59c466dd FF	261	}
	262	EXPORT_SYMBOL_GPL(flow_offload_teardown);
	263
ac2a6666 PNA	264	struct flow_offload_tuple_rhash *
	265	flow_offload_lookup(struct nf_flowtable *flow_table,
	266	struct flow_offload_tuple *tuple)
	267	{
ba03137f FF	268	struct flow_offload_tuple_rhash *tuplehash;
ba03137f FF	269	struct flow_offload *flow;
8cd2bc98	270	struct flow_offload_entry *e;
ba03137f FF	271	int dir;
ba03137f FF	272
a2d88182 TY	273	tuplehash = rhashtable_lookup(&flow_table->rhashtable, tuple,
a2d88182 TY	274	nf_flow_offload_rhash_params);
ba03137f FF	275	if (!tuplehash)
	276	return NULL;
	277
	278	dir = tuplehash->tuple.dir;
	279	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	280	if (flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN))
	281	return NULL;
	282
8cd2bc98 TY	283	e = container_of(flow, struct flow_offload_entry, flow);
	284	if (unlikely(nf_ct_is_dying(e->ct)))
	285	return NULL;
	286
ba03137f	287	return tuplehash;
ac2a6666 PNA	288	}
	289	EXPORT_SYMBOL_GPL(flow_offload_lookup);
	290
49de9c09 TY	291	static int
	292	nf_flow_table_iterate(struct nf_flowtable *flow_table,
	293	void (iter)(struct flow_offload flow, void *data),
	294	void *data)
ac2a6666 PNA	295	{
	296	struct flow_offload_tuple_rhash *tuplehash;
	297	struct rhashtable_iter hti;
	298	struct flow_offload *flow;
0de22baa	299	int err = 0;
ac2a6666	300
0de22baa	301	rhashtable_walk_enter(&flow_table->rhashtable, &hti);
ac2a6666 PNA	302	rhashtable_walk_start(&hti);
	303
	304	while ((tuplehash = rhashtable_walk_next(&hti))) {
	305	if (IS_ERR(tuplehash)) {
0de22baa TY	306	if (PTR_ERR(tuplehash) != -EAGAIN) {
	307	err = PTR_ERR(tuplehash);
	308	break;
	309	}
ac2a6666 PNA	310	continue;
	311	}
	312	if (tuplehash->tuple.dir)
	313	continue;
	314
	315	flow = container_of(tuplehash, struct flow_offload, tuplehash[0]);
	316
	317	iter(flow, data);
	318	}
ac2a6666 PNA	319	rhashtable_walk_stop(&hti);
	320	rhashtable_walk_exit(&hti);
	321
	322	return err;
	323	}
ac2a6666	324
b9660987	325	static void nf_flow_offload_gc_step(struct flow_offload flow, void data)
ac2a6666	326	{
b9660987	327	struct nf_flowtable *flow_table = data;
8cd2bc98	328	struct flow_offload_entry *e;
ac2a6666	329
8cd2bc98 TY	330	e = container_of(flow, struct flow_offload_entry, flow);
8cd2bc98 TY	331	if (nf_flow_has_expired(flow) \|\| nf_ct_is_dying(e->ct) \|\|
b9660987 TY	332	(flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN)))
b9660987 TY	333	flow_offload_del(flow_table, flow);
b408c5b0 PNA	334	}
b408c5b0 PNA	335
a268de77	336	static void nf_flow_offload_work_gc(struct work_struct *work)
b408c5b0 PNA	337	{
	338	struct nf_flowtable *flow_table;
	339
	340	flow_table = container_of(work, struct nf_flowtable, gc_work.work);
b9660987	341	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
ac2a6666 PNA	342	queue_delayed_work(system_power_efficient_wq, &flow_table->gc_work, HZ);
ac2a6666 PNA	343	}
ac2a6666 PNA	344
	345	static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
	346	__be16 port, __be16 new_port)
	347	{
	348	struct tcphdr *tcph;
	349
	350	if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) \|\|
	351	skb_try_make_writable(skb, thoff + sizeof(*tcph)))
	352	return -1;
	353
	354	tcph = (void *)(skb_network_header(skb) + thoff);
	355	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, true);
	356
	357	return 0;
	358	}
	359
	360	static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
	361	__be16 port, __be16 new_port)
	362	{
	363	struct udphdr *udph;
	364
	365	if (!pskb_may_pull(skb, thoff + sizeof(*udph)) \|\|
	366	skb_try_make_writable(skb, thoff + sizeof(*udph)))
	367	return -1;
	368
	369	udph = (void *)(skb_network_header(skb) + thoff);
	370	if (udph->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	371	inet_proto_csum_replace2(&udph->check, skb, port,
	372	new_port, true);
	373	if (!udph->check)
	374	udph->check = CSUM_MANGLED_0;
	375	}
	376
	377	return 0;
	378	}
	379
	380	static int nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
	381	u8 protocol, __be16 port, __be16 new_port)
	382	{
	383	switch (protocol) {
	384	case IPPROTO_TCP:
	385	if (nf_flow_nat_port_tcp(skb, thoff, port, new_port) < 0)
	386	return NF_DROP;
	387	break;
	388	case IPPROTO_UDP:
	389	if (nf_flow_nat_port_udp(skb, thoff, port, new_port) < 0)
	390	return NF_DROP;
	391	break;
	392	}
	393
	394	return 0;
	395	}
	396
	397	int nf_flow_snat_port(const struct flow_offload *flow,
	398	struct sk_buff *skb, unsigned int thoff,
	399	u8 protocol, enum flow_offload_tuple_dir dir)
	400	{
	401	struct flow_ports *hdr;
	402	__be16 port, new_port;
	403
	404	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
	405	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
	406	return -1;
	407
408	hdr = (void *)(skb_network_header(skb) + thoff);
409
410	switch (dir) {
411	case FLOW_OFFLOAD_DIR_ORIGINAL:
412	port = hdr->source;
413	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_port;
414	hdr->source = new_port;
415	break;
416	case FLOW_OFFLOAD_DIR_REPLY:
417	port = hdr->dest;
418	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_port;
419	hdr->dest = new_port;
420	break;
421	default:
422	return -1;
423	}
424
425	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
426	}
427	EXPORT_SYMBOL_GPL(nf_flow_snat_port);
428
429	int nf_flow_dnat_port(const struct flow_offload *flow,
430	struct sk_buff *skb, unsigned int thoff,
431	u8 protocol, enum flow_offload_tuple_dir dir)
432	{
433	struct flow_ports *hdr;
434	__be16 port, new_port;
435
436	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
437	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
438	return -1;
439
440	hdr = (void *)(skb_network_header(skb) + thoff);
441
442	switch (dir) {
443	case FLOW_OFFLOAD_DIR_ORIGINAL:
444	port = hdr->dest;
445	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_port;
446	hdr->dest = new_port;
447	break;
448	case FLOW_OFFLOAD_DIR_REPLY:
449	port = hdr->source;
450	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_port;
451	hdr->source = new_port;
452	break;
453	default:
454	return -1;
455	}
456
457	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
458	}
459	EXPORT_SYMBOL_GPL(nf_flow_dnat_port);
460
a268de77 FF	461	int nf_flow_table_init(struct nf_flowtable *flowtable)
	462	{
	463	int err;
	464
	465	INIT_DEFERRABLE_WORK(&flowtable->gc_work, nf_flow_offload_work_gc);
	466
	467	err = rhashtable_init(&flowtable->rhashtable,
	468	&nf_flow_offload_rhash_params);
	469	if (err < 0)
	470	return err;
	471
	472	queue_delayed_work(system_power_efficient_wq,
	473	&flowtable->gc_work, HZ);
	474
84453a90 FF	475	mutex_lock(&flowtable_lock);
	476	list_add(&flowtable->list, &flowtables);
	477	mutex_unlock(&flowtable_lock);
	478
a268de77 FF	479	return 0;
	480	}
	481	EXPORT_SYMBOL_GPL(nf_flow_table_init);
	482
c0ea1bcb PNA	483	static void nf_flow_table_do_cleanup(struct flow_offload flow, void data)
	484	{
	485	struct net_device *dev = data;
a3fb3698 TY	486	struct flow_offload_entry *e;
	487
	488	e = container_of(flow, struct flow_offload_entry, flow);
c0ea1bcb	489
59c466dd FF	490	if (!dev) {
59c466dd FF	491	flow_offload_teardown(flow);
c0ea1bcb	492	return;
59c466dd	493	}
a3fb3698 TY	494	if (net_eq(nf_ct_net(e->ct), dev_net(dev)) &&
	495	(flow->tuplehash[0].tuple.iifidx == dev->ifindex \|\|
	496	flow->tuplehash[1].tuple.iifidx == dev->ifindex))
59c466dd	497	flow_offload_dead(flow);
c0ea1bcb PNA	498	}
	499
	500	static void nf_flow_table_iterate_cleanup(struct nf_flowtable *flowtable,
84453a90	501	struct net_device *dev)
c0ea1bcb	502	{
84453a90	503	nf_flow_table_iterate(flowtable, nf_flow_table_do_cleanup, dev);
c0ea1bcb PNA	504	flush_delayed_work(&flowtable->gc_work);
	505	}
	506
5f1be84a	507	void nf_flow_table_cleanup(struct net_device *dev)
c0ea1bcb	508	{
84453a90 FF	509	struct nf_flowtable *flowtable;
	510
	511	mutex_lock(&flowtable_lock);
	512	list_for_each_entry(flowtable, &flowtables, list)
	513	nf_flow_table_iterate_cleanup(flowtable, dev);
	514	mutex_unlock(&flowtable_lock);
c0ea1bcb PNA	515	}
	516	EXPORT_SYMBOL_GPL(nf_flow_table_cleanup);
	517
b408c5b0 PNA	518	void nf_flow_table_free(struct nf_flowtable *flow_table)
b408c5b0 PNA	519	{
84453a90 FF	520	mutex_lock(&flowtable_lock);
	521	list_del(&flow_table->list);
	522	mutex_unlock(&flowtable_lock);
a268de77	523	cancel_delayed_work_sync(&flow_table->gc_work);
b408c5b0	524	nf_flow_table_iterate(flow_table, nf_flow_table_do_cleanup, NULL);
b9660987	525	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
a268de77	526	rhashtable_destroy(&flow_table->rhashtable);
b408c5b0 PNA	527	}
	528	EXPORT_SYMBOL_GPL(nf_flow_table_free);
	529
ac2a6666 PNA	530	MODULE_LICENSE("GPL");
ac2a6666 PNA	531	MODULE_AUTHOR("Pablo Neira Ayuso <pablo@netfilter.org>");