[mirror_ubuntu-bionic-kernel.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/fib_rules.h>

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

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 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);

static void 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);
	}
}

int fib_rules_unregister(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);
			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 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 (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(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();

	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 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 (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 fib_rules_ops *ops;
	int idx = 0, family;

	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, pid, ops->nlgroup, nlh, GFP_KERNEL);
errout:
	if (err < 0)
		rtnl_set_sk_err(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;

	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>
	14	#include <net/fib_rules.h>
	15
	16	static LIST_HEAD(rules_ops);
	17	static DEFINE_SPINLOCK(rules_mod_lock);
	18
	19	static void notify_rule_change(int event, struct fib_rule *rule,
c17084d2 TG	20	struct fib_rules_ops ops, struct nlmsghdr nlh,
c17084d2 TG	21	u32 pid);
14c0b97d TG	22
	23	static struct fib_rules_ops *lookup_rules_ops(int family)
	24	{
	25	struct fib_rules_ops *ops;
	26
	27	rcu_read_lock();
	28	list_for_each_entry_rcu(ops, &rules_ops, list) {
	29	if (ops->family == family) {
	30	if (!try_module_get(ops->owner))
	31	ops = NULL;
	32	rcu_read_unlock();
	33	return ops;
	34	}
	35	}
	36	rcu_read_unlock();
	37
	38	return NULL;
	39	}
	40
	41	static void rules_ops_put(struct fib_rules_ops *ops)
	42	{
	43	if (ops)
	44	module_put(ops->owner);
	45	}
	46
73417f61 TG	47	static void flush_route_cache(struct fib_rules_ops *ops)
	48	{
	49	if (ops->flush_cache)
	50	ops->flush_cache();
	51	}
	52
14c0b97d TG	53	int fib_rules_register(struct fib_rules_ops *ops)
	54	{
	55	int err = -EEXIST;
	56	struct fib_rules_ops *o;
	57
	58	if (ops->rule_size < sizeof(struct fib_rule))
	59	return -EINVAL;
	60
	61	if (ops->match == NULL \|\| ops->configure == NULL \|\|
	62	ops->compare == NULL \|\| ops->fill == NULL \|\|
	63	ops->action == NULL)
	64	return -EINVAL;
	65
	66	spin_lock(&rules_mod_lock);
	67	list_for_each_entry(o, &rules_ops, list)
	68	if (ops->family == o->family)
	69	goto errout;
	70
	71	list_add_tail_rcu(&ops->list, &rules_ops);
	72	err = 0;
	73	errout:
	74	spin_unlock(&rules_mod_lock);
	75
	76	return err;
	77	}
	78
	79	EXPORT_SYMBOL_GPL(fib_rules_register);
	80
	81	static void cleanup_ops(struct fib_rules_ops *ops)
	82	{
	83	struct fib_rule rule, tmp;
	84
	85	list_for_each_entry_safe(rule, tmp, ops->rules_list, list) {
	86	list_del_rcu(&rule->list);
	87	fib_rule_put(rule);
	88	}
	89	}
	90
	91	int fib_rules_unregister(struct fib_rules_ops *ops)
	92	{
	93	int err = 0;
	94	struct fib_rules_ops *o;
	95
	96	spin_lock(&rules_mod_lock);
	97	list_for_each_entry(o, &rules_ops, list) {
	98	if (o == ops) {
	99	list_del_rcu(&o->list);
	100	cleanup_ops(ops);
	101	goto out;
	102	}
	103	}
	104
	105	err = -ENOENT;
	106	out:
	107	spin_unlock(&rules_mod_lock);
	108
	109	synchronize_rcu();
	110
	111	return err;
	112	}
	113
	114	EXPORT_SYMBOL_GPL(fib_rules_unregister);
	115
3dfbcc41 TG	116	static int fib_rule_match(struct fib_rule rule, struct fib_rules_ops ops,
	117	struct flowi *fl, int flags)
	118	{
	119	int ret = 0;
	120
	121	if (rule->ifindex && (rule->ifindex != fl->iif))
	122	goto out;
	123
	124	if ((rule->mark ^ fl->mark) & rule->mark_mask)
	125	goto out;
	126
	127	ret = ops->match(rule, fl, flags);
	128	out:
	129	return (rule->flags & FIB_RULE_INVERT) ? !ret : ret;
	130	}
	131
14c0b97d TG	132	int fib_rules_lookup(struct fib_rules_ops ops, struct flowi fl,
	133	int flags, struct fib_lookup_arg *arg)
	134	{
	135	struct fib_rule *rule;
	136	int err;
	137
	138	rcu_read_lock();
	139
	140	list_for_each_entry_rcu(rule, ops->rules_list, list) {
0947c9fe	141	jumped:
3dfbcc41	142	if (!fib_rule_match(rule, ops, fl, flags))
14c0b97d TG	143	continue;
14c0b97d TG	144
0947c9fe TG	145	if (rule->action == FR_ACT_GOTO) {
	146	struct fib_rule *target;
	147
	148	target = rcu_dereference(rule->ctarget);
	149	if (target == NULL) {
	150	continue;
	151	} else {
	152	rule = target;
	153	goto jumped;
	154	}
fa0b2d1d TG	155	} else if (rule->action == FR_ACT_NOP)
	156	continue;
	157	else
0947c9fe TG	158	err = ops->action(rule, fl, flags, arg);
0947c9fe TG	159
14c0b97d TG	160	if (err != -EAGAIN) {
	161	fib_rule_get(rule);
	162	arg->rule = rule;
	163	goto out;
	164	}
	165	}
	166
83886b6b	167	err = -ESRCH;
14c0b97d TG	168	out:
	169	rcu_read_unlock();
	170
	171	return err;
	172	}
	173
	174	EXPORT_SYMBOL_GPL(fib_rules_lookup);
	175
e1701c68 TG	176	static int validate_rulemsg(struct fib_rule_hdr frh, struct nlattr *tb,
	177	struct fib_rules_ops *ops)
	178	{
	179	int err = -EINVAL;
	180
	181	if (frh->src_len)
	182	if (tb[FRA_SRC] == NULL \|\|
	183	frh->src_len > (ops->addr_size * 8) \|\|
	184	nla_len(tb[FRA_SRC]) != ops->addr_size)
	185	goto errout;
	186
	187	if (frh->dst_len)
	188	if (tb[FRA_DST] == NULL \|\|
	189	frh->dst_len > (ops->addr_size * 8) \|\|
	190	nla_len(tb[FRA_DST]) != ops->addr_size)
	191	goto errout;
	192
	193	err = 0;
	194	errout:
	195	return err;
	196	}
	197
9d9e6a58	198	static int fib_nl_newrule(struct sk_buff skb, struct nlmsghdr nlh, void *arg)
14c0b97d TG	199	{
	200	struct fib_rule_hdr *frh = nlmsg_data(nlh);
	201	struct fib_rules_ops *ops = NULL;
	202	struct fib_rule rule, r, *last = NULL;
	203	struct nlattr *tb[FRA_MAX+1];
0947c9fe	204	int err = -EINVAL, unresolved = 0;
14c0b97d TG	205
	206	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
	207	goto errout;
	208
	209	ops = lookup_rules_ops(frh->family);
	210	if (ops == NULL) {
	211	err = EAFNOSUPPORT;
	212	goto errout;
	213	}
	214
	215	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	216	if (err < 0)
	217	goto errout;
	218
e1701c68 TG	219	err = validate_rulemsg(frh, tb, ops);
	220	if (err < 0)
	221	goto errout;
	222
14c0b97d TG	223	rule = kzalloc(ops->rule_size, GFP_KERNEL);
	224	if (rule == NULL) {
	225	err = -ENOMEM;
	226	goto errout;
	227	}
	228
	229	if (tb[FRA_PRIORITY])
	230	rule->pref = nla_get_u32(tb[FRA_PRIORITY]);
	231
	232	if (tb[FRA_IFNAME]) {
	233	struct net_device *dev;
	234
	235	rule->ifindex = -1;
5176f91e	236	nla_strlcpy(rule->ifname, tb[FRA_IFNAME], IFNAMSIZ);
14c0b97d TG	237	dev = __dev_get_by_name(rule->ifname);
	238	if (dev)
	239	rule->ifindex = dev->ifindex;
	240	}
	241
b8964ed9 TG	242	if (tb[FRA_FWMARK]) {
	243	rule->mark = nla_get_u32(tb[FRA_FWMARK]);
	244	if (rule->mark)
	245	/* compatibility: if the mark value is non-zero all bits
	246	* are compared unless a mask is explicitly specified.
	247	*/
	248	rule->mark_mask = 0xFFFFFFFF;
	249	}
	250
	251	if (tb[FRA_FWMASK])
	252	rule->mark_mask = nla_get_u32(tb[FRA_FWMASK]);
	253
14c0b97d TG	254	rule->action = frh->action;
14c0b97d TG	255	rule->flags = frh->flags;
9e762a4a	256	rule->table = frh_get_table(frh, tb);
14c0b97d TG	257
	258	if (!rule->pref && ops->default_pref)
	259	rule->pref = ops->default_pref();
	260
0947c9fe TG	261	err = -EINVAL;
	262	if (tb[FRA_GOTO]) {
	263	if (rule->action != FR_ACT_GOTO)
	264	goto errout_free;
	265
	266	rule->target = nla_get_u32(tb[FRA_GOTO]);
	267	/* Backward jumps are prohibited to avoid endless loops */
	268	if (rule->target <= rule->pref)
	269	goto errout_free;
	270
	271	list_for_each_entry(r, ops->rules_list, list) {
	272	if (r->pref == rule->target) {
	273	rule->ctarget = r;
	274	break;
	275	}
	276	}
	277
	278	if (rule->ctarget == NULL)
	279	unresolved = 1;
	280	} else if (rule->action == FR_ACT_GOTO)
	281	goto errout_free;
	282
14c0b97d TG	283	err = ops->configure(rule, skb, nlh, frh, tb);
	284	if (err < 0)
	285	goto errout_free;
	286
	287	list_for_each_entry(r, ops->rules_list, list) {
	288	if (r->pref > rule->pref)
	289	break;
	290	last = r;
	291	}
	292
	293	fib_rule_get(rule);
	294
0947c9fe TG	295	if (ops->unresolved_rules) {
	296	/*
	297	* There are unresolved goto rules in the list, check if
	298	* any of them are pointing to this new rule.
	299	*/
	300	list_for_each_entry(r, ops->rules_list, list) {
	301	if (r->action == FR_ACT_GOTO &&
	302	r->target == rule->pref) {
	303	BUG_ON(r->ctarget != NULL);
	304	rcu_assign_pointer(r->ctarget, rule);
	305	if (--ops->unresolved_rules == 0)
	306	break;
	307	}
	308	}
	309	}
	310
	311	if (rule->action == FR_ACT_GOTO)
	312	ops->nr_goto_rules++;
	313
	314	if (unresolved)
	315	ops->unresolved_rules++;
	316
14c0b97d TG	317	if (last)
	318	list_add_rcu(&rule->list, &last->list);
	319	else
	320	list_add_rcu(&rule->list, ops->rules_list);
	321
c17084d2	322	notify_rule_change(RTM_NEWRULE, rule, ops, nlh, NETLINK_CB(skb).pid);
73417f61	323	flush_route_cache(ops);
14c0b97d TG	324	rules_ops_put(ops);
	325	return 0;
	326
	327	errout_free:
	328	kfree(rule);
	329	errout:
	330	rules_ops_put(ops);
	331	return err;
	332	}
	333
9d9e6a58	334	static int fib_nl_delrule(struct sk_buff skb, struct nlmsghdr nlh, void *arg)
14c0b97d TG	335	{
	336	struct fib_rule_hdr *frh = nlmsg_data(nlh);
	337	struct fib_rules_ops *ops = NULL;
0947c9fe	338	struct fib_rule rule, tmp;
14c0b97d TG	339	struct nlattr *tb[FRA_MAX+1];
	340	int err = -EINVAL;
	341
	342	if (nlh->nlmsg_len < nlmsg_msg_size(sizeof(*frh)))
	343	goto errout;
	344
	345	ops = lookup_rules_ops(frh->family);
	346	if (ops == NULL) {
	347	err = EAFNOSUPPORT;
	348	goto errout;
	349	}
	350
	351	err = nlmsg_parse(nlh, sizeof(*frh), tb, FRA_MAX, ops->policy);
	352	if (err < 0)
	353	goto errout;
	354
e1701c68 TG	355	err = validate_rulemsg(frh, tb, ops);
	356	if (err < 0)
	357	goto errout;
	358
14c0b97d TG	359	list_for_each_entry(rule, ops->rules_list, list) {
	360	if (frh->action && (frh->action != rule->action))
	361	continue;
	362
9e762a4a	363	if (frh->table && (frh_get_table(frh, tb) != rule->table))
14c0b97d TG	364	continue;
	365
	366	if (tb[FRA_PRIORITY] &&
	367	(rule->pref != nla_get_u32(tb[FRA_PRIORITY])))
	368	continue;
	369
	370	if (tb[FRA_IFNAME] &&
	371	nla_strcmp(tb[FRA_IFNAME], rule->ifname))
	372	continue;
	373
b8964ed9 TG	374	if (tb[FRA_FWMARK] &&
	375	(rule->mark != nla_get_u32(tb[FRA_FWMARK])))
	376	continue;
	377
	378	if (tb[FRA_FWMASK] &&
	379	(rule->mark_mask != nla_get_u32(tb[FRA_FWMASK])))
	380	continue;
	381
14c0b97d TG	382	if (!ops->compare(rule, frh, tb))
	383	continue;
	384
	385	if (rule->flags & FIB_RULE_PERMANENT) {
	386	err = -EPERM;
	387	goto errout;
	388	}
	389
	390	list_del_rcu(&rule->list);
0947c9fe TG	391
	392	if (rule->action == FR_ACT_GOTO)
	393	ops->nr_goto_rules--;
	394
	395	/*
	396	* Check if this rule is a target to any of them. If so,
	397	* disable them. As this operation is eventually very
	398	* expensive, it is only performed if goto rules have
	399	* actually been added.
	400	*/
	401	if (ops->nr_goto_rules > 0) {
	402	list_for_each_entry(tmp, ops->rules_list, list) {
	403	if (tmp->ctarget == rule) {
	404	rcu_assign_pointer(tmp->ctarget, NULL);
	405	ops->unresolved_rules++;
	406	}
	407	}
	408	}
	409
14c0b97d	410	synchronize_rcu();
c17084d2 TG	411	notify_rule_change(RTM_DELRULE, rule, ops, nlh,
c17084d2 TG	412	NETLINK_CB(skb).pid);
14c0b97d	413	fib_rule_put(rule);
73417f61	414	flush_route_cache(ops);
14c0b97d TG	415	rules_ops_put(ops);
	416	return 0;
	417	}
	418
	419	err = -ENOENT;
	420	errout:
	421	rules_ops_put(ops);
	422	return err;
	423	}
	424
339bf98f TG	425	static inline size_t fib_rule_nlmsg_size(struct fib_rules_ops *ops,
	426	struct fib_rule *rule)
	427	{
	428	size_t payload = NLMSG_ALIGN(sizeof(struct fib_rule_hdr))
	429	+ nla_total_size(IFNAMSIZ) /* FRA_IFNAME */
	430	+ nla_total_size(4) /* FRA_PRIORITY */
	431	+ nla_total_size(4) /* FRA_TABLE */
	432	+ nla_total_size(4) /* FRA_FWMARK */
	433	+ nla_total_size(4); /* FRA_FWMASK */
	434
	435	if (ops->nlmsg_payload)
	436	payload += ops->nlmsg_payload(rule);
	437
	438	return payload;
	439	}
	440
14c0b97d TG	441	static int fib_nl_fill_rule(struct sk_buff skb, struct fib_rule rule,
	442	u32 pid, u32 seq, int type, int flags,
	443	struct fib_rules_ops *ops)
	444	{
	445	struct nlmsghdr *nlh;
	446	struct fib_rule_hdr *frh;
	447
	448	nlh = nlmsg_put(skb, pid, seq, type, sizeof(*frh), flags);
	449	if (nlh == NULL)
26932566	450	return -EMSGSIZE;
14c0b97d TG	451
	452	frh = nlmsg_data(nlh);
	453	frh->table = rule->table;
9e762a4a	454	NLA_PUT_U32(skb, FRA_TABLE, rule->table);
14c0b97d TG	455	frh->res1 = 0;
	456	frh->res2 = 0;
	457	frh->action = rule->action;
	458	frh->flags = rule->flags;
	459
0947c9fe TG	460	if (rule->action == FR_ACT_GOTO && rule->ctarget == NULL)
	461	frh->flags \|= FIB_RULE_UNRESOLVED;
	462
2b443683	463	if (rule->ifname[0]) {
14c0b97d TG	464	NLA_PUT_STRING(skb, FRA_IFNAME, rule->ifname);
14c0b97d TG	465
2b443683 TG	466	if (rule->ifindex == -1)
	467	frh->flags \|= FIB_RULE_DEV_DETACHED;
	468	}
	469
14c0b97d TG	470	if (rule->pref)
	471	NLA_PUT_U32(skb, FRA_PRIORITY, rule->pref);
	472
b8964ed9 TG	473	if (rule->mark)
	474	NLA_PUT_U32(skb, FRA_FWMARK, rule->mark);
	475
	476	if (rule->mark_mask \|\| rule->mark)
	477	NLA_PUT_U32(skb, FRA_FWMASK, rule->mark_mask);
	478
0947c9fe TG	479	if (rule->target)
	480	NLA_PUT_U32(skb, FRA_GOTO, rule->target);
	481
14c0b97d TG	482	if (ops->fill(rule, skb, nlh, frh) < 0)
	483	goto nla_put_failure;
	484
	485	return nlmsg_end(skb, nlh);
	486
	487	nla_put_failure:
26932566 PM	488	nlmsg_cancel(skb, nlh);
26932566 PM	489	return -EMSGSIZE;
14c0b97d TG	490	}
14c0b97d TG	491
c454673d TG	492	static int dump_rules(struct sk_buff skb, struct netlink_callback cb,
c454673d TG	493	struct fib_rules_ops *ops)
14c0b97d TG	494	{
	495	int idx = 0;
	496	struct fib_rule *rule;
14c0b97d	497
6313c1e0	498	list_for_each_entry(rule, ops->rules_list, list) {
c454673d	499	if (idx < cb->args[1])
14c0b97d TG	500	goto skip;
	501
	502	if (fib_nl_fill_rule(skb, rule, NETLINK_CB(cb->skb).pid,
	503	cb->nlh->nlmsg_seq, RTM_NEWRULE,
	504	NLM_F_MULTI, ops) < 0)
	505	break;
	506	skip:
	507	idx++;
	508	}
c454673d	509	cb->args[1] = idx;
14c0b97d TG	510	rules_ops_put(ops);
	511
	512	return skb->len;
	513	}
	514
c454673d TG	515	static int fib_nl_dumprule(struct sk_buff skb, struct netlink_callback cb)
	516	{
	517	struct fib_rules_ops *ops;
	518	int idx = 0, family;
	519
	520	family = rtnl_msg_family(cb->nlh);
	521	if (family != AF_UNSPEC) {
	522	/* Protocol specific dump request */
	523	ops = lookup_rules_ops(family);
	524	if (ops == NULL)
	525	return -EAFNOSUPPORT;
	526
	527	return dump_rules(skb, cb, ops);
	528	}
	529
	530	rcu_read_lock();
	531	list_for_each_entry_rcu(ops, &rules_ops, list) {
	532	if (idx < cb->args[0] \|\| !try_module_get(ops->owner))
	533	goto skip;
	534
	535	if (dump_rules(skb, cb, ops) < 0)
	536	break;
	537
	538	cb->args[1] = 0;
	539	skip:
	540	idx++;
	541	}
	542	rcu_read_unlock();
	543	cb->args[0] = idx;
	544
	545	return skb->len;
	546	}
14c0b97d TG	547
14c0b97d TG	548	static void notify_rule_change(int event, struct fib_rule *rule,
c17084d2 TG	549	struct fib_rules_ops ops, struct nlmsghdr nlh,
c17084d2 TG	550	u32 pid)
14c0b97d	551	{
c17084d2 TG	552	struct sk_buff *skb;
c17084d2 TG	553	int err = -ENOBUFS;
14c0b97d	554
339bf98f	555	skb = nlmsg_new(fib_rule_nlmsg_size(ops, rule), GFP_KERNEL);
14c0b97d	556	if (skb == NULL)
c17084d2 TG	557	goto errout;
	558
	559	err = fib_nl_fill_rule(skb, rule, pid, nlh->nlmsg_seq, event, 0, ops);
26932566 PM	560	if (err < 0) {
	561	/* -EMSGSIZE implies BUG in fib_rule_nlmsg_size() */
	562	WARN_ON(err == -EMSGSIZE);
	563	kfree_skb(skb);
	564	goto errout;
	565	}
c17084d2 TG	566	err = rtnl_notify(skb, pid, ops->nlgroup, nlh, GFP_KERNEL);
	567	errout:
	568	if (err < 0)
	569	rtnl_set_sk_err(ops->nlgroup, err);
14c0b97d TG	570	}
	571
	572	static void attach_rules(struct list_head rules, struct net_device dev)
	573	{
	574	struct fib_rule *rule;
	575
	576	list_for_each_entry(rule, rules, list) {
	577	if (rule->ifindex == -1 &&
	578	strcmp(dev->name, rule->ifname) == 0)
	579	rule->ifindex = dev->ifindex;
	580	}
	581	}
	582
	583	static void detach_rules(struct list_head rules, struct net_device dev)
	584	{
	585	struct fib_rule *rule;
	586
	587	list_for_each_entry(rule, rules, list)
	588	if (rule->ifindex == dev->ifindex)
	589	rule->ifindex = -1;
	590	}
	591
	592
	593	static int fib_rules_event(struct notifier_block *this, unsigned long event,
	594	void *ptr)
	595	{
	596	struct net_device *dev = ptr;
	597	struct fib_rules_ops *ops;
	598
	599	ASSERT_RTNL();
	600	rcu_read_lock();
	601
	602	switch (event) {
	603	case NETDEV_REGISTER:
	604	list_for_each_entry(ops, &rules_ops, list)
	605	attach_rules(ops->rules_list, dev);
	606	break;
	607
	608	case NETDEV_UNREGISTER:
	609	list_for_each_entry(ops, &rules_ops, list)
	610	detach_rules(ops->rules_list, dev);
	611	break;
	612	}
	613
	614	rcu_read_unlock();
	615
	616	return NOTIFY_DONE;
	617	}
	618
	619	static struct notifier_block fib_rules_notifier = {
	620	.notifier_call = fib_rules_event,
	621	};
	622
	623	static int __init fib_rules_init(void)
	624	{
9d9e6a58 TG	625	rtnl_register(PF_UNSPEC, RTM_NEWRULE, fib_nl_newrule, NULL);
9d9e6a58 TG	626	rtnl_register(PF_UNSPEC, RTM_DELRULE, fib_nl_delrule, NULL);
c454673d	627	rtnl_register(PF_UNSPEC, RTM_GETRULE, NULL, fib_nl_dumprule);
9d9e6a58	628
14c0b97d TG	629	return register_netdevice_notifier(&fib_rules_notifier);
	630	}
	631
	632	subsys_initcall(fib_rules_init);