[net-next-2.6.git] / net / sched / cls_u32.c

/*
 * net/sched/cls_u32.c	Ugly (or Universal) 32bit key Packet Classifier.
 *
 *		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; either version
 *		2 of the License, or (at your option) any later version.
 *
 * Authors:	Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
 *
 *	The filters are packed to hash tables of key nodes
 *	with a set of 32bit key/mask pairs at every node.
 *	Nodes reference next level hash tables etc.
 *
 *	This scheme is the best universal classifier I managed to
 *	invent; it is not super-fast, but it is not slow (provided you
 *	program it correctly), and general enough.  And its relative
 *	speed grows as the number of rules becomes larger.
 *
 *	It seems that it represents the best middle point between
 *	speed and manageability both by human and by machine.
 *
 *	It is especially useful for link sharing combined with QoS;
 *	pure RSVP doesn't need such a general approach and can use
 *	much simpler (and faster) schemes, sort of cls_rsvp.c.
 *
 *	JHS: We should remove the CONFIG_NET_CLS_IND from here
 *	eventually when the meta match extension is made available
 *
 *	nfmark match added by Catalin(ux aka Dino) BOIE <catab at umbrella.ro>
 */

#include <linux/module.h>
#include <linux/slab.h>
#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/rtnetlink.h>
#include <linux/skbuff.h>
#include <net/netlink.h>
#include <net/act_api.h>
#include <net/pkt_cls.h>

struct tc_u_knode
{
	struct tc_u_knode	*next;
	u32			handle;
	struct tc_u_hnode	*ht_up;
	struct tcf_exts		exts;
#ifdef CONFIG_NET_CLS_IND
	char                     indev[IFNAMSIZ];
#endif
	u8			fshift;
	struct tcf_result	res;
	struct tc_u_hnode	*ht_down;
#ifdef CONFIG_CLS_U32_PERF
	struct tc_u32_pcnt	*pf;
#endif
#ifdef CONFIG_CLS_U32_MARK
	struct tc_u32_mark	mark;
#endif
	struct tc_u32_sel	sel;
};

struct tc_u_hnode
{
	struct tc_u_hnode	*next;
	u32			handle;
	u32			prio;
	struct tc_u_common	*tp_c;
	int			refcnt;
	unsigned		divisor;
	struct tc_u_knode	*ht[1];
};

struct tc_u_common
{
	struct tc_u_hnode	*hlist;
	struct Qdisc		*q;
	int			refcnt;
	u32			hgenerator;
};

static const struct tcf_ext_map u32_ext_map = {
	.action = TCA_U32_ACT,
	.police = TCA_U32_POLICE
};

static __inline__ unsigned u32_hash_fold(__be32 key, struct tc_u32_sel *sel, u8 fshift)
{
	unsigned h = ntohl(key & sel->hmask)>>fshift;

	return h;
}

static int u32_classify(struct sk_buff *skb, struct tcf_proto *tp, struct tcf_result *res)
{
	struct {
		struct tc_u_knode *knode;
		unsigned int	  off;
	} stack[TC_U32_MAXDEPTH];

	struct tc_u_hnode *ht = (struct tc_u_hnode*)tp->root;
	unsigned int off = skb_network_offset(skb);
	struct tc_u_knode *n;
	int sdepth = 0;
	int off2 = 0;
	int sel = 0;
#ifdef CONFIG_CLS_U32_PERF
	int j;
#endif
	int i, r;

next_ht:
	n = ht->ht[sel];

next_knode:
	if (n) {
		struct tc_u32_key *key = n->sel.keys;

#ifdef CONFIG_CLS_U32_PERF
		n->pf->rcnt +=1;
		j = 0;
#endif

#ifdef CONFIG_CLS_U32_MARK
		if ((skb->mark & n->mark.mask) != n->mark.val) {
			n = n->next;
			goto next_knode;
		} else {
			n->mark.success++;
		}
#endif

		for (i = n->sel.nkeys; i>0; i--, key++) {
			unsigned int toff;
			__be32 *data, _data;

			toff = off + key->off + (off2 & key->offmask);
			data = skb_header_pointer(skb, toff, 4, &_data);
			if (!data)
				goto out;
			if ((*data ^ key->val) & key->mask) {
				n = n->next;
				goto next_knode;
			}
#ifdef CONFIG_CLS_U32_PERF
			n->pf->kcnts[j] +=1;
			j++;
#endif
		}
		if (n->ht_down == NULL) {
check_terminal:
			if (n->sel.flags&TC_U32_TERMINAL) {

				*res = n->res;
#ifdef CONFIG_NET_CLS_IND
				if (!tcf_match_indev(skb, n->indev)) {
					n = n->next;
					goto next_knode;
				}
#endif
#ifdef CONFIG_CLS_U32_PERF
				n->pf->rhit +=1;
#endif
				r = tcf_exts_exec(skb, &n->exts, res);
				if (r < 0) {
					n = n->next;
					goto next_knode;
				}

				return r;
			}
			n = n->next;
			goto next_knode;
		}

		/* PUSH */
		if (sdepth >= TC_U32_MAXDEPTH)
			goto deadloop;
		stack[sdepth].knode = n;
		stack[sdepth].off = off;
		sdepth++;

		ht = n->ht_down;
		sel = 0;
		if (ht->divisor) {
			__be32 *data, _data;

			data = skb_header_pointer(skb, off + n->sel.hoff, 4,
						  &_data);
			if (!data)
				goto out;
			sel = ht->divisor & u32_hash_fold(*data, &n->sel,
							  n->fshift);
		}
		if (!(n->sel.flags&(TC_U32_VAROFFSET|TC_U32_OFFSET|TC_U32_EAT)))
			goto next_ht;

		if (n->sel.flags&(TC_U32_OFFSET|TC_U32_VAROFFSET)) {
			off2 = n->sel.off + 3;
			if (n->sel.flags & TC_U32_VAROFFSET) {
				__be16 *data, _data;

				data = skb_header_pointer(skb,
							  off + n->sel.offoff,
							  2, &_data);
				if (!data)
					goto out;
				off2 += ntohs(n->sel.offmask & *data) >>
					n->sel.offshift;
			}
			off2 &= ~3;
		}
		if (n->sel.flags&TC_U32_EAT) {
			off += off2;
			off2 = 0;
		}

		if (off < skb->len)
			goto next_ht;
	}

	/* POP */
	if (sdepth--) {
		n = stack[sdepth].knode;
		ht = n->ht_up;
		off = stack[sdepth].off;
		goto check_terminal;
	}
out:
	return -1;

deadloop:
	if (net_ratelimit())
		printk(KERN_WARNING "cls_u32: dead loop\n");
	return -1;
}

static __inline__ struct tc_u_hnode *
u32_lookup_ht(struct tc_u_common *tp_c, u32 handle)
{
	struct tc_u_hnode *ht;

	for (ht = tp_c->hlist; ht; ht = ht->next)
		if (ht->handle == handle)
			break;

	return ht;
}

static __inline__ struct tc_u_knode *
u32_lookup_key(struct tc_u_hnode *ht, u32 handle)
{
	unsigned sel;
	struct tc_u_knode *n = NULL;

	sel = TC_U32_HASH(handle);
	if (sel > ht->divisor)
		goto out;

	for (n = ht->ht[sel]; n; n = n->next)
		if (n->handle == handle)
			break;
out:
	return n;
}


static unsigned long u32_get(struct tcf_proto *tp, u32 handle)
{
	struct tc_u_hnode *ht;
	struct tc_u_common *tp_c = tp->data;

	if (TC_U32_HTID(handle) == TC_U32_ROOT)
		ht = tp->root;
	else
		ht = u32_lookup_ht(tp_c, TC_U32_HTID(handle));

	if (!ht)
		return 0;

	if (TC_U32_KEY(handle) == 0)
		return (unsigned long)ht;

	return (unsigned long)u32_lookup_key(ht, handle);
}

static void u32_put(struct tcf_proto *tp, unsigned long f)
{
}

static u32 gen_new_htid(struct tc_u_common *tp_c)
{
	int i = 0x800;

	do {
		if (++tp_c->hgenerator == 0x7FF)
			tp_c->hgenerator = 1;
	} while (--i>0 && u32_lookup_ht(tp_c, (tp_c->hgenerator|0x800)<<20));

	return i > 0 ? (tp_c->hgenerator|0x800)<<20 : 0;
}

static int u32_init(struct tcf_proto *tp)
{
	struct tc_u_hnode *root_ht;
	struct tc_u_common *tp_c;

	tp_c = tp->q->u32_node;

	root_ht = kzalloc(sizeof(*root_ht), GFP_KERNEL);
	if (root_ht == NULL)
		return -ENOBUFS;

	root_ht->divisor = 0;
	root_ht->refcnt++;
	root_ht->handle = tp_c ? gen_new_htid(tp_c) : 0x80000000;
	root_ht->prio = tp->prio;

	if (tp_c == NULL) {
		tp_c = kzalloc(sizeof(*tp_c), GFP_KERNEL);
		if (tp_c == NULL) {
			kfree(root_ht);
			return -ENOBUFS;
		}
		tp_c->q = tp->q;
		tp->q->u32_node = tp_c;
	}

	tp_c->refcnt++;
	root_ht->next = tp_c->hlist;
	tp_c->hlist = root_ht;
	root_ht->tp_c = tp_c;

	tp->root = root_ht;
	tp->data = tp_c;
	return 0;
}

static int u32_destroy_key(struct tcf_proto *tp, struct tc_u_knode *n)
{
	tcf_unbind_filter(tp, &n->res);
	tcf_exts_destroy(tp, &n->exts);
	if (n->ht_down)
		n->ht_down->refcnt--;
#ifdef CONFIG_CLS_U32_PERF
	kfree(n->pf);
#endif
	kfree(n);
	return 0;
}

static int u32_delete_key(struct tcf_proto *tp, struct tc_u_knode* key)
{
	struct tc_u_knode **kp;
	struct tc_u_hnode *ht = key->ht_up;

	if (ht) {
		for (kp = &ht->ht[TC_U32_HASH(key->handle)]; *kp; kp = &(*kp)->next) {
			if (*kp == key) {
				tcf_tree_lock(tp);
				*kp = key->next;
				tcf_tree_unlock(tp);

				u32_destroy_key(tp, key);
				return 0;
			}
		}
	}
	WARN_ON(1);
	return 0;
}

static void u32_clear_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
{
	struct tc_u_knode *n;
	unsigned h;

	for (h=0; h<=ht->divisor; h++) {
		while ((n = ht->ht[h]) != NULL) {
			ht->ht[h] = n->next;

			u32_destroy_key(tp, n);
		}
	}
}

static int u32_destroy_hnode(struct tcf_proto *tp, struct tc_u_hnode *ht)
{
	struct tc_u_common *tp_c = tp->data;
	struct tc_u_hnode **hn;

	WARN_ON(ht->refcnt);

	u32_clear_hnode(tp, ht);

	for (hn = &tp_c->hlist; *hn; hn = &(*hn)->next) {
		if (*hn == ht) {
			*hn = ht->next;
			kfree(ht);
			return 0;
		}
	}

	WARN_ON(1);
	return -ENOENT;
}

static void u32_destroy(struct tcf_proto *tp)
{
	struct tc_u_common *tp_c = tp->data;
	struct tc_u_hnode *root_ht = tp->root;

	WARN_ON(root_ht == NULL);

	if (root_ht && --root_ht->refcnt == 0)
		u32_destroy_hnode(tp, root_ht);

	if (--tp_c->refcnt == 0) {
		struct tc_u_hnode *ht;

		tp->q->u32_node = NULL;

		for (ht = tp_c->hlist; ht; ht = ht->next) {
			ht->refcnt--;
			u32_clear_hnode(tp, ht);
		}

		while ((ht = tp_c->hlist) != NULL) {
			tp_c->hlist = ht->next;

			WARN_ON(ht->refcnt != 0);

			kfree(ht);
		}

		kfree(tp_c);
	}

	tp->data = NULL;
}

static int u32_delete(struct tcf_proto *tp, unsigned long arg)
{
	struct tc_u_hnode *ht = (struct tc_u_hnode*)arg;

	if (ht == NULL)
		return 0;

	if (TC_U32_KEY(ht->handle))
		return u32_delete_key(tp, (struct tc_u_knode*)ht);

	if (tp->root == ht)
		return -EINVAL;

	if (ht->refcnt == 1) {
		ht->refcnt--;
		u32_destroy_hnode(tp, ht);
	} else {
		return -EBUSY;
	}

	return 0;
}

static u32 gen_new_kid(struct tc_u_hnode *ht, u32 handle)
{
	struct tc_u_knode *n;
	unsigned i = 0x7FF;

	for (n=ht->ht[TC_U32_HASH(handle)]; n; n = n->next)
		if (i < TC_U32_NODE(n->handle))
			i = TC_U32_NODE(n->handle);
	i++;

	return handle|(i>0xFFF ? 0xFFF : i);
}

static const struct nla_policy u32_policy[TCA_U32_MAX + 1] = {
	[TCA_U32_CLASSID]	= { .type = NLA_U32 },
	[TCA_U32_HASH]		= { .type = NLA_U32 },
	[TCA_U32_LINK]		= { .type = NLA_U32 },
	[TCA_U32_DIVISOR]	= { .type = NLA_U32 },
	[TCA_U32_SEL]		= { .len = sizeof(struct tc_u32_sel) },
	[TCA_U32_INDEV]		= { .type = NLA_STRING, .len = IFNAMSIZ },
	[TCA_U32_MARK]		= { .len = sizeof(struct tc_u32_mark) },
};

static int u32_set_parms(struct tcf_proto *tp, unsigned long base,
			 struct tc_u_hnode *ht,
			 struct tc_u_knode *n, struct nlattr **tb,
			 struct nlattr *est)
{
	int err;
	struct tcf_exts e;

	err = tcf_exts_validate(tp, tb, est, &e, &u32_ext_map);
	if (err < 0)
		return err;

	err = -EINVAL;
	if (tb[TCA_U32_LINK]) {
		u32 handle = nla_get_u32(tb[TCA_U32_LINK]);
		struct tc_u_hnode *ht_down = NULL, *ht_old;

		if (TC_U32_KEY(handle))
			goto errout;

		if (handle) {
			ht_down = u32_lookup_ht(ht->tp_c, handle);

			if (ht_down == NULL)
				goto errout;
			ht_down->refcnt++;
		}

		tcf_tree_lock(tp);
		ht_old = n->ht_down;
		n->ht_down = ht_down;
		tcf_tree_unlock(tp);

		if (ht_old)
			ht_old->refcnt--;
	}
	if (tb[TCA_U32_CLASSID]) {
		n->res.classid = nla_get_u32(tb[TCA_U32_CLASSID]);
		tcf_bind_filter(tp, &n->res, base);
	}

#ifdef CONFIG_NET_CLS_IND
	if (tb[TCA_U32_INDEV]) {
		err = tcf_change_indev(tp, n->indev, tb[TCA_U32_INDEV]);
		if (err < 0)
			goto errout;
	}
#endif
	tcf_exts_change(tp, &n->exts, &e);

	return 0;
errout:
	tcf_exts_destroy(tp, &e);
	return err;
}

static int u32_change(struct tcf_proto *tp, unsigned long base, u32 handle,
		      struct nlattr **tca,
		      unsigned long *arg)
{
	struct tc_u_common *tp_c = tp->data;
	struct tc_u_hnode *ht;
	struct tc_u_knode *n;
	struct tc_u32_sel *s;
	struct nlattr *opt = tca[TCA_OPTIONS];
	struct nlattr *tb[TCA_U32_MAX + 1];
	u32 htid;
	int err;

	if (opt == NULL)
		return handle ? -EINVAL : 0;

	err = nla_parse_nested(tb, TCA_U32_MAX, opt, u32_policy);
	if (err < 0)
		return err;

	if ((n = (struct tc_u_knode*)*arg) != NULL) {
		if (TC_U32_KEY(n->handle) == 0)
			return -EINVAL;

		return u32_set_parms(tp, base, n->ht_up, n, tb, tca[TCA_RATE]);
	}

	if (tb[TCA_U32_DIVISOR]) {
		unsigned divisor = nla_get_u32(tb[TCA_U32_DIVISOR]);

		if (--divisor > 0x100)
			return -EINVAL;
		if (TC_U32_KEY(handle))
			return -EINVAL;
		if (handle == 0) {
			handle = gen_new_htid(tp->data);
			if (handle == 0)
				return -ENOMEM;
		}
		ht = kzalloc(sizeof(*ht) + divisor*sizeof(void*), GFP_KERNEL);
		if (ht == NULL)
			return -ENOBUFS;
		ht->tp_c = tp_c;
		ht->refcnt = 1;
		ht->divisor = divisor;
		ht->handle = handle;
		ht->prio = tp->prio;
		ht->next = tp_c->hlist;
		tp_c->hlist = ht;
		*arg = (unsigned long)ht;
		return 0;
	}

	if (tb[TCA_U32_HASH]) {
		htid = nla_get_u32(tb[TCA_U32_HASH]);
		if (TC_U32_HTID(htid) == TC_U32_ROOT) {
			ht = tp->root;
			htid = ht->handle;
		} else {
			ht = u32_lookup_ht(tp->data, TC_U32_HTID(htid));
			if (ht == NULL)
				return -EINVAL;
		}
	} else {
		ht = tp->root;
		htid = ht->handle;
	}

	if (ht->divisor < TC_U32_HASH(htid))
		return -EINVAL;

	if (handle) {
		if (TC_U32_HTID(handle) && TC_U32_HTID(handle^htid))
			return -EINVAL;
		handle = htid | TC_U32_NODE(handle);
	} else
		handle = gen_new_kid(ht, htid);

	if (tb[TCA_U32_SEL] == NULL)
		return -EINVAL;

	s = nla_data(tb[TCA_U32_SEL]);

	n = kzalloc(sizeof(*n) + s->nkeys*sizeof(struct tc_u32_key), GFP_KERNEL);
	if (n == NULL)
		return -ENOBUFS;

#ifdef CONFIG_CLS_U32_PERF
	n->pf = kzalloc(sizeof(struct tc_u32_pcnt) + s->nkeys*sizeof(u64), GFP_KERNEL);
	if (n->pf == NULL) {
		kfree(n);
		return -ENOBUFS;
	}
#endif

	memcpy(&n->sel, s, sizeof(*s) + s->nkeys*sizeof(struct tc_u32_key));
	n->ht_up = ht;
	n->handle = handle;
	n->fshift = s->hmask ? ffs(ntohl(s->hmask)) - 1 : 0;

#ifdef CONFIG_CLS_U32_MARK
	if (tb[TCA_U32_MARK]) {
		struct tc_u32_mark *mark;

		mark = nla_data(tb[TCA_U32_MARK]);
		memcpy(&n->mark, mark, sizeof(struct tc_u32_mark));
		n->mark.success = 0;
	}
#endif

	err = u32_set_parms(tp, base, ht, n, tb, tca[TCA_RATE]);
	if (err == 0) {
		struct tc_u_knode **ins;
		for (ins = &ht->ht[TC_U32_HASH(handle)]; *ins; ins = &(*ins)->next)
			if (TC_U32_NODE(handle) < TC_U32_NODE((*ins)->handle))
				break;

		n->next = *ins;
		tcf_tree_lock(tp);
		*ins = n;
		tcf_tree_unlock(tp);

		*arg = (unsigned long)n;
		return 0;
	}
#ifdef CONFIG_CLS_U32_PERF
	kfree(n->pf);
#endif
	kfree(n);
	return err;
}

static void u32_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
	struct tc_u_common *tp_c = tp->data;
	struct tc_u_hnode *ht;
	struct tc_u_knode *n;
	unsigned h;

	if (arg->stop)
		return;

	for (ht = tp_c->hlist; ht; ht = ht->next) {
		if (ht->prio != tp->prio)
			continue;
		if (arg->count >= arg->skip) {
			if (arg->fn(tp, (unsigned long)ht, arg) < 0) {
				arg->stop = 1;
				return;
			}
		}
		arg->count++;
		for (h = 0; h <= ht->divisor; h++) {
			for (n = ht->ht[h]; n; n = n->next) {
				if (arg->count < arg->skip) {
					arg->count++;
					continue;
				}
				if (arg->fn(tp, (unsigned long)n, arg) < 0) {
					arg->stop = 1;
					return;
				}
				arg->count++;
			}
		}
	}
}

static int u32_dump(struct tcf_proto *tp, unsigned long fh,
		     struct sk_buff *skb, struct tcmsg *t)
{
	struct tc_u_knode *n = (struct tc_u_knode*)fh;
	struct nlattr *nest;

	if (n == NULL)
		return skb->len;

	t->tcm_handle = n->handle;

	nest = nla_nest_start(skb, TCA_OPTIONS);
	if (nest == NULL)
		goto nla_put_failure;

	if (TC_U32_KEY(n->handle) == 0) {
		struct tc_u_hnode *ht = (struct tc_u_hnode*)fh;
		u32 divisor = ht->divisor+1;
		NLA_PUT_U32(skb, TCA_U32_DIVISOR, divisor);
	} else {
		NLA_PUT(skb, TCA_U32_SEL,
			sizeof(n->sel) + n->sel.nkeys*sizeof(struct tc_u32_key),
			&n->sel);
		if (n->ht_up) {
			u32 htid = n->handle & 0xFFFFF000;
			NLA_PUT_U32(skb, TCA_U32_HASH, htid);
		}
		if (n->res.classid)
			NLA_PUT_U32(skb, TCA_U32_CLASSID, n->res.classid);
		if (n->ht_down)
			NLA_PUT_U32(skb, TCA_U32_LINK, n->ht_down->handle);

#ifdef CONFIG_CLS_U32_MARK
		if (n->mark.val || n->mark.mask)
			NLA_PUT(skb, TCA_U32_MARK, sizeof(n->mark), &n->mark);
#endif

		if (tcf_exts_dump(skb, &n->exts, &u32_ext_map) < 0)
			goto nla_put_failure;

#ifdef CONFIG_NET_CLS_IND
		if(strlen(n->indev))
			NLA_PUT_STRING(skb, TCA_U32_INDEV, n->indev);
#endif
#ifdef CONFIG_CLS_U32_PERF
		NLA_PUT(skb, TCA_U32_PCNT,
		sizeof(struct tc_u32_pcnt) + n->sel.nkeys*sizeof(u64),
			n->pf);
#endif
	}

	nla_nest_end(skb, nest);

	if (TC_U32_KEY(n->handle))
		if (tcf_exts_dump_stats(skb, &n->exts, &u32_ext_map) < 0)
			goto nla_put_failure;
	return skb->len;

nla_put_failure:
	nla_nest_cancel(skb, nest);
	return -1;
}

static struct tcf_proto_ops cls_u32_ops __read_mostly = {
	.kind		=	"u32",
	.classify	=	u32_classify,
	.init		=	u32_init,
	.destroy	=	u32_destroy,
	.get		=	u32_get,
	.put		=	u32_put,
	.change		=	u32_change,
	.delete		=	u32_delete,
	.walk		=	u32_walk,
	.dump		=	u32_dump,
	.owner		=	THIS_MODULE,
};

static int __init init_u32(void)
{
	pr_info("u32 classifier\n");
#ifdef CONFIG_CLS_U32_PERF
	pr_info("    Performance counters on\n");
#endif
#ifdef CONFIG_NET_CLS_IND
	pr_info("    input device check on\n");
#endif
#ifdef CONFIG_NET_CLS_ACT
	pr_info("    Actions configured\n");
#endif
	return register_tcf_proto_ops(&cls_u32_ops);
}

static void __exit exit_u32(void)
{
	unregister_tcf_proto_ops(&cls_u32_ops);
}

module_init(init_u32)
module_exit(exit_u32)
MODULE_LICENSE("GPL");
Commit	Line	Data
1da177e4 LT	1	/*
	2	* net/sched/cls_u32.c Ugly (or Universal) 32bit key Packet Classifier.
	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
	6	* as published by the Free Software Foundation; either version
	7	* 2 of the License, or (at your option) any later version.
	8	*
	9	* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
	10	*
	11	* The filters are packed to hash tables of key nodes
	12	* with a set of 32bit key/mask pairs at every node.
	13	* Nodes reference next level hash tables etc.
	14	*
	15	* This scheme is the best universal classifier I managed to
	16	* invent; it is not super-fast, but it is not slow (provided you
	17	* program it correctly), and general enough. And its relative
	18	* speed grows as the number of rules becomes larger.
	19	*
	20	* It seems that it represents the best middle point between
	21	* speed and manageability both by human and by machine.
	22	*
	23	* It is especially useful for link sharing combined with QoS;
	24	* pure RSVP doesn't need such a general approach and can use
	25	* much simpler (and faster) schemes, sort of cls_rsvp.c.
	26	*
	27	* JHS: We should remove the CONFIG_NET_CLS_IND from here
	28	* eventually when the meta match extension is made available
	29	*
	30	* nfmark match added by Catalin(ux aka Dino) BOIE <catab at umbrella.ro>
	31	*/
	32
1da177e4	33	#include <linux/module.h>
5a0e3ad6	34	#include <linux/slab.h>
1da177e4 LT	35	#include <linux/types.h>
1da177e4 LT	36	#include <linux/kernel.h>
1da177e4	37	#include <linux/string.h>
1da177e4	38	#include <linux/errno.h>
1da177e4	39	#include <linux/rtnetlink.h>
1da177e4	40	#include <linux/skbuff.h>
0ba48053	41	#include <net/netlink.h>
1da177e4 LT	42	#include <net/act_api.h>
	43	#include <net/pkt_cls.h>
	44
	45	struct tc_u_knode
	46	{
	47	struct tc_u_knode *next;
	48	u32 handle;
	49	struct tc_u_hnode *ht_up;
	50	struct tcf_exts exts;
	51	#ifdef CONFIG_NET_CLS_IND
	52	char indev[IFNAMSIZ];
	53	#endif
	54	u8 fshift;
	55	struct tcf_result res;
	56	struct tc_u_hnode *ht_down;
	57	#ifdef CONFIG_CLS_U32_PERF
	58	struct tc_u32_pcnt *pf;
	59	#endif
	60	#ifdef CONFIG_CLS_U32_MARK
	61	struct tc_u32_mark mark;
	62	#endif
	63	struct tc_u32_sel sel;
	64	};
	65
	66	struct tc_u_hnode
	67	{
	68	struct tc_u_hnode *next;
	69	u32 handle;
	70	u32 prio;
	71	struct tc_u_common *tp_c;
	72	int refcnt;
	73	unsigned divisor;
	74	struct tc_u_knode *ht[1];
	75	};
	76
	77	struct tc_u_common
	78	{
1da177e4 LT	79	struct tc_u_hnode *hlist;
	80	struct Qdisc *q;
	81	int refcnt;
	82	u32 hgenerator;
	83	};
	84
5239008b	85	static const struct tcf_ext_map u32_ext_map = {
1da177e4 LT	86	.action = TCA_U32_ACT,
	87	.police = TCA_U32_POLICE
	88	};
	89
0382b9c3	90	static __inline__ unsigned u32_hash_fold(__be32 key, struct tc_u32_sel *sel, u8 fshift)
1da177e4	91	{
543821c6	92	unsigned h = ntohl(key & sel->hmask)>>fshift;
1da177e4 LT	93
	94	return h;
	95	}
	96
	97	static int u32_classify(struct sk_buff skb, struct tcf_proto tp, struct tcf_result *res)
	98	{
	99	struct {
	100	struct tc_u_knode *knode;
fbc2e7d9	101	unsigned int off;
1da177e4 LT	102	} stack[TC_U32_MAXDEPTH];
	103
	104	struct tc_u_hnode ht = (struct tc_u_hnode)tp->root;
fbc2e7d9	105	unsigned int off = skb_network_offset(skb);
1da177e4 LT	106	struct tc_u_knode *n;
	107	int sdepth = 0;
	108	int off2 = 0;
	109	int sel = 0;
	110	#ifdef CONFIG_CLS_U32_PERF
	111	int j;
	112	#endif
	113	int i, r;
	114
	115	next_ht:
	116	n = ht->ht[sel];
	117
	118	next_knode:
	119	if (n) {
	120	struct tc_u32_key *key = n->sel.keys;
	121
	122	#ifdef CONFIG_CLS_U32_PERF
	123	n->pf->rcnt +=1;
	124	j = 0;
	125	#endif
	126
	127	#ifdef CONFIG_CLS_U32_MARK
82e91ffe	128	if ((skb->mark & n->mark.mask) != n->mark.val) {
1da177e4 LT	129	n = n->next;
	130	goto next_knode;
	131	} else {
	132	n->mark.success++;
	133	}
	134	#endif
	135
	136	for (i = n->sel.nkeys; i>0; i--, key++) {
fbc2e7d9 CG	137	unsigned int toff;
	138	__be32 *data, _data;
	139
	140	toff = off + key->off + (off2 & key->offmask);
	141	data = skb_header_pointer(skb, toff, 4, &_data);
	142	if (!data)
	143	goto out;
	144	if ((*data ^ key->val) & key->mask) {
1da177e4 LT	145	n = n->next;
	146	goto next_knode;
	147	}
	148	#ifdef CONFIG_CLS_U32_PERF
	149	n->pf->kcnts[j] +=1;
	150	j++;
	151	#endif
	152	}
	153	if (n->ht_down == NULL) {
	154	check_terminal:
	155	if (n->sel.flags&TC_U32_TERMINAL) {
	156
	157	*res = n->res;
	158	#ifdef CONFIG_NET_CLS_IND
	159	if (!tcf_match_indev(skb, n->indev)) {
	160	n = n->next;
	161	goto next_knode;
	162	}
	163	#endif
	164	#ifdef CONFIG_CLS_U32_PERF
	165	n->pf->rhit +=1;
	166	#endif
	167	r = tcf_exts_exec(skb, &n->exts, res);
	168	if (r < 0) {
	169	n = n->next;
	170	goto next_knode;
	171	}
	172
	173	return r;
	174	}
	175	n = n->next;
	176	goto next_knode;
	177	}
	178
	179	/* PUSH */
	180	if (sdepth >= TC_U32_MAXDEPTH)
	181	goto deadloop;
	182	stack[sdepth].knode = n;
fbc2e7d9	183	stack[sdepth].off = off;
1da177e4 LT	184	sdepth++;
	185
	186	ht = n->ht_down;
	187	sel = 0;
fbc2e7d9 CG	188	if (ht->divisor) {
	189	__be32 *data, _data;
	190
	191	data = skb_header_pointer(skb, off + n->sel.hoff, 4,
	192	&_data);
	193	if (!data)
	194	goto out;
	195	sel = ht->divisor & u32_hash_fold(*data, &n->sel,
	196	n->fshift);
	197	}
1da177e4 LT	198	if (!(n->sel.flags&(TC_U32_VAROFFSET\|TC_U32_OFFSET\|TC_U32_EAT)))
	199	goto next_ht;
	200
	201	if (n->sel.flags&(TC_U32_OFFSET\|TC_U32_VAROFFSET)) {
	202	off2 = n->sel.off + 3;
fbc2e7d9 CG	203	if (n->sel.flags & TC_U32_VAROFFSET) {
	204	__be16 *data, _data;
	205
	206	data = skb_header_pointer(skb,
	207	off + n->sel.offoff,
	208	2, &_data);
	209	if (!data)
	210	goto out;
	211	off2 += ntohs(n->sel.offmask & *data) >>
	212	n->sel.offshift;
	213	}
1da177e4 LT	214	off2 &= ~3;
	215	}
	216	if (n->sel.flags&TC_U32_EAT) {
fbc2e7d9	217	off += off2;
1da177e4 LT	218	off2 = 0;
	219	}
	220
fbc2e7d9	221	if (off < skb->len)
1da177e4 LT	222	goto next_ht;
	223	}
	224
	225	/* POP */
	226	if (sdepth--) {
	227	n = stack[sdepth].knode;
	228	ht = n->ht_up;
fbc2e7d9	229	off = stack[sdepth].off;
1da177e4 LT	230	goto check_terminal;
1da177e4 LT	231	}
fbc2e7d9	232	out:
1da177e4 LT	233	return -1;
	234
	235	deadloop:
	236	if (net_ratelimit())
6ff9c364	237	printk(KERN_WARNING "cls_u32: dead loop\n");
1da177e4 LT	238	return -1;
	239	}
	240
	241	static __inline__ struct tc_u_hnode *
	242	u32_lookup_ht(struct tc_u_common *tp_c, u32 handle)
	243	{
	244	struct tc_u_hnode *ht;
	245
	246	for (ht = tp_c->hlist; ht; ht = ht->next)
	247	if (ht->handle == handle)
	248	break;
	249
	250	return ht;
	251	}
	252
	253	static __inline__ struct tc_u_knode *
	254	u32_lookup_key(struct tc_u_hnode *ht, u32 handle)
	255	{
	256	unsigned sel;
	257	struct tc_u_knode *n = NULL;
	258
	259	sel = TC_U32_HASH(handle);
	260	if (sel > ht->divisor)
	261	goto out;
	262
	263	for (n = ht->ht[sel]; n; n = n->next)
	264	if (n->handle == handle)
	265	break;
	266	out:
	267	return n;
	268	}
	269
	270
	271	static unsigned long u32_get(struct tcf_proto *tp, u32 handle)
	272	{
	273	struct tc_u_hnode *ht;
	274	struct tc_u_common *tp_c = tp->data;
	275
	276	if (TC_U32_HTID(handle) == TC_U32_ROOT)
	277	ht = tp->root;
	278	else
	279	ht = u32_lookup_ht(tp_c, TC_U32_HTID(handle));
	280
	281	if (!ht)
	282	return 0;
	283
	284	if (TC_U32_KEY(handle) == 0)
	285	return (unsigned long)ht;
	286
	287	return (unsigned long)u32_lookup_key(ht, handle);
	288	}
	289
	290	static void u32_put(struct tcf_proto *tp, unsigned long f)
	291	{
	292	}
	293
	294	static u32 gen_new_htid(struct tc_u_common *tp_c)
	295	{
	296	int i = 0x800;
	297
	298	do {
	299	if (++tp_c->hgenerator == 0x7FF)
	300	tp_c->hgenerator = 1;
	301	} while (--i>0 && u32_lookup_ht(tp_c, (tp_c->hgenerator\|0x800)<<20));
302
303	return i > 0 ? (tp_c->hgenerator\|0x800)<<20 : 0;
304	}
305
306	static int u32_init(struct tcf_proto *tp)
307	{
308	struct tc_u_hnode *root_ht;
309	struct tc_u_common *tp_c;
310
72b25a91	311	tp_c = tp->q->u32_node;
1da177e4	312
0da974f4	313	root_ht = kzalloc(sizeof(*root_ht), GFP_KERNEL);
1da177e4 LT	314	if (root_ht == NULL)
	315	return -ENOBUFS;
	316
1da177e4 LT	317	root_ht->divisor = 0;
	318	root_ht->refcnt++;
	319	root_ht->handle = tp_c ? gen_new_htid(tp_c) : 0x80000000;
	320	root_ht->prio = tp->prio;
	321
	322	if (tp_c == NULL) {
0da974f4	323	tp_c = kzalloc(sizeof(*tp_c), GFP_KERNEL);
1da177e4 LT	324	if (tp_c == NULL) {
	325	kfree(root_ht);
	326	return -ENOBUFS;
	327	}
1da177e4	328	tp_c->q = tp->q;
72b25a91	329	tp->q->u32_node = tp_c;
1da177e4 LT	330	}
	331
	332	tp_c->refcnt++;
	333	root_ht->next = tp_c->hlist;
	334	tp_c->hlist = root_ht;
	335	root_ht->tp_c = tp_c;
	336
	337	tp->root = root_ht;
	338	tp->data = tp_c;
	339	return 0;
	340	}
	341
	342	static int u32_destroy_key(struct tcf_proto tp, struct tc_u_knode n)
	343	{
	344	tcf_unbind_filter(tp, &n->res);
	345	tcf_exts_destroy(tp, &n->exts);
	346	if (n->ht_down)
	347	n->ht_down->refcnt--;
	348	#ifdef CONFIG_CLS_U32_PERF
1ae39a43	349	kfree(n->pf);
1da177e4 LT	350	#endif
	351	kfree(n);
	352	return 0;
	353	}
	354
	355	static int u32_delete_key(struct tcf_proto tp, struct tc_u_knode key)
	356	{
	357	struct tc_u_knode **kp;
	358	struct tc_u_hnode *ht = key->ht_up;
	359
	360	if (ht) {
	361	for (kp = &ht->ht[TC_U32_HASH(key->handle)]; kp; kp = &(kp)->next) {
	362	if (*kp == key) {
	363	tcf_tree_lock(tp);
	364	*kp = key->next;
	365	tcf_tree_unlock(tp);
	366
	367	u32_destroy_key(tp, key);
	368	return 0;
	369	}
	370	}
	371	}
547b792c	372	WARN_ON(1);
1da177e4 LT	373	return 0;
	374	}
	375
	376	static void u32_clear_hnode(struct tcf_proto tp, struct tc_u_hnode ht)
	377	{
	378	struct tc_u_knode *n;
	379	unsigned h;
	380
	381	for (h=0; h<=ht->divisor; h++) {
	382	while ((n = ht->ht[h]) != NULL) {
	383	ht->ht[h] = n->next;
	384
	385	u32_destroy_key(tp, n);
	386	}
	387	}
	388	}
	389
	390	static int u32_destroy_hnode(struct tcf_proto tp, struct tc_u_hnode ht)
	391	{
	392	struct tc_u_common *tp_c = tp->data;
	393	struct tc_u_hnode **hn;
	394
547b792c	395	WARN_ON(ht->refcnt);
1da177e4 LT	396
	397	u32_clear_hnode(tp, ht);
	398
	399	for (hn = &tp_c->hlist; hn; hn = &(hn)->next) {
	400	if (*hn == ht) {
	401	*hn = ht->next;
	402	kfree(ht);
	403	return 0;
	404	}
	405	}
	406
547b792c	407	WARN_ON(1);
1da177e4 LT	408	return -ENOENT;
	409	}
	410
	411	static void u32_destroy(struct tcf_proto *tp)
	412	{
	413	struct tc_u_common *tp_c = tp->data;
47a1a1d4	414	struct tc_u_hnode *root_ht = tp->root;
1da177e4	415
547b792c	416	WARN_ON(root_ht == NULL);
1da177e4 LT	417
	418	if (root_ht && --root_ht->refcnt == 0)
	419	u32_destroy_hnode(tp, root_ht);
	420
	421	if (--tp_c->refcnt == 0) {
	422	struct tc_u_hnode *ht;
1da177e4	423
72b25a91	424	tp->q->u32_node = NULL;
1da177e4	425
e56cfad1 JP	426	for (ht = tp_c->hlist; ht; ht = ht->next) {
e56cfad1 JP	427	ht->refcnt--;
1da177e4	428	u32_clear_hnode(tp, ht);
e56cfad1	429	}
1da177e4 LT	430
	431	while ((ht = tp_c->hlist) != NULL) {
	432	tp_c->hlist = ht->next;
	433
547b792c	434	WARN_ON(ht->refcnt != 0);
1da177e4 LT	435
1da177e4 LT	436	kfree(ht);
3ff50b79	437	}
1da177e4 LT	438
	439	kfree(tp_c);
	440	}
	441
	442	tp->data = NULL;
	443	}
	444
	445	static int u32_delete(struct tcf_proto *tp, unsigned long arg)
	446	{
	447	struct tc_u_hnode ht = (struct tc_u_hnode)arg;
	448
	449	if (ht == NULL)
	450	return 0;
	451
	452	if (TC_U32_KEY(ht->handle))
	453	return u32_delete_key(tp, (struct tc_u_knode*)ht);
	454
	455	if (tp->root == ht)
	456	return -EINVAL;
	457
e56cfad1 JP	458	if (ht->refcnt == 1) {
e56cfad1 JP	459	ht->refcnt--;
1da177e4	460	u32_destroy_hnode(tp, ht);
e56cfad1 JP	461	} else {
	462	return -EBUSY;
	463	}
1da177e4 LT	464
	465	return 0;
	466	}
	467
	468	static u32 gen_new_kid(struct tc_u_hnode *ht, u32 handle)
	469	{
	470	struct tc_u_knode *n;
	471	unsigned i = 0x7FF;
	472
	473	for (n=ht->ht[TC_U32_HASH(handle)]; n; n = n->next)
	474	if (i < TC_U32_NODE(n->handle))
	475	i = TC_U32_NODE(n->handle);
	476	i++;
	477
	478	return handle\|(i>0xFFF ? 0xFFF : i);
	479	}
	480
6fa8c014 PM	481	static const struct nla_policy u32_policy[TCA_U32_MAX + 1] = {
	482	[TCA_U32_CLASSID] = { .type = NLA_U32 },
	483	[TCA_U32_HASH] = { .type = NLA_U32 },
	484	[TCA_U32_LINK] = { .type = NLA_U32 },
	485	[TCA_U32_DIVISOR] = { .type = NLA_U32 },
	486	[TCA_U32_SEL] = { .len = sizeof(struct tc_u32_sel) },
	487	[TCA_U32_INDEV] = { .type = NLA_STRING, .len = IFNAMSIZ },
	488	[TCA_U32_MARK] = { .len = sizeof(struct tc_u32_mark) },
	489	};
	490
1da177e4 LT	491	static int u32_set_parms(struct tcf_proto *tp, unsigned long base,
1da177e4 LT	492	struct tc_u_hnode *ht,
add93b61 PM	493	struct tc_u_knode n, struct nlattr *tb,
add93b61 PM	494	struct nlattr *est)
1da177e4 LT	495	{
	496	int err;
	497	struct tcf_exts e;
	498
	499	err = tcf_exts_validate(tp, tb, est, &e, &u32_ext_map);
	500	if (err < 0)
	501	return err;
	502
	503	err = -EINVAL;
add93b61	504	if (tb[TCA_U32_LINK]) {
1587bac4	505	u32 handle = nla_get_u32(tb[TCA_U32_LINK]);
47a1a1d4	506	struct tc_u_hnode ht_down = NULL, ht_old;
1da177e4 LT	507
	508	if (TC_U32_KEY(handle))
	509	goto errout;
	510
	511	if (handle) {
	512	ht_down = u32_lookup_ht(ht->tp_c, handle);
	513
	514	if (ht_down == NULL)
	515	goto errout;
	516	ht_down->refcnt++;
	517	}
	518
	519	tcf_tree_lock(tp);
47a1a1d4 PM	520	ht_old = n->ht_down;
47a1a1d4 PM	521	n->ht_down = ht_down;
1da177e4 LT	522	tcf_tree_unlock(tp);
1da177e4 LT	523
47a1a1d4 PM	524	if (ht_old)
47a1a1d4 PM	525	ht_old->refcnt--;
1da177e4	526	}
add93b61	527	if (tb[TCA_U32_CLASSID]) {
1587bac4	528	n->res.classid = nla_get_u32(tb[TCA_U32_CLASSID]);
1da177e4 LT	529	tcf_bind_filter(tp, &n->res, base);
	530	}
	531
	532	#ifdef CONFIG_NET_CLS_IND
add93b61 PM	533	if (tb[TCA_U32_INDEV]) {
add93b61 PM	534	err = tcf_change_indev(tp, n->indev, tb[TCA_U32_INDEV]);
1da177e4 LT	535	if (err < 0)
	536	goto errout;
	537	}
	538	#endif
	539	tcf_exts_change(tp, &n->exts, &e);
	540
	541	return 0;
	542	errout:
	543	tcf_exts_destroy(tp, &e);
	544	return err;
	545	}
	546
	547	static int u32_change(struct tcf_proto *tp, unsigned long base, u32 handle,
add93b61	548	struct nlattr **tca,
1da177e4 LT	549	unsigned long *arg)
	550	{
	551	struct tc_u_common *tp_c = tp->data;
	552	struct tc_u_hnode *ht;
	553	struct tc_u_knode *n;
	554	struct tc_u32_sel *s;
add93b61 PM	555	struct nlattr *opt = tca[TCA_OPTIONS];
add93b61 PM	556	struct nlattr *tb[TCA_U32_MAX + 1];
1da177e4 LT	557	u32 htid;
	558	int err;
	559
	560	if (opt == NULL)
	561	return handle ? -EINVAL : 0;
	562
6fa8c014	563	err = nla_parse_nested(tb, TCA_U32_MAX, opt, u32_policy);
cee63723 PM	564	if (err < 0)
cee63723 PM	565	return err;
1da177e4 LT	566
	567	if ((n = (struct tc_u_knode)arg) != NULL) {
	568	if (TC_U32_KEY(n->handle) == 0)
	569	return -EINVAL;
	570
add93b61	571	return u32_set_parms(tp, base, n->ht_up, n, tb, tca[TCA_RATE]);
1da177e4 LT	572	}
1da177e4 LT	573
add93b61	574	if (tb[TCA_U32_DIVISOR]) {
1587bac4	575	unsigned divisor = nla_get_u32(tb[TCA_U32_DIVISOR]);
1da177e4 LT	576
	577	if (--divisor > 0x100)
	578	return -EINVAL;
	579	if (TC_U32_KEY(handle))
	580	return -EINVAL;
	581	if (handle == 0) {
	582	handle = gen_new_htid(tp->data);
	583	if (handle == 0)
	584	return -ENOMEM;
	585	}
0da974f4	586	ht = kzalloc(sizeof(ht) + divisorsizeof(void*), GFP_KERNEL);
1da177e4 LT	587	if (ht == NULL)
1da177e4 LT	588	return -ENOBUFS;
1da177e4	589	ht->tp_c = tp_c;
e56cfad1	590	ht->refcnt = 1;
1da177e4 LT	591	ht->divisor = divisor;
	592	ht->handle = handle;
	593	ht->prio = tp->prio;
	594	ht->next = tp_c->hlist;
	595	tp_c->hlist = ht;
	596	*arg = (unsigned long)ht;
	597	return 0;
	598	}
	599
add93b61	600	if (tb[TCA_U32_HASH]) {
1587bac4	601	htid = nla_get_u32(tb[TCA_U32_HASH]);
1da177e4 LT	602	if (TC_U32_HTID(htid) == TC_U32_ROOT) {
	603	ht = tp->root;
	604	htid = ht->handle;
	605	} else {
	606	ht = u32_lookup_ht(tp->data, TC_U32_HTID(htid));
	607	if (ht == NULL)
	608	return -EINVAL;
	609	}
	610	} else {
	611	ht = tp->root;
	612	htid = ht->handle;
	613	}
	614
	615	if (ht->divisor < TC_U32_HASH(htid))
	616	return -EINVAL;
	617
	618	if (handle) {
	619	if (TC_U32_HTID(handle) && TC_U32_HTID(handle^htid))
	620	return -EINVAL;
	621	handle = htid \| TC_U32_NODE(handle);
	622	} else
	623	handle = gen_new_kid(ht, htid);
	624
6fa8c014	625	if (tb[TCA_U32_SEL] == NULL)
1da177e4 LT	626	return -EINVAL;
1da177e4 LT	627
add93b61	628	s = nla_data(tb[TCA_U32_SEL]);
1da177e4	629
0da974f4	630	n = kzalloc(sizeof(n) + s->nkeyssizeof(struct tc_u32_key), GFP_KERNEL);
1da177e4 LT	631	if (n == NULL)
	632	return -ENOBUFS;
	633
1da177e4	634	#ifdef CONFIG_CLS_U32_PERF
0da974f4	635	n->pf = kzalloc(sizeof(struct tc_u32_pcnt) + s->nkeys*sizeof(u64), GFP_KERNEL);
1da177e4 LT	636	if (n->pf == NULL) {
	637	kfree(n);
	638	return -ENOBUFS;
	639	}
1da177e4 LT	640	#endif
	641
	642	memcpy(&n->sel, s, sizeof(s) + s->nkeyssizeof(struct tc_u32_key));
	643	n->ht_up = ht;
	644	n->handle = handle;
b2268016	645	n->fshift = s->hmask ? ffs(ntohl(s->hmask)) - 1 : 0;
1da177e4 LT	646
1da177e4 LT	647	#ifdef CONFIG_CLS_U32_MARK
add93b61	648	if (tb[TCA_U32_MARK]) {
1da177e4 LT	649	struct tc_u32_mark *mark;
1da177e4 LT	650
add93b61	651	mark = nla_data(tb[TCA_U32_MARK]);
1da177e4 LT	652	memcpy(&n->mark, mark, sizeof(struct tc_u32_mark));
	653	n->mark.success = 0;
	654	}
	655	#endif
	656
add93b61	657	err = u32_set_parms(tp, base, ht, n, tb, tca[TCA_RATE]);
1da177e4 LT	658	if (err == 0) {
	659	struct tc_u_knode **ins;
	660	for (ins = &ht->ht[TC_U32_HASH(handle)]; ins; ins = &(ins)->next)
	661	if (TC_U32_NODE(handle) < TC_U32_NODE((*ins)->handle))
	662	break;
	663
	664	n->next = *ins;
6f573214	665	tcf_tree_lock(tp);
1da177e4	666	*ins = n;
6f573214	667	tcf_tree_unlock(tp);
1da177e4 LT	668
	669	*arg = (unsigned long)n;
	670	return 0;
	671	}
	672	#ifdef CONFIG_CLS_U32_PERF
1ae39a43	673	kfree(n->pf);
1da177e4 LT	674	#endif
	675	kfree(n);
	676	return err;
	677	}
	678
	679	static void u32_walk(struct tcf_proto tp, struct tcf_walker arg)
	680	{
	681	struct tc_u_common *tp_c = tp->data;
	682	struct tc_u_hnode *ht;
	683	struct tc_u_knode *n;
	684	unsigned h;
	685
	686	if (arg->stop)
	687	return;
	688
	689	for (ht = tp_c->hlist; ht; ht = ht->next) {
	690	if (ht->prio != tp->prio)
	691	continue;
	692	if (arg->count >= arg->skip) {
	693	if (arg->fn(tp, (unsigned long)ht, arg) < 0) {
	694	arg->stop = 1;
	695	return;
	696	}
	697	}
	698	arg->count++;
	699	for (h = 0; h <= ht->divisor; h++) {
	700	for (n = ht->ht[h]; n; n = n->next) {
	701	if (arg->count < arg->skip) {
	702	arg->count++;
	703	continue;
	704	}
	705	if (arg->fn(tp, (unsigned long)n, arg) < 0) {
	706	arg->stop = 1;
	707	return;
	708	}
	709	arg->count++;
	710	}
	711	}
	712	}
	713	}
	714
	715	static int u32_dump(struct tcf_proto *tp, unsigned long fh,
	716	struct sk_buff skb, struct tcmsg t)
	717	{
	718	struct tc_u_knode n = (struct tc_u_knode)fh;
4b3550ef	719	struct nlattr *nest;
1da177e4 LT	720
	721	if (n == NULL)
	722	return skb->len;
	723
	724	t->tcm_handle = n->handle;
	725
4b3550ef PM	726	nest = nla_nest_start(skb, TCA_OPTIONS);
	727	if (nest == NULL)
	728	goto nla_put_failure;
1da177e4 LT	729
	730	if (TC_U32_KEY(n->handle) == 0) {
	731	struct tc_u_hnode ht = (struct tc_u_hnode)fh;
	732	u32 divisor = ht->divisor+1;
24beeab5	733	NLA_PUT_U32(skb, TCA_U32_DIVISOR, divisor);
1da177e4	734	} else {
add93b61	735	NLA_PUT(skb, TCA_U32_SEL,
1da177e4 LT	736	sizeof(n->sel) + n->sel.nkeys*sizeof(struct tc_u32_key),
	737	&n->sel);
	738	if (n->ht_up) {
	739	u32 htid = n->handle & 0xFFFFF000;
24beeab5	740	NLA_PUT_U32(skb, TCA_U32_HASH, htid);
1da177e4 LT	741	}
1da177e4 LT	742	if (n->res.classid)
24beeab5	743	NLA_PUT_U32(skb, TCA_U32_CLASSID, n->res.classid);
1da177e4	744	if (n->ht_down)
24beeab5	745	NLA_PUT_U32(skb, TCA_U32_LINK, n->ht_down->handle);
1da177e4 LT	746
	747	#ifdef CONFIG_CLS_U32_MARK
	748	if (n->mark.val \|\| n->mark.mask)
add93b61	749	NLA_PUT(skb, TCA_U32_MARK, sizeof(n->mark), &n->mark);
1da177e4 LT	750	#endif
	751
	752	if (tcf_exts_dump(skb, &n->exts, &u32_ext_map) < 0)
add93b61	753	goto nla_put_failure;
1da177e4 LT	754
	755	#ifdef CONFIG_NET_CLS_IND
	756	if(strlen(n->indev))
57e1c487	757	NLA_PUT_STRING(skb, TCA_U32_INDEV, n->indev);
1da177e4 LT	758	#endif
1da177e4 LT	759	#ifdef CONFIG_CLS_U32_PERF
add93b61	760	NLA_PUT(skb, TCA_U32_PCNT,
1da177e4 LT	761	sizeof(struct tc_u32_pcnt) + n->sel.nkeys*sizeof(u64),
	762	n->pf);
	763	#endif
	764	}
	765
4b3550ef PM	766	nla_nest_end(skb, nest);
4b3550ef PM	767
1da177e4 LT	768	if (TC_U32_KEY(n->handle))
1da177e4 LT	769	if (tcf_exts_dump_stats(skb, &n->exts, &u32_ext_map) < 0)
add93b61	770	goto nla_put_failure;
1da177e4 LT	771	return skb->len;
1da177e4 LT	772
add93b61	773	nla_put_failure:
4b3550ef	774	nla_nest_cancel(skb, nest);
1da177e4 LT	775	return -1;
	776	}
	777
2eb9d75c	778	static struct tcf_proto_ops cls_u32_ops __read_mostly = {
1da177e4 LT	779	.kind = "u32",
	780	.classify = u32_classify,
	781	.init = u32_init,
	782	.destroy = u32_destroy,
	783	.get = u32_get,
	784	.put = u32_put,
	785	.change = u32_change,
	786	.delete = u32_delete,
	787	.walk = u32_walk,
	788	.dump = u32_dump,
	789	.owner = THIS_MODULE,
	790	};
	791
	792	static int __init init_u32(void)
	793	{
6ff9c364	794	pr_info("u32 classifier\n");
1da177e4	795	#ifdef CONFIG_CLS_U32_PERF
6ff9c364	796	pr_info(" Performance counters on\n");
1da177e4	797	#endif
1da177e4	798	#ifdef CONFIG_NET_CLS_IND
6ff9c364	799	pr_info(" input device check on\n");
1da177e4 LT	800	#endif
1da177e4 LT	801	#ifdef CONFIG_NET_CLS_ACT
6ff9c364	802	pr_info(" Actions configured\n");
1da177e4 LT	803	#endif
	804	return register_tcf_proto_ops(&cls_u32_ops);
	805	}
	806
10297b99	807	static void __exit exit_u32(void)
1da177e4 LT	808	{
	809	unregister_tcf_proto_ops(&cls_u32_ops);
	810	}
	811
	812	module_init(init_u32)
	813	module_exit(exit_u32)
	814	MODULE_LICENSE("GPL");