[net-next-2.6.git] / net / core / fib_rules.c

/*
 * net/core/fib_rules.c		Generic Routing Rules
 *
 *	This program is free software; you can redistribute it and/or
 *	modify it under the terms of the GNU General Public License as
 *	published by the Free Software Foundation, version 2.
 *
 * Authors:	Thomas Graf <tgraf@suug.ch>
 */

#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/list.h>
#include <net/net_namespace.h>
#include <net/sock.h>
#include <net/fib_rules.h>

static LIST_HEAD(rules_ops);
static DEFINE_SPINLOCK(rules_mod_lock);

int fib_default_rule_add(struct fib_rules_ops *ops,
			 u32 pref, u32 table, u32 flags)
{
	struct fib_rule *r;

	r = kzalloc(ops->rule_size, GFP_KERNEL);
	if (r == NULL)
		return -ENOMEM;

	atomic_set(&r->refcnt, 1);
	r->action = FR_ACT_TO_TBL;
	r->pref = pref;
	r->table = table;
	r->flags = flags;

	/* The lock is not required here, the list in unreacheable
	 * at the moment this function is called */
	list_add_tail(&r->list, &ops->rules_list);
	return 0;
}
EXPORT_SYMBOL(fib_default_rule_add);

static void notify_rule_change(int event, struct fib_rule *rule,
			       struct fib_rules_ops *ops, struct nlmsghdr *nlh,
			       u32 pid);

static struct fib_rules_ops *lookup_rules_ops(int family)
{
	struct fib_rules_ops *ops;

	rcu_read_lock();
	list_for_each_entry_rcu(ops, &rules_ops, list) {
		if (ops->family == family) {
			if (!try_module_get(ops->owner))
				ops = NULL;
			rcu_read_unlock();
			return ops;
		}
	}
	rcu_read_unlock();

	return NULL;
}

static void rules_ops_put(struct fib_rules_ops *ops)
{
	if (ops)
		module_put(ops->owner);
}

static void flush_route_cache(struct fib_rules_ops *ops)
{
	if (ops->flush_cache)
		ops->flush_cache();
}

int fib_rules_register(struct net *net, struct fib_rules_ops *ops)
{
	int err = -EEXIST;
	struct fib_rules_ops *o;

	if (ops->rule_size < sizeof(struct fib_rule))
		return -EINVAL;

	if (ops->match == NULL || ops->configure == NULL ||
	    ops->compare == NULL || ops->fill == NULL ||
	    ops->action == NULL)
		return -EINVAL;

	spin_lock(&rules_mod_lock);
	list_for_each_entry(o, &rules_ops, list)
		if (ops->family == o->family)
			goto errout;

	list_add_tail_rcu(&ops->list, &rules_ops);
	err = 0;
errout:
	spin_unlock(&rules_mod_lock);

	return err;
}

EXPORT_SYMBOL_GPL(fib_rules_register);

void fib_rules_cleanup_ops(struct fib_rules_ops *ops)
{
	struct fib_rule *rule, *tmp;

	list_for_each_entry_safe(rule, tmp, &ops->rules_list, list) {
		list_del_rcu(&rule->list);
		fib_rule_put(rule);
	}
}
EXPORT_SYMBOL_GPL(fib_rules_cleanup_ops);

int fib_rules_unregister(struct net *net, struct fib_rules_ops *ops)
{
	int err = 0;
	struct fib_rules_ops *o;

	spin_lock(&rules_mod_lock);
	list_for_each_entry(o, &rules_ops, list) {
		if (o == ops) {
			list_del_rcu(&o->list);
			fib_rules_cleanup_ops(ops);
			goto out;
		}
	}

	err = -ENOENT;
out:
	spin_unlock(&rules_mod_lock);

	synchronize_rcu();

	return err;
}

EXPORT_SYMBOL_GPL(fib_rules_unregister);

static int fib_rule_match(struct fib_rule *rule, struct fib_rules_ops *ops,
			  struct flowi *fl, int flags)
{
	int ret = 0;

	if (rule->ifindex && (rule->ifindex != fl->iif))
		goto out;

	if ((rule->mark ^ fl->mark) & rule->mark_mask)
		goto out;

	ret = ops->match(rule, fl, flags);
out:
	return (rule->flags & FIB_RULE_INVERT) ? !ret : ret;
}

int fib_rules_lookup(struct fib_rules_ops *ops, struct flowi *fl,
		     int flags, struct fib_lookup_arg *arg)
{
	struct fib_rule *rule;
	int err;

	rcu_read_lock();

	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
jumped:
		if (!fib_rule_match(rule, ops, fl, flags))
			continue;

		if (rule->action == FR_ACT_GOTO) {
			struct fib_rule *target;

			target = rcu_dereference(rule->ctarget);
			if (target == NULL) {
				continue;
			} else {
				rule = target;
				goto jumped;
			}
		} else if (rule->action == FR_ACT_NOP)
			continue;
		else
			err = ops->action(rule, fl, flags, arg);

		if (err != -EAGAIN) {
			fib_rule_get(rule);
			arg->rule = rule;
			goto out;
		}
	}

	err = -ESRCH;
out:
	rcu_read_unlock();

	return err;
}

EXPORT_SYMBOL_GPL(fib_rules_lookup);

static int validate_rulemsg(struct fib_rule_hdr *frh, struct nlattr **tb,
			    struct fib_rules_ops *ops)
{
	int err = -EINVAL;

	if (frh->src_len)
		if (tb[FRA_SRC] == NULL ||
		    frh->src_len > (ops->addr_size * 8) ||
		    nla_len(tb[FRA_SRC]) != ops->addr_size)
			goto errout;

	if (frh->dst_len)
		if (tb[FRA_DST] == NULL ||
		    frh->dst_len > (ops->addr_size * 8) ||
		    nla_len(tb[FRA_DST]) != ops->addr_size)
			goto errout;

	err = 0;
errout:
	return err;
}

static int fib_nl_newrule(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
{
	struct net *net = skb->sk->sk_net;
	struct fib_rule_hdr *frh = nlmsg_data(nlh);
	struct fib_rules_ops *ops = NULL;
	struct fib_rule *rule, *r, *last = NULL;
	struct nlattr *tb[FRA_MAX+1];
	int err = -EINVAL, unresolved = 0;

	if (net != &init_net)
		return -EINVAL;

	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
		goto errout;

	ops = lookup_rules_ops(frh->family);
	if (ops == NULL) {
		err = EAFNOSUPPORT;
		goto errout;
	}

	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	if (err < 0)
		goto errout;

	err = validate_rulemsg(frh, tb, ops);
	if (err < 0)
		goto errout;

	rule = kzalloc(ops->rule_size, GFP_KERNEL);
	if (rule == NULL) {
		err = -ENOMEM;
		goto errout;
	}

	if (tb[FRA_PRIORITY])
		rule->pref = nla_get_u32(tb[FRA_PRIORITY]);

	if (tb[FRA_IFNAME]) {
		struct net_device *dev;

		rule->ifindex = -1;
		nla_strlcpy(rule->ifname, tb[FRA_IFNAME], IFNAMSIZ);
		dev = __dev_get_by_name(net, rule->ifname);
		if (dev)
			rule->ifindex = dev->ifindex;
	}

	if (tb[FRA_FWMARK]) {
		rule->mark = nla_get_u32(tb[FRA_FWMARK]);
		if (rule->mark)
			/* compatibility: if the mark value is non-zero all bits
			 * are compared unless a mask is explicitly specified.
			 */
			rule->mark_mask = 0xFFFFFFFF;
	}

	if (tb[FRA_FWMASK])
		rule->mark_mask = nla_get_u32(tb[FRA_FWMASK]);

	rule->action = frh->action;
	rule->flags = frh->flags;
	rule->table = frh_get_table(frh, tb);

	if (!rule->pref && ops->default_pref)
		rule->pref = ops->default_pref(ops);

	err = -EINVAL;
	if (tb[FRA_GOTO]) {
		if (rule->action != FR_ACT_GOTO)
			goto errout_free;

		rule->target = nla_get_u32(tb[FRA_GOTO]);
		/* Backward jumps are prohibited to avoid endless loops */
		if (rule->target <= rule->pref)
			goto errout_free;

		list_for_each_entry(r, &ops->rules_list, list) {
			if (r->pref == rule->target) {
				rule->ctarget = r;
				break;
			}
		}

		if (rule->ctarget == NULL)
			unresolved = 1;
	} else if (rule->action == FR_ACT_GOTO)
		goto errout_free;

	err = ops->configure(rule, skb, nlh, frh, tb);
	if (err < 0)
		goto errout_free;

	list_for_each_entry(r, &ops->rules_list, list) {
		if (r->pref > rule->pref)
			break;
		last = r;
	}

	fib_rule_get(rule);

	if (ops->unresolved_rules) {
		/*
		 * There are unresolved goto rules in the list, check if
		 * any of them are pointing to this new rule.
		 */
		list_for_each_entry(r, &ops->rules_list, list) {
			if (r->action == FR_ACT_GOTO &&
			    r->target == rule->pref) {
				BUG_ON(r->ctarget != NULL);
				rcu_assign_pointer(r->ctarget, rule);
				if (--ops->unresolved_rules == 0)
					break;
			}
		}
	}

	if (rule->action == FR_ACT_GOTO)
		ops->nr_goto_rules++;

	if (unresolved)
		ops->unresolved_rules++;

	if (last)
		list_add_rcu(&rule->list, &last->list);
	else
		list_add_rcu(&rule->list, &ops->rules_list);

	notify_rule_change(RTM_NEWRULE, rule, ops, nlh, NETLINK_CB(skb).pid);
	flush_route_cache(ops);
	rules_ops_put(ops);
	return 0;

errout_free:
	kfree(rule);
errout:
	rules_ops_put(ops);
	return err;
}

static int fib_nl_delrule(struct sk_buff *skb, struct nlmsghdr* nlh, void *arg)
{
	struct net *net = skb->sk->sk_net;
	struct fib_rule_hdr *frh = nlmsg_data(nlh);
	struct fib_rules_ops *ops = NULL;
	struct fib_rule *rule, *tmp;
	struct nlattr *tb[FRA_MAX+1];
	int err = -EINVAL;

	if (net != &init_net)
		return -EINVAL;

	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
		goto errout;

	ops = lookup_rules_ops(frh->family);
	if (ops == NULL) {
		err = EAFNOSUPPORT;
		goto errout;
	}

	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	if (err < 0)
		goto errout;

	err = validate_rulemsg(frh, tb, ops);
	if (err < 0)
		goto errout;

	list_for_each_entry(rule, &ops->rules_list, list) {
		if (frh->action && (frh->action != rule->action))
			continue;

		if (frh->table && (frh_get_table(frh, tb) != rule->table))
			continue;

		if (tb[FRA_PRIORITY] &&
		    (rule->pref != nla_get_u32(tb[FRA_PRIORITY])))
			continue;

		if (tb[FRA_IFNAME] &&
		    nla_strcmp(tb[FRA_IFNAME], rule->ifname))
			continue;

		if (tb[FRA_FWMARK] &&
		    (rule->mark != nla_get_u32(tb[FRA_FWMARK])))
			continue;

		if (tb[FRA_FWMASK] &&
		    (rule->mark_mask != nla_get_u32(tb[FRA_FWMASK])))
			continue;

		if (!ops->compare(rule, frh, tb))
			continue;

		if (rule->flags & FIB_RULE_PERMANENT) {
			err = -EPERM;
			goto errout;
		}

		list_del_rcu(&rule->list);

		if (rule->action == FR_ACT_GOTO)
			ops->nr_goto_rules--;

		/*
		 * Check if this rule is a target to any of them. If so,
		 * disable them. As this operation is eventually very
		 * expensive, it is only performed if goto rules have
		 * actually been added.
		 */
		if (ops->nr_goto_rules > 0) {
			list_for_each_entry(tmp, &ops->rules_list, list) {
				if (tmp->ctarget == rule) {
					rcu_assign_pointer(tmp->ctarget, NULL);
					ops->unresolved_rules++;
				}
			}
		}

		synchronize_rcu();
		notify_rule_change(RTM_DELRULE, rule, ops, nlh,
				   NETLINK_CB(skb).pid);
		fib_rule_put(rule);
		flush_route_cache(ops);
		rules_ops_put(ops);
		return 0;
	}

	err = -ENOENT;
errout:
	rules_ops_put(ops);
	return err;
}

static inline size_t fib_rule_nlmsg_size(struct fib_rules_ops *ops,
					 struct fib_rule *rule)
{
	size_t payload = NLMSG_ALIGN(sizeof(struct fib_rule_hdr))
			 + nla_total_size(IFNAMSIZ) /* FRA_IFNAME */
			 + nla_total_size(4) /* FRA_PRIORITY */
			 + nla_total_size(4) /* FRA_TABLE */
			 + nla_total_size(4) /* FRA_FWMARK */
			 + nla_total_size(4); /* FRA_FWMASK */

	if (ops->nlmsg_payload)
		payload += ops->nlmsg_payload(rule);

	return payload;
}

static int fib_nl_fill_rule(struct sk_buff *skb, struct fib_rule *rule,
			    u32 pid, u32 seq, int type, int flags,
			    struct fib_rules_ops *ops)
{
	struct nlmsghdr *nlh;
	struct fib_rule_hdr *frh;

	nlh = nlmsg_put(skb, pid, seq, type, sizeof(*frh), flags);
	if (nlh == NULL)
		return -EMSGSIZE;

	frh = nlmsg_data(nlh);
	frh->table = rule->table;
	NLA_PUT_U32(skb, FRA_TABLE, rule->table);
	frh->res1 = 0;
	frh->res2 = 0;
	frh->action = rule->action;
	frh->flags = rule->flags;

	if (rule->action == FR_ACT_GOTO && rule->ctarget == NULL)
		frh->flags |= FIB_RULE_UNRESOLVED;

	if (rule->ifname[0]) {
		NLA_PUT_STRING(skb, FRA_IFNAME, rule->ifname);

		if (rule->ifindex == -1)
			frh->flags |= FIB_RULE_DEV_DETACHED;
	}

	if (rule->pref)
		NLA_PUT_U32(skb, FRA_PRIORITY, rule->pref);

	if (rule->mark)
		NLA_PUT_U32(skb, FRA_FWMARK, rule->mark);

	if (rule->mark_mask || rule->mark)
		NLA_PUT_U32(skb, FRA_FWMASK, rule->mark_mask);

	if (rule->target)
		NLA_PUT_U32(skb, FRA_GOTO, rule->target);

	if (ops->fill(rule, skb, nlh, frh) < 0)
		goto nla_put_failure;

	return nlmsg_end(skb, nlh);

nla_put_failure:
	nlmsg_cancel(skb, nlh);
	return -EMSGSIZE;
}

static int dump_rules(struct sk_buff *skb, struct netlink_callback *cb,
		      struct fib_rules_ops *ops)
{
	int idx = 0;
	struct fib_rule *rule;

	list_for_each_entry(rule, &ops->rules_list, list) {
		if (idx < cb->args[1])
			goto skip;

		if (fib_nl_fill_rule(skb, rule, NETLINK_CB(cb->skb).pid,
				     cb->nlh->nlmsg_seq, RTM_NEWRULE,
				     NLM_F_MULTI, ops) < 0)
			break;
skip:
		idx++;
	}
	cb->args[1] = idx;
	rules_ops_put(ops);

	return skb->len;
}

static int fib_nl_dumprule(struct sk_buff *skb, struct netlink_callback *cb)
{
	struct net *net = skb->sk->sk_net;
	struct fib_rules_ops *ops;
	int idx = 0, family;

	if (net != &init_net)
		return -EINVAL;

	family = rtnl_msg_family(cb->nlh);
	if (family != AF_UNSPEC) {
		/* Protocol specific dump request */
		ops = lookup_rules_ops(family);
		if (ops == NULL)
			return -EAFNOSUPPORT;

		return dump_rules(skb, cb, ops);
	}

	rcu_read_lock();
	list_for_each_entry_rcu(ops, &rules_ops, list) {
		if (idx < cb->args[0] || !try_module_get(ops->owner))
			goto skip;

		if (dump_rules(skb, cb, ops) < 0)
			break;

		cb->args[1] = 0;
	skip:
		idx++;
	}
	rcu_read_unlock();
	cb->args[0] = idx;

	return skb->len;
}

static void notify_rule_change(int event, struct fib_rule *rule,
			       struct fib_rules_ops *ops, struct nlmsghdr *nlh,
			       u32 pid)
{
	struct sk_buff *skb;
	int err = -ENOBUFS;

	skb = nlmsg_new(fib_rule_nlmsg_size(ops, rule), GFP_KERNEL);
	if (skb == NULL)
		goto errout;

	err = fib_nl_fill_rule(skb, rule, pid, nlh->nlmsg_seq, event, 0, ops);
	if (err < 0) {
		/* -EMSGSIZE implies BUG in fib_rule_nlmsg_size() */
		WARN_ON(err == -EMSGSIZE);
		kfree_skb(skb);
		goto errout;
	}
	err = rtnl_notify(skb, &init_net, pid, ops->nlgroup, nlh, GFP_KERNEL);
errout:
	if (err < 0)
		rtnl_set_sk_err(&init_net, ops->nlgroup, err);
}

static void attach_rules(struct list_head *rules, struct net_device *dev)
{
	struct fib_rule *rule;

	list_for_each_entry(rule, rules, list) {
		if (rule->ifindex == -1 &&
		    strcmp(dev->name, rule->ifname) == 0)
			rule->ifindex = dev->ifindex;
	}
}

static void detach_rules(struct list_head *rules, struct net_device *dev)
{
	struct fib_rule *rule;

	list_for_each_entry(rule, rules, list)
		if (rule->ifindex == dev->ifindex)
			rule->ifindex = -1;
}


static int fib_rules_event(struct notifier_block *this, unsigned long event,
			    void *ptr)
{
	struct net_device *dev = ptr;
	struct fib_rules_ops *ops;

	if (dev->nd_net != &init_net)
		return NOTIFY_DONE;

	ASSERT_RTNL();
	rcu_read_lock();

	switch (event) {
	case NETDEV_REGISTER:
		list_for_each_entry(ops, &rules_ops, list)
			attach_rules(&ops->rules_list, dev);
		break;

	case NETDEV_UNREGISTER:
		list_for_each_entry(ops, &rules_ops, list)
			detach_rules(&ops->rules_list, dev);
		break;
	}

	rcu_read_unlock();

	return NOTIFY_DONE;
}

static struct notifier_block fib_rules_notifier = {
	.notifier_call = fib_rules_event,
};

static int __init fib_rules_init(void)
{
	rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL);
	rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL);
	rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule);

	return register_netdevice_notifier(&fib_rules_notifier);
}

subsys_initcall(fib_rules_init);
Commit	Line	Data
14c0b97d TG	1	/*
	2	* net/core/fib_rules.c Generic Routing Rules
	3	*
	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of the GNU General Public License as
	6	* published by the Free Software Foundation, version 2.
	7	*
	8	* Authors: Thomas Graf <tgraf@suug.ch>
	9	*/
	10
14c0b97d TG	11	#include <linux/types.h>
	12	#include <linux/kernel.h>
	13	#include <linux/list.h>
e9dc8653	14	#include <net/net_namespace.h>
881d966b	15	#include <net/sock.h>
14c0b97d TG	16	#include <net/fib_rules.h>
	17
	18	static LIST_HEAD(rules_ops);
	19	static DEFINE_SPINLOCK(rules_mod_lock);
	20
2994c638 DL	21	int fib_default_rule_add(struct fib_rules_ops *ops,
	22	u32 pref, u32 table, u32 flags)
	23	{
	24	struct fib_rule *r;
	25
	26	r = kzalloc(ops->rule_size, GFP_KERNEL);
	27	if (r == NULL)
	28	return -ENOMEM;
	29
	30	atomic_set(&r->refcnt, 1);
	31	r->action = FR_ACT_TO_TBL;
	32	r->pref = pref;
	33	r->table = table;
	34	r->flags = flags;
	35
	36	/* The lock is not required here, the list in unreacheable
	37	* at the moment this function is called */
	38	list_add_tail(&r->list, &ops->rules_list);
	39	return 0;
	40	}
	41	EXPORT_SYMBOL(fib_default_rule_add);
	42
14c0b97d	43	static void notify_rule_change(int event, struct fib_rule *rule,
c17084d2 TG	44	struct fib_rules_ops ops, struct nlmsghdr nlh,
c17084d2 TG	45	u32 pid);
14c0b97d TG	46
	47	static struct fib_rules_ops *lookup_rules_ops(int family)
	48	{
	49	struct fib_rules_ops *ops;
	50
	51	rcu_read_lock();
	52	list_for_each_entry_rcu(ops, &rules_ops, list) {
	53	if (ops->family == family) {
	54	if (!try_module_get(ops->owner))
	55	ops = NULL;
	56	rcu_read_unlock();
	57	return ops;
	58	}
	59	}
	60	rcu_read_unlock();
	61
	62	return NULL;
	63	}
	64
	65	static void rules_ops_put(struct fib_rules_ops *ops)
	66	{
	67	if (ops)
	68	module_put(ops->owner);
	69	}
	70
73417f61 TG	71	static void flush_route_cache(struct fib_rules_ops *ops)
	72	{
	73	if (ops->flush_cache)
	74	ops->flush_cache();
	75	}
	76
f8c26b8d	77	int fib_rules_register(struct net net, struct fib_rules_ops ops)
14c0b97d TG	78	{
	79	int err = -EEXIST;
	80	struct fib_rules_ops *o;
	81
	82	if (ops->rule_size < sizeof(struct fib_rule))
	83	return -EINVAL;
	84
	85	if (ops->match == NULL \|\| ops->configure == NULL \|\|
	86	ops->compare == NULL \|\| ops->fill == NULL \|\|
	87	ops->action == NULL)
	88	return -EINVAL;
	89
	90	spin_lock(&rules_mod_lock);
	91	list_for_each_entry(o, &rules_ops, list)
	92	if (ops->family == o->family)
	93	goto errout;
	94
	95	list_add_tail_rcu(&ops->list, &rules_ops);
	96	err = 0;
	97	errout:
	98	spin_unlock(&rules_mod_lock);
	99
	100	return err;
	101	}
	102
	103	EXPORT_SYMBOL_GPL(fib_rules_register);
	104
9eb87f3f	105	void fib_rules_cleanup_ops(struct fib_rules_ops *ops)
14c0b97d TG	106	{
	107	struct fib_rule rule, tmp;
	108
76c72d4f	109	list_for_each_entry_safe(rule, tmp, &ops->rules_list, list) {
14c0b97d TG	110	list_del_rcu(&rule->list);
	111	fib_rule_put(rule);
	112	}
	113	}
9eb87f3f	114	EXPORT_SYMBOL_GPL(fib_rules_cleanup_ops);
14c0b97d	115
f8c26b8d	116	int fib_rules_unregister(struct net net, struct fib_rules_ops ops)
14c0b97d TG	117	{
	118	int err = 0;
	119	struct fib_rules_ops *o;
	120
	121	spin_lock(&rules_mod_lock);
	122	list_for_each_entry(o, &rules_ops, list) {
	123	if (o == ops) {
	124	list_del_rcu(&o->list);
9eb87f3f	125	fib_rules_cleanup_ops(ops);
14c0b97d TG	126	goto out;
	127	}
	128	}
	129
	130	err = -ENOENT;
	131	out:
	132	spin_unlock(&rules_mod_lock);
	133
	134	synchronize_rcu();
	135
	136	return err;
	137	}
	138
	139	EXPORT_SYMBOL_GPL(fib_rules_unregister);
	140
3dfbcc41 TG	141	static int fib_rule_match(struct fib_rule rule, struct fib_rules_ops ops,
	142	struct flowi *fl, int flags)
	143	{
	144	int ret = 0;
	145
	146	if (rule->ifindex && (rule->ifindex != fl->iif))
	147	goto out;
	148
	149	if ((rule->mark ^ fl->mark) & rule->mark_mask)
	150	goto out;
	151
	152	ret = ops->match(rule, fl, flags);
	153	out:
	154	return (rule->flags & FIB_RULE_INVERT) ? !ret : ret;
	155	}
	156
14c0b97d TG	157	int fib_rules_lookup(struct fib_rules_ops ops, struct flowi fl,
	158	int flags, struct fib_lookup_arg *arg)
	159	{
	160	struct fib_rule *rule;
	161	int err;
	162
	163	rcu_read_lock();
	164
76c72d4f	165	list_for_each_entry_rcu(rule, &ops->rules_list, list) {
0947c9fe	166	jumped:
3dfbcc41	167	if (!fib_rule_match(rule, ops, fl, flags))
14c0b97d TG	168	continue;
14c0b97d TG	169
0947c9fe TG	170	if (rule->action == FR_ACT_GOTO) {
	171	struct fib_rule *target;
	172
	173	target = rcu_dereference(rule->ctarget);
	174	if (target == NULL) {
	175	continue;
	176	} else {
	177	rule = target;
	178	goto jumped;
	179	}
fa0b2d1d TG	180	} else if (rule->action == FR_ACT_NOP)
	181	continue;
	182	else
0947c9fe TG	183	err = ops->action(rule, fl, flags, arg);
0947c9fe TG	184
14c0b97d TG	185	if (err != -EAGAIN) {
	186	fib_rule_get(rule);
	187	arg->rule = rule;
	188	goto out;
	189	}
	190	}
	191
83886b6b	192	err = -ESRCH;
14c0b97d TG	193	out:
	194	rcu_read_unlock();
	195
	196	return err;
	197	}
	198
	199	EXPORT_SYMBOL_GPL(fib_rules_lookup);
	200
e1701c68 TG	201	static int validate_rulemsg(struct fib_rule_hdr frh, struct nlattr *tb,
	202	struct fib_rules_ops *ops)
	203	{
	204	int err = -EINVAL;
	205
	206	if (frh->src_len)
	207	if (tb[FRA_SRC] == NULL \|\|
	208	frh->src_len > (ops->addr_size * 8) \|\|
	209	nla_len(tb[FRA_SRC]) != ops->addr_size)
	210	goto errout;
	211
	212	if (frh->dst_len)
	213	if (tb[FRA_DST] == NULL \|\|
	214	frh->dst_len > (ops->addr_size * 8) \|\|
	215	nla_len(tb[FRA_DST]) != ops->addr_size)
	216	goto errout;
	217
	218	err = 0;
	219	errout:
	220	return err;
	221	}
	222
9d9e6a58	223	static int fib_nl_newrule(struct sk_buff skb, struct nlmsghdr nlh, void *arg)
14c0b97d	224	{
881d966b	225	struct net *net = skb->sk->sk_net;
14c0b97d TG	226	struct fib_rule_hdr *frh = nlmsg_data(nlh);
	227	struct fib_rules_ops *ops = NULL;
	228	struct fib_rule rule, r, *last = NULL;
	229	struct nlattr *tb[FRA_MAX+1];
0947c9fe	230	int err = -EINVAL, unresolved = 0;
14c0b97d	231
b854272b DL	232	if (net != &init_net)
	233	return -EINVAL;
	234
14c0b97d TG	235	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
	236	goto errout;
	237
	238	ops = lookup_rules_ops(frh->family);
	239	if (ops == NULL) {
	240	err = EAFNOSUPPORT;
	241	goto errout;
	242	}
	243
	244	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	245	if (err < 0)
	246	goto errout;
	247
e1701c68 TG	248	err = validate_rulemsg(frh, tb, ops);
	249	if (err < 0)
	250	goto errout;
	251
14c0b97d TG	252	rule = kzalloc(ops->rule_size, GFP_KERNEL);
	253	if (rule == NULL) {
	254	err = -ENOMEM;
	255	goto errout;
	256	}
	257
	258	if (tb[FRA_PRIORITY])
	259	rule->pref = nla_get_u32(tb[FRA_PRIORITY]);
	260
	261	if (tb[FRA_IFNAME]) {
	262	struct net_device *dev;
	263
	264	rule->ifindex = -1;
5176f91e	265	nla_strlcpy(rule->ifname, tb[FRA_IFNAME], IFNAMSIZ);
881d966b	266	dev = __dev_get_by_name(net, rule->ifname);
14c0b97d TG	267	if (dev)
	268	rule->ifindex = dev->ifindex;
	269	}
	270
b8964ed9 TG	271	if (tb[FRA_FWMARK]) {
	272	rule->mark = nla_get_u32(tb[FRA_FWMARK]);
	273	if (rule->mark)
	274	/* compatibility: if the mark value is non-zero all bits
	275	* are compared unless a mask is explicitly specified.
	276	*/
	277	rule->mark_mask = 0xFFFFFFFF;
	278	}
	279
	280	if (tb[FRA_FWMASK])
	281	rule->mark_mask = nla_get_u32(tb[FRA_FWMASK]);
	282
14c0b97d TG	283	rule->action = frh->action;
14c0b97d TG	284	rule->flags = frh->flags;
9e762a4a	285	rule->table = frh_get_table(frh, tb);
14c0b97d TG	286
14c0b97d TG	287	if (!rule->pref && ops->default_pref)
868d13ac	288	rule->pref = ops->default_pref(ops);
14c0b97d	289
0947c9fe TG	290	err = -EINVAL;
	291	if (tb[FRA_GOTO]) {
	292	if (rule->action != FR_ACT_GOTO)
	293	goto errout_free;
	294
	295	rule->target = nla_get_u32(tb[FRA_GOTO]);
	296	/* Backward jumps are prohibited to avoid endless loops */
	297	if (rule->target <= rule->pref)
	298	goto errout_free;
	299
76c72d4f	300	list_for_each_entry(r, &ops->rules_list, list) {
0947c9fe TG	301	if (r->pref == rule->target) {
	302	rule->ctarget = r;
	303	break;
	304	}
	305	}
	306
	307	if (rule->ctarget == NULL)
	308	unresolved = 1;
	309	} else if (rule->action == FR_ACT_GOTO)
	310	goto errout_free;
	311
14c0b97d TG	312	err = ops->configure(rule, skb, nlh, frh, tb);
	313	if (err < 0)
	314	goto errout_free;
	315
76c72d4f	316	list_for_each_entry(r, &ops->rules_list, list) {
14c0b97d TG	317	if (r->pref > rule->pref)
	318	break;
	319	last = r;
	320	}
	321
	322	fib_rule_get(rule);
	323
0947c9fe TG	324	if (ops->unresolved_rules) {
	325	/*
	326	* There are unresolved goto rules in the list, check if
	327	* any of them are pointing to this new rule.
	328	*/
76c72d4f	329	list_for_each_entry(r, &ops->rules_list, list) {
0947c9fe TG	330	if (r->action == FR_ACT_GOTO &&
	331	r->target == rule->pref) {
	332	BUG_ON(r->ctarget != NULL);
	333	rcu_assign_pointer(r->ctarget, rule);
	334	if (--ops->unresolved_rules == 0)
	335	break;
	336	}
	337	}
	338	}
	339
	340	if (rule->action == FR_ACT_GOTO)
	341	ops->nr_goto_rules++;
	342
	343	if (unresolved)
	344	ops->unresolved_rules++;
	345
14c0b97d TG	346	if (last)
	347	list_add_rcu(&rule->list, &last->list);
	348	else
76c72d4f	349	list_add_rcu(&rule->list, &ops->rules_list);
14c0b97d	350
c17084d2	351	notify_rule_change(RTM_NEWRULE, rule, ops, nlh, NETLINK_CB(skb).pid);
73417f61	352	flush_route_cache(ops);
14c0b97d TG	353	rules_ops_put(ops);
	354	return 0;
	355
	356	errout_free:
	357	kfree(rule);
	358	errout:
	359	rules_ops_put(ops);
	360	return err;
	361	}
	362
9d9e6a58	363	static int fib_nl_delrule(struct sk_buff skb, struct nlmsghdr nlh, void *arg)
14c0b97d	364	{
b854272b	365	struct net *net = skb->sk->sk_net;
14c0b97d TG	366	struct fib_rule_hdr *frh = nlmsg_data(nlh);
14c0b97d TG	367	struct fib_rules_ops *ops = NULL;
0947c9fe	368	struct fib_rule rule, tmp;
14c0b97d TG	369	struct nlattr *tb[FRA_MAX+1];
	370	int err = -EINVAL;
	371
b854272b DL	372	if (net != &init_net)
	373	return -EINVAL;
	374
14c0b97d TG	375	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
	376	goto errout;
	377
	378	ops = lookup_rules_ops(frh->family);
	379	if (ops == NULL) {
	380	err = EAFNOSUPPORT;
	381	goto errout;
	382	}
	383
	384	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	385	if (err < 0)
	386	goto errout;
	387
e1701c68 TG	388	err = validate_rulemsg(frh, tb, ops);
	389	if (err < 0)
	390	goto errout;
	391
76c72d4f	392	list_for_each_entry(rule, &ops->rules_list, list) {
14c0b97d TG	393	if (frh->action && (frh->action != rule->action))
	394	continue;
	395
9e762a4a	396	if (frh->table && (frh_get_table(frh, tb) != rule->table))
14c0b97d TG	397	continue;
	398
	399	if (tb[FRA_PRIORITY] &&
	400	(rule->pref != nla_get_u32(tb[FRA_PRIORITY])))
	401	continue;
	402
	403	if (tb[FRA_IFNAME] &&
	404	nla_strcmp(tb[FRA_IFNAME], rule->ifname))
	405	continue;
	406
b8964ed9 TG	407	if (tb[FRA_FWMARK] &&
	408	(rule->mark != nla_get_u32(tb[FRA_FWMARK])))
	409	continue;
	410
	411	if (tb[FRA_FWMASK] &&
	412	(rule->mark_mask != nla_get_u32(tb[FRA_FWMASK])))
	413	continue;
	414
14c0b97d TG	415	if (!ops->compare(rule, frh, tb))
	416	continue;
	417
	418	if (rule->flags & FIB_RULE_PERMANENT) {
	419	err = -EPERM;
	420	goto errout;
	421	}
	422
	423	list_del_rcu(&rule->list);
0947c9fe TG	424
	425	if (rule->action == FR_ACT_GOTO)
	426	ops->nr_goto_rules--;
	427
	428	/*
	429	* Check if this rule is a target to any of them. If so,
	430	* disable them. As this operation is eventually very
	431	* expensive, it is only performed if goto rules have
	432	* actually been added.
	433	*/
	434	if (ops->nr_goto_rules > 0) {
76c72d4f	435	list_for_each_entry(tmp, &ops->rules_list, list) {
0947c9fe TG	436	if (tmp->ctarget == rule) {
	437	rcu_assign_pointer(tmp->ctarget, NULL);
	438	ops->unresolved_rules++;
	439	}
	440	}
	441	}
	442
14c0b97d	443	synchronize_rcu();
c17084d2 TG	444	notify_rule_change(RTM_DELRULE, rule, ops, nlh,
c17084d2 TG	445	NETLINK_CB(skb).pid);
14c0b97d	446	fib_rule_put(rule);
73417f61	447	flush_route_cache(ops);
14c0b97d TG	448	rules_ops_put(ops);
	449	return 0;
	450	}
	451
	452	err = -ENOENT;
	453	errout:
	454	rules_ops_put(ops);
	455	return err;
	456	}
	457
339bf98f TG	458	static inline size_t fib_rule_nlmsg_size(struct fib_rules_ops *ops,
	459	struct fib_rule *rule)
	460	{
	461	size_t payload = NLMSG_ALIGN(sizeof(struct fib_rule_hdr))
	462	+ nla_total_size(IFNAMSIZ) /* FRA_IFNAME */
	463	+ nla_total_size(4) /* FRA_PRIORITY */
	464	+ nla_total_size(4) /* FRA_TABLE */
	465	+ nla_total_size(4) /* FRA_FWMARK */
	466	+ nla_total_size(4); /* FRA_FWMASK */
	467
	468	if (ops->nlmsg_payload)
	469	payload += ops->nlmsg_payload(rule);
	470
	471	return payload;
	472	}
	473
14c0b97d TG	474	static int fib_nl_fill_rule(struct sk_buff skb, struct fib_rule rule,
	475	u32 pid, u32 seq, int type, int flags,
	476	struct fib_rules_ops *ops)
	477	{
	478	struct nlmsghdr *nlh;
	479	struct fib_rule_hdr *frh;
	480
	481	nlh = nlmsg_put(skb, pid, seq, type, sizeof(*frh), flags);
	482	if (nlh == NULL)
26932566	483	return -EMSGSIZE;
14c0b97d TG	484
	485	frh = nlmsg_data(nlh);
	486	frh->table = rule->table;
9e762a4a	487	NLA_PUT_U32(skb, FRA_TABLE, rule->table);
14c0b97d TG	488	frh->res1 = 0;
	489	frh->res2 = 0;
	490	frh->action = rule->action;
	491	frh->flags = rule->flags;
	492
0947c9fe TG	493	if (rule->action == FR_ACT_GOTO && rule->ctarget == NULL)
	494	frh->flags \|= FIB_RULE_UNRESOLVED;
	495
2b443683	496	if (rule->ifname[0]) {
14c0b97d TG	497	NLA_PUT_STRING(skb, FRA_IFNAME, rule->ifname);
14c0b97d TG	498
2b443683 TG	499	if (rule->ifindex == -1)
	500	frh->flags \|= FIB_RULE_DEV_DETACHED;
	501	}
	502
14c0b97d TG	503	if (rule->pref)
	504	NLA_PUT_U32(skb, FRA_PRIORITY, rule->pref);
	505
b8964ed9 TG	506	if (rule->mark)
	507	NLA_PUT_U32(skb, FRA_FWMARK, rule->mark);
	508
	509	if (rule->mark_mask \|\| rule->mark)
	510	NLA_PUT_U32(skb, FRA_FWMASK, rule->mark_mask);
	511
0947c9fe TG	512	if (rule->target)
	513	NLA_PUT_U32(skb, FRA_GOTO, rule->target);
	514
14c0b97d TG	515	if (ops->fill(rule, skb, nlh, frh) < 0)
	516	goto nla_put_failure;
	517
	518	return nlmsg_end(skb, nlh);
	519
	520	nla_put_failure:
26932566 PM	521	nlmsg_cancel(skb, nlh);
26932566 PM	522	return -EMSGSIZE;
14c0b97d TG	523	}
14c0b97d TG	524
c454673d TG	525	static int dump_rules(struct sk_buff skb, struct netlink_callback cb,
c454673d TG	526	struct fib_rules_ops *ops)
14c0b97d TG	527	{
	528	int idx = 0;
	529	struct fib_rule *rule;
14c0b97d	530
76c72d4f	531	list_for_each_entry(rule, &ops->rules_list, list) {
c454673d	532	if (idx < cb->args[1])
14c0b97d TG	533	goto skip;
	534
	535	if (fib_nl_fill_rule(skb, rule, NETLINK_CB(cb->skb).pid,
	536	cb->nlh->nlmsg_seq, RTM_NEWRULE,
	537	NLM_F_MULTI, ops) < 0)
	538	break;
	539	skip:
	540	idx++;
	541	}
c454673d	542	cb->args[1] = idx;
14c0b97d TG	543	rules_ops_put(ops);
	544
	545	return skb->len;
	546	}
	547
c454673d TG	548	static int fib_nl_dumprule(struct sk_buff skb, struct netlink_callback cb)
c454673d TG	549	{
b854272b	550	struct net *net = skb->sk->sk_net;
c454673d TG	551	struct fib_rules_ops *ops;
	552	int idx = 0, family;
	553
b854272b DL	554	if (net != &init_net)
	555	return -EINVAL;
	556
c454673d TG	557	family = rtnl_msg_family(cb->nlh);
	558	if (family != AF_UNSPEC) {
	559	/* Protocol specific dump request */
	560	ops = lookup_rules_ops(family);
	561	if (ops == NULL)
	562	return -EAFNOSUPPORT;
	563
	564	return dump_rules(skb, cb, ops);
	565	}
	566
	567	rcu_read_lock();
	568	list_for_each_entry_rcu(ops, &rules_ops, list) {
	569	if (idx < cb->args[0] \|\| !try_module_get(ops->owner))
	570	goto skip;
	571
	572	if (dump_rules(skb, cb, ops) < 0)
	573	break;
	574
	575	cb->args[1] = 0;
	576	skip:
	577	idx++;
	578	}
	579	rcu_read_unlock();
	580	cb->args[0] = idx;
	581
	582	return skb->len;
	583	}
14c0b97d TG	584
14c0b97d TG	585	static void notify_rule_change(int event, struct fib_rule *rule,
c17084d2 TG	586	struct fib_rules_ops ops, struct nlmsghdr nlh,
c17084d2 TG	587	u32 pid)
14c0b97d	588	{
c17084d2 TG	589	struct sk_buff *skb;
c17084d2 TG	590	int err = -ENOBUFS;
14c0b97d	591
339bf98f	592	skb = nlmsg_new(fib_rule_nlmsg_size(ops, rule), GFP_KERNEL);
14c0b97d	593	if (skb == NULL)
c17084d2 TG	594	goto errout;
	595
	596	err = fib_nl_fill_rule(skb, rule, pid, nlh->nlmsg_seq, event, 0, ops);
26932566 PM	597	if (err < 0) {
	598	/* -EMSGSIZE implies BUG in fib_rule_nlmsg_size() */
	599	WARN_ON(err == -EMSGSIZE);
	600	kfree_skb(skb);
	601	goto errout;
	602	}
97c53cac	603	err = rtnl_notify(skb, &init_net, pid, ops->nlgroup, nlh, GFP_KERNEL);
c17084d2 TG	604	errout:
c17084d2 TG	605	if (err < 0)
97c53cac	606	rtnl_set_sk_err(&init_net, ops->nlgroup, err);
14c0b97d TG	607	}
	608
	609	static void attach_rules(struct list_head rules, struct net_device dev)
	610	{
	611	struct fib_rule *rule;
	612
	613	list_for_each_entry(rule, rules, list) {
	614	if (rule->ifindex == -1 &&
	615	strcmp(dev->name, rule->ifname) == 0)
	616	rule->ifindex = dev->ifindex;
	617	}
	618	}
	619
	620	static void detach_rules(struct list_head rules, struct net_device dev)
	621	{
	622	struct fib_rule *rule;
	623
	624	list_for_each_entry(rule, rules, list)
	625	if (rule->ifindex == dev->ifindex)
	626	rule->ifindex = -1;
	627	}
	628
	629
	630	static int fib_rules_event(struct notifier_block *this, unsigned long event,
	631	void *ptr)
	632	{
	633	struct net_device *dev = ptr;
	634	struct fib_rules_ops *ops;
	635
e9dc8653 EB	636	if (dev->nd_net != &init_net)
	637	return NOTIFY_DONE;
	638
14c0b97d TG	639	ASSERT_RTNL();
	640	rcu_read_lock();
	641
	642	switch (event) {
	643	case NETDEV_REGISTER:
	644	list_for_each_entry(ops, &rules_ops, list)
76c72d4f	645	attach_rules(&ops->rules_list, dev);
14c0b97d TG	646	break;
	647
	648	case NETDEV_UNREGISTER:
	649	list_for_each_entry(ops, &rules_ops, list)
76c72d4f	650	detach_rules(&ops->rules_list, dev);
14c0b97d TG	651	break;
	652	}
	653
	654	rcu_read_unlock();
	655
	656	return NOTIFY_DONE;
	657	}
	658
	659	static struct notifier_block fib_rules_notifier = {
	660	.notifier_call = fib_rules_event,
	661	};
	662
	663	static int __init fib_rules_init(void)
	664	{
9d9e6a58 TG	665	rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL);
9d9e6a58 TG	666	rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL);
c454673d	667	rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule);
9d9e6a58	668
14c0b97d TG	669	return register_netdevice_notifier(&fib_rules_notifier);
	670	}
	671
	672	subsys_initcall(fib_rules_init);