[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_l4proto.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 + NF_FLOW_TIMEOUT;
	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>
40d102cd	14	#include <net/netfilter/nf_conntrack_l4proto.h>
ac2a6666 PNA	15	#include <net/netfilter/nf_conntrack_tuple.h>
	16
	17	struct flow_offload_entry {
	18	struct flow_offload flow;
	19	struct nf_conn *ct;
	20	struct rcu_head rcu_head;
	21	};
	22
84453a90 FF	23	static DEFINE_MUTEX(flowtable_lock);
	24	static LIST_HEAD(flowtables);
	25
047b300e FF	26	static void
	27	flow_offload_fill_dir(struct flow_offload flow, struct nf_conn ct,
	28	struct nf_flow_route *route,
	29	enum flow_offload_tuple_dir dir)
	30	{
	31	struct flow_offload_tuple *ft = &flow->tuplehash[dir].tuple;
	32	struct nf_conntrack_tuple *ctt = &ct->tuplehash[dir].tuple;
10f4e765	33	struct dst_entry *other_dst = route->tuple[!dir].dst;
4f3780c0	34	struct dst_entry *dst = route->tuple[dir].dst;
047b300e FF	35
	36	ft->dir = dir;
	37
	38	switch (ctt->src.l3num) {
	39	case NFPROTO_IPV4:
	40	ft->src_v4 = ctt->src.u3.in;
	41	ft->dst_v4 = ctt->dst.u3.in;
4f3780c0	42	ft->mtu = ip_dst_mtu_maybe_forward(dst, true);
047b300e FF	43	break;
	44	case NFPROTO_IPV6:
	45	ft->src_v6 = ctt->src.u3.in6;
	46	ft->dst_v6 = ctt->dst.u3.in6;
4f3780c0	47	ft->mtu = ip6_dst_mtu_forward(dst);
047b300e FF	48	break;
	49	}
	50
	51	ft->l3proto = ctt->src.l3num;
	52	ft->l4proto = ctt->dst.protonum;
	53	ft->src_port = ctt->src.u.tcp.port;
	54	ft->dst_port = ctt->dst.u.tcp.port;
	55
10f4e765	56	ft->iifidx = other_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
1e5b2471 PNA	115	static inline __s32 nf_flow_timeout_delta(unsigned int timeout)
	116	{
	117	return (__s32)(timeout - (u32)jiffies);
	118	}
	119
	120	static void flow_offload_fixup_ct_timeout(struct nf_conn *ct)
da5984e5 FF	121	{
da5984e5 FF	122	const struct nf_conntrack_l4proto *l4proto;
1e5b2471	123	int l4num = nf_ct_protonum(ct);
da5984e5	124	unsigned int timeout;
da5984e5	125
4a60dc74	126	l4proto = nf_ct_l4proto_find(l4num);
da5984e5 FF	127	if (!l4proto)
	128	return;
	129
da5984e5	130	if (l4num == IPPROTO_TCP)
e97d9404	131	timeout = NF_FLOWTABLE_TCP_PICKUP_TIMEOUT;
da5984e5	132	else if (l4num == IPPROTO_UDP)
e97d9404	133	timeout = NF_FLOWTABLE_UDP_PICKUP_TIMEOUT;
da5984e5 FF	134	else
	135	return;
	136
1e5b2471 PNA	137	if (nf_flow_timeout_delta(ct->timeout) > (__s32)timeout)
	138	ct->timeout = nfct_time_stamp + timeout;
	139	}
	140
	141	static void flow_offload_fixup_ct_state(struct nf_conn *ct)
	142	{
	143	if (nf_ct_protonum(ct) == IPPROTO_TCP)
	144	flow_offload_fixup_tcp(&ct->proto.tcp);
	145	}
	146
	147	static void flow_offload_fixup_ct(struct nf_conn *ct)
	148	{
	149	flow_offload_fixup_ct_state(ct);
	150	flow_offload_fixup_ct_timeout(ct);
da5984e5 FF	151	}
da5984e5 FF	152
ac2a6666 PNA	153	void flow_offload_free(struct flow_offload *flow)
	154	{
	155	struct flow_offload_entry *e;
	156
	157	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_cache);
	158	dst_release(flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_cache);
	159	e = container_of(flow, struct flow_offload_entry, flow);
da5984e5 FF	160	if (flow->flags & FLOW_OFFLOAD_DYING)
da5984e5 FF	161	nf_ct_delete(e->ct, 0, 0);
0ff90b6c FF	162	nf_ct_put(e->ct);
0ff90b6c FF	163	kfree_rcu(e, rcu_head);
ac2a6666 PNA	164	}
	165	EXPORT_SYMBOL_GPL(flow_offload_free);
	166
a268de77 FF	167	static u32 flow_offload_hash(const void *data, u32 len, u32 seed)
	168	{
	169	const struct flow_offload_tuple *tuple = data;
	170
	171	return jhash(tuple, offsetof(struct flow_offload_tuple, dir), seed);
	172	}
	173
	174	static u32 flow_offload_hash_obj(const void *data, u32 len, u32 seed)
	175	{
	176	const struct flow_offload_tuple_rhash *tuplehash = data;
	177
	178	return jhash(&tuplehash->tuple, offsetof(struct flow_offload_tuple, dir), seed);
	179	}
	180
	181	static int flow_offload_hash_cmp(struct rhashtable_compare_arg *arg,
	182	const void *ptr)
	183	{
	184	const struct flow_offload_tuple *tuple = arg->key;
	185	const struct flow_offload_tuple_rhash *x = ptr;
	186
	187	if (memcmp(&x->tuple, tuple, offsetof(struct flow_offload_tuple, dir)))
	188	return 1;
	189
	190	return 0;
	191	}
	192
	193	static const struct rhashtable_params nf_flow_offload_rhash_params = {
	194	.head_offset = offsetof(struct flow_offload_tuple_rhash, node),
	195	.hashfn = flow_offload_hash,
	196	.obj_hashfn = flow_offload_hash_obj,
	197	.obj_cmpfn = flow_offload_hash_cmp,
	198	.automatic_shrinking = true,
	199	};
	200
ac2a6666 PNA	201	int flow_offload_add(struct nf_flowtable flow_table, struct flow_offload flow)
ac2a6666 PNA	202	{
43c8f131 TY	203	int err;
	204
	205	err = rhashtable_insert_fast(&flow_table->rhashtable,
	206	&flow->tuplehash[0].node,
	207	nf_flow_offload_rhash_params);
	208	if (err < 0)
	209	return err;
	210
	211	err = rhashtable_insert_fast(&flow_table->rhashtable,
	212	&flow->tuplehash[1].node,
	213	nf_flow_offload_rhash_params);
	214	if (err < 0) {
	215	rhashtable_remove_fast(&flow_table->rhashtable,
	216	&flow->tuplehash[0].node,
	217	nf_flow_offload_rhash_params);
	218	return err;
	219	}
ac2a6666	220
110e4872	221	flow->timeout = (u32)jiffies + NF_FLOW_TIMEOUT;
ac2a6666 PNA	222	return 0;
	223	}
	224	EXPORT_SYMBOL_GPL(flow_offload_add);
	225
3e68db2f PNA	226	static inline bool nf_flow_has_expired(const struct flow_offload *flow)
3e68db2f PNA	227	{
1e5b2471	228	return nf_flow_timeout_delta(flow->timeout) <= 0;
3e68db2f PNA	229	}
3e68db2f PNA	230
0ff90b6c FF	231	static void flow_offload_del(struct nf_flowtable *flow_table,
0ff90b6c FF	232	struct flow_offload *flow)
ac2a6666	233	{
da5984e5 FF	234	struct flow_offload_entry *e;
da5984e5 FF	235
ac2a6666 PNA	236	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	237	&flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].node,
a268de77	238	nf_flow_offload_rhash_params);
ac2a6666 PNA	239	rhashtable_remove_fast(&flow_table->rhashtable,
ac2a6666 PNA	240	&flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].node,
a268de77	241	nf_flow_offload_rhash_params);
ac2a6666	242
da5984e5 FF	243	e = container_of(flow, struct flow_offload_entry, flow);
	244	clear_bit(IPS_OFFLOAD_BIT, &e->ct->status);
	245
3e68db2f PNA	246	if (nf_flow_has_expired(flow))
3e68db2f PNA	247	flow_offload_fixup_ct(e->ct);
1e5b2471 PNA	248	else if (flow->flags & FLOW_OFFLOAD_TEARDOWN)
1e5b2471 PNA	249	flow_offload_fixup_ct_timeout(e->ct);
3e68db2f	250
0ff90b6c	251	flow_offload_free(flow);
ac2a6666	252	}
ac2a6666	253
59c466dd FF	254	void flow_offload_teardown(struct flow_offload *flow)
59c466dd FF	255	{
da5984e5 FF	256	struct flow_offload_entry *e;
da5984e5 FF	257
59c466dd	258	flow->flags \|= FLOW_OFFLOAD_TEARDOWN;
da5984e5 FF	259
da5984e5 FF	260	e = container_of(flow, struct flow_offload_entry, flow);
1e5b2471	261	flow_offload_fixup_ct_state(e->ct);
59c466dd FF	262	}
	263	EXPORT_SYMBOL_GPL(flow_offload_teardown);
	264
ac2a6666 PNA	265	struct flow_offload_tuple_rhash *
	266	flow_offload_lookup(struct nf_flowtable *flow_table,
	267	struct flow_offload_tuple *tuple)
	268	{
ba03137f FF	269	struct flow_offload_tuple_rhash *tuplehash;
ba03137f FF	270	struct flow_offload *flow;
8cd2bc98	271	struct flow_offload_entry *e;
ba03137f FF	272	int dir;
ba03137f FF	273
a2d88182 TY	274	tuplehash = rhashtable_lookup(&flow_table->rhashtable, tuple,
a2d88182 TY	275	nf_flow_offload_rhash_params);
ba03137f FF	276	if (!tuplehash)
	277	return NULL;
	278
	279	dir = tuplehash->tuple.dir;
	280	flow = container_of(tuplehash, struct flow_offload, tuplehash[dir]);
	281	if (flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN))
	282	return NULL;
	283
8cd2bc98 TY	284	e = container_of(flow, struct flow_offload_entry, flow);
	285	if (unlikely(nf_ct_is_dying(e->ct)))
	286	return NULL;
	287
ba03137f	288	return tuplehash;
ac2a6666 PNA	289	}
	290	EXPORT_SYMBOL_GPL(flow_offload_lookup);
	291
49de9c09 TY	292	static int
	293	nf_flow_table_iterate(struct nf_flowtable *flow_table,
	294	void (iter)(struct flow_offload flow, void *data),
	295	void *data)
ac2a6666 PNA	296	{
	297	struct flow_offload_tuple_rhash *tuplehash;
	298	struct rhashtable_iter hti;
	299	struct flow_offload *flow;
0de22baa	300	int err = 0;
ac2a6666	301
0de22baa	302	rhashtable_walk_enter(&flow_table->rhashtable, &hti);
ac2a6666 PNA	303	rhashtable_walk_start(&hti);
	304
	305	while ((tuplehash = rhashtable_walk_next(&hti))) {
	306	if (IS_ERR(tuplehash)) {
0de22baa TY	307	if (PTR_ERR(tuplehash) != -EAGAIN) {
	308	err = PTR_ERR(tuplehash);
	309	break;
	310	}
ac2a6666 PNA	311	continue;
	312	}
	313	if (tuplehash->tuple.dir)
	314	continue;
	315
	316	flow = container_of(tuplehash, struct flow_offload, tuplehash[0]);
	317
	318	iter(flow, data);
	319	}
ac2a6666 PNA	320	rhashtable_walk_stop(&hti);
	321	rhashtable_walk_exit(&hti);
	322
	323	return err;
	324	}
ac2a6666	325
b9660987	326	static void nf_flow_offload_gc_step(struct flow_offload flow, void data)
ac2a6666	327	{
b9660987	328	struct nf_flowtable *flow_table = data;
8cd2bc98	329	struct flow_offload_entry *e;
ac2a6666	330
8cd2bc98 TY	331	e = container_of(flow, struct flow_offload_entry, flow);
8cd2bc98 TY	332	if (nf_flow_has_expired(flow) \|\| nf_ct_is_dying(e->ct) \|\|
b9660987 TY	333	(flow->flags & (FLOW_OFFLOAD_DYING \| FLOW_OFFLOAD_TEARDOWN)))
b9660987 TY	334	flow_offload_del(flow_table, flow);
b408c5b0 PNA	335	}
b408c5b0 PNA	336
a268de77	337	static void nf_flow_offload_work_gc(struct work_struct *work)
b408c5b0 PNA	338	{
	339	struct nf_flowtable *flow_table;
	340
	341	flow_table = container_of(work, struct nf_flowtable, gc_work.work);
b9660987	342	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
ac2a6666 PNA	343	queue_delayed_work(system_power_efficient_wq, &flow_table->gc_work, HZ);
ac2a6666 PNA	344	}
ac2a6666 PNA	345
	346	static int nf_flow_nat_port_tcp(struct sk_buff *skb, unsigned int thoff,
	347	__be16 port, __be16 new_port)
	348	{
	349	struct tcphdr *tcph;
	350
	351	if (!pskb_may_pull(skb, thoff + sizeof(*tcph)) \|\|
	352	skb_try_make_writable(skb, thoff + sizeof(*tcph)))
	353	return -1;
	354
	355	tcph = (void *)(skb_network_header(skb) + thoff);
	356	inet_proto_csum_replace2(&tcph->check, skb, port, new_port, true);
	357
	358	return 0;
	359	}
	360
	361	static int nf_flow_nat_port_udp(struct sk_buff *skb, unsigned int thoff,
	362	__be16 port, __be16 new_port)
	363	{
	364	struct udphdr *udph;
	365
	366	if (!pskb_may_pull(skb, thoff + sizeof(*udph)) \|\|
	367	skb_try_make_writable(skb, thoff + sizeof(*udph)))
	368	return -1;
	369
	370	udph = (void *)(skb_network_header(skb) + thoff);
	371	if (udph->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	372	inet_proto_csum_replace2(&udph->check, skb, port,
	373	new_port, true);
	374	if (!udph->check)
	375	udph->check = CSUM_MANGLED_0;
	376	}
	377
	378	return 0;
	379	}
	380
	381	static int nf_flow_nat_port(struct sk_buff *skb, unsigned int thoff,
	382	u8 protocol, __be16 port, __be16 new_port)
	383	{
	384	switch (protocol) {
	385	case IPPROTO_TCP:
	386	if (nf_flow_nat_port_tcp(skb, thoff, port, new_port) < 0)
	387	return NF_DROP;
	388	break;
	389	case IPPROTO_UDP:
	390	if (nf_flow_nat_port_udp(skb, thoff, port, new_port) < 0)
	391	return NF_DROP;
	392	break;
	393	}
	394
	395	return 0;
	396	}
	397
	398	int nf_flow_snat_port(const struct flow_offload *flow,
	399	struct sk_buff *skb, unsigned int thoff,
	400	u8 protocol, enum flow_offload_tuple_dir dir)
	401	{
	402	struct flow_ports *hdr;
	403	__be16 port, new_port;
	404
	405	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
	406	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
	407	return -1;
	408
409	hdr = (void *)(skb_network_header(skb) + thoff);
410
411	switch (dir) {
412	case FLOW_OFFLOAD_DIR_ORIGINAL:
413	port = hdr->source;
414	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.dst_port;
415	hdr->source = new_port;
416	break;
417	case FLOW_OFFLOAD_DIR_REPLY:
418	port = hdr->dest;
419	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.src_port;
420	hdr->dest = new_port;
421	break;
422	default:
423	return -1;
424	}
425
426	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
427	}
428	EXPORT_SYMBOL_GPL(nf_flow_snat_port);
429
430	int nf_flow_dnat_port(const struct flow_offload *flow,
431	struct sk_buff *skb, unsigned int thoff,
432	u8 protocol, enum flow_offload_tuple_dir dir)
433	{
434	struct flow_ports *hdr;
435	__be16 port, new_port;
436
437	if (!pskb_may_pull(skb, thoff + sizeof(*hdr)) \|\|
438	skb_try_make_writable(skb, thoff + sizeof(*hdr)))
439	return -1;
440
441	hdr = (void *)(skb_network_header(skb) + thoff);
442
443	switch (dir) {
444	case FLOW_OFFLOAD_DIR_ORIGINAL:
445	port = hdr->dest;
446	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_REPLY].tuple.src_port;
447	hdr->dest = new_port;
448	break;
449	case FLOW_OFFLOAD_DIR_REPLY:
450	port = hdr->source;
451	new_port = flow->tuplehash[FLOW_OFFLOAD_DIR_ORIGINAL].tuple.dst_port;
452	hdr->source = new_port;
453	break;
454	default:
455	return -1;
456	}
457
458	return nf_flow_nat_port(skb, thoff, protocol, port, new_port);
459	}
460	EXPORT_SYMBOL_GPL(nf_flow_dnat_port);
461
a268de77 FF	462	int nf_flow_table_init(struct nf_flowtable *flowtable)
	463	{
	464	int err;
	465
	466	INIT_DEFERRABLE_WORK(&flowtable->gc_work, nf_flow_offload_work_gc);
	467
	468	err = rhashtable_init(&flowtable->rhashtable,
	469	&nf_flow_offload_rhash_params);
	470	if (err < 0)
	471	return err;
	472
	473	queue_delayed_work(system_power_efficient_wq,
	474	&flowtable->gc_work, HZ);
	475
84453a90 FF	476	mutex_lock(&flowtable_lock);
	477	list_add(&flowtable->list, &flowtables);
	478	mutex_unlock(&flowtable_lock);
	479
a268de77 FF	480	return 0;
	481	}
	482	EXPORT_SYMBOL_GPL(nf_flow_table_init);
	483
c0ea1bcb PNA	484	static void nf_flow_table_do_cleanup(struct flow_offload flow, void data)
	485	{
	486	struct net_device *dev = data;
a3fb3698 TY	487	struct flow_offload_entry *e;
	488
	489	e = container_of(flow, struct flow_offload_entry, flow);
c0ea1bcb	490
59c466dd FF	491	if (!dev) {
59c466dd FF	492	flow_offload_teardown(flow);
c0ea1bcb	493	return;
59c466dd	494	}
a3fb3698 TY	495	if (net_eq(nf_ct_net(e->ct), dev_net(dev)) &&
	496	(flow->tuplehash[0].tuple.iifidx == dev->ifindex \|\|
	497	flow->tuplehash[1].tuple.iifidx == dev->ifindex))
59c466dd	498	flow_offload_dead(flow);
c0ea1bcb PNA	499	}
	500
	501	static void nf_flow_table_iterate_cleanup(struct nf_flowtable *flowtable,
84453a90	502	struct net_device *dev)
c0ea1bcb	503	{
84453a90	504	nf_flow_table_iterate(flowtable, nf_flow_table_do_cleanup, dev);
c0ea1bcb PNA	505	flush_delayed_work(&flowtable->gc_work);
	506	}
	507
5f1be84a	508	void nf_flow_table_cleanup(struct net_device *dev)
c0ea1bcb	509	{
84453a90 FF	510	struct nf_flowtable *flowtable;
	511
	512	mutex_lock(&flowtable_lock);
	513	list_for_each_entry(flowtable, &flowtables, list)
	514	nf_flow_table_iterate_cleanup(flowtable, dev);
	515	mutex_unlock(&flowtable_lock);
c0ea1bcb PNA	516	}
	517	EXPORT_SYMBOL_GPL(nf_flow_table_cleanup);
	518
b408c5b0 PNA	519	void nf_flow_table_free(struct nf_flowtable *flow_table)
b408c5b0 PNA	520	{
84453a90 FF	521	mutex_lock(&flowtable_lock);
	522	list_del(&flow_table->list);
	523	mutex_unlock(&flowtable_lock);
a268de77	524	cancel_delayed_work_sync(&flow_table->gc_work);
b408c5b0	525	nf_flow_table_iterate(flow_table, nf_flow_table_do_cleanup, NULL);
b9660987	526	nf_flow_table_iterate(flow_table, nf_flow_offload_gc_step, flow_table);
a268de77	527	rhashtable_destroy(&flow_table->rhashtable);
b408c5b0 PNA	528	}
	529	EXPORT_SYMBOL_GPL(nf_flow_table_free);
	530
ac2a6666 PNA	531	MODULE_LICENSE("GPL");
ac2a6666 PNA	532	MODULE_AUTHOR("Pablo Neira Ayuso <pablo@netfilter.org>");