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

/*
 * net/sched/cls_flow.c		Generic flow classifier
 *
 * Copyright (c) 2007, 2008 Patrick McHardy <kaber@trash.net>
 *
 * 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.
 */

#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/list.h>
#include <linux/jhash.h>
#include <linux/random.h>
#include <linux/pkt_cls.h>
#include <linux/skbuff.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/if_vlan.h>
#include <linux/slab.h>

#include <net/pkt_cls.h>
#include <net/ip.h>
#include <net/route.h>
#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
#include <net/netfilter/nf_conntrack.h>
#endif

struct flow_head {
	struct list_head	filters;
};

struct flow_filter {
	struct list_head	list;
	struct tcf_exts		exts;
	struct tcf_ematch_tree	ematches;
	struct timer_list	perturb_timer;
	u32			perturb_period;
	u32			handle;

	u32			nkeys;
	u32			keymask;
	u32			mode;
	u32			mask;
	u32			xor;
	u32			rshift;
	u32			addend;
	u32			divisor;
	u32			baseclass;
	u32			hashrnd;
};

static const struct tcf_ext_map flow_ext_map = {
	.action	= TCA_FLOW_ACT,
	.police	= TCA_FLOW_POLICE,
};

static inline u32 addr_fold(void *addr)
{
	unsigned long a = (unsigned long)addr;

	return (a & 0xFFFFFFFF) ^ (BITS_PER_LONG > 32 ? a >> 32 : 0);
}

static u32 flow_get_src(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP):
		if (pskb_network_may_pull(skb, sizeof(struct iphdr)))
			return ntohl(ip_hdr(skb)->saddr);
		break;
	case htons(ETH_P_IPV6):
		if (pskb_network_may_pull(skb, sizeof(struct ipv6hdr)))
			return ntohl(ipv6_hdr(skb)->saddr.s6_addr32[3]);
		break;
	}

	return addr_fold(skb->sk);
}

static u32 flow_get_dst(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP):
		if (pskb_network_may_pull(skb, sizeof(struct iphdr)))
			return ntohl(ip_hdr(skb)->daddr);
		break;
	case htons(ETH_P_IPV6):
		if (pskb_network_may_pull(skb, sizeof(struct ipv6hdr)))
			return ntohl(ipv6_hdr(skb)->daddr.s6_addr32[3]);
		break;
	}

	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
}

static u32 flow_get_proto(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP):
		return pskb_network_may_pull(skb, sizeof(struct iphdr)) ?
		       ip_hdr(skb)->protocol : 0;
	case htons(ETH_P_IPV6):
		return pskb_network_may_pull(skb, sizeof(struct ipv6hdr)) ?
		       ipv6_hdr(skb)->nexthdr : 0;
	default:
		return 0;
	}
}

static u32 flow_get_proto_src(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP): {
		struct iphdr *iph;
		int poff;

		if (!pskb_network_may_pull(skb, sizeof(*iph)))
			break;
		iph = ip_hdr(skb);
		if (iph->frag_off & htons(IP_MF|IP_OFFSET))
			break;
		poff = proto_ports_offset(iph->protocol);
		if (poff >= 0 &&
		    pskb_network_may_pull(skb, iph->ihl * 4 + 2 + poff)) {
			iph = ip_hdr(skb);
			return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4 +
						 poff));
		}
		break;
	}
	case htons(ETH_P_IPV6): {
		struct ipv6hdr *iph;
		int poff;

		if (!pskb_network_may_pull(skb, sizeof(*iph)))
			break;
		iph = ipv6_hdr(skb);
		poff = proto_ports_offset(iph->nexthdr);
		if (poff >= 0 &&
		    pskb_network_may_pull(skb, sizeof(*iph) + poff + 2)) {
			iph = ipv6_hdr(skb);
			return ntohs(*(__be16 *)((void *)iph + sizeof(*iph) +
						 poff));
		}
		break;
	}
	}

	return addr_fold(skb->sk);
}

static u32 flow_get_proto_dst(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP): {
		struct iphdr *iph;
		int poff;

		if (!pskb_network_may_pull(skb, sizeof(*iph)))
			break;
		iph = ip_hdr(skb);
		if (iph->frag_off & htons(IP_MF|IP_OFFSET))
			break;
		poff = proto_ports_offset(iph->protocol);
		if (poff >= 0 &&
		    pskb_network_may_pull(skb, iph->ihl * 4 + 4 + poff)) {
			iph = ip_hdr(skb);
			return ntohs(*(__be16 *)((void *)iph + iph->ihl * 4 +
						 2 + poff));
		}
		break;
	}
	case htons(ETH_P_IPV6): {
		struct ipv6hdr *iph;
		int poff;

		if (!pskb_network_may_pull(skb, sizeof(*iph)))
			break;
		iph = ipv6_hdr(skb);
		poff = proto_ports_offset(iph->nexthdr);
		if (poff >= 0 &&
		    pskb_network_may_pull(skb, sizeof(*iph) + poff + 4)) {
			iph = ipv6_hdr(skb);
			return ntohs(*(__be16 *)((void *)iph + sizeof(*iph) +
						 poff + 2));
		}
		break;
	}
	}

	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
}

static u32 flow_get_iif(const struct sk_buff *skb)
{
	return skb->skb_iif;
}

static u32 flow_get_priority(const struct sk_buff *skb)
{
	return skb->priority;
}

static u32 flow_get_mark(const struct sk_buff *skb)
{
	return skb->mark;
}

static u32 flow_get_nfct(const struct sk_buff *skb)
{
#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
	return addr_fold(skb->nfct);
#else
	return 0;
#endif
}

#if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE)
#define CTTUPLE(skb, member)						\
({									\
	enum ip_conntrack_info ctinfo;					\
	struct nf_conn *ct = nf_ct_get(skb, &ctinfo);			\
	if (ct == NULL)							\
		goto fallback;						\
	ct->tuplehash[CTINFO2DIR(ctinfo)].tuple.member;			\
})
#else
#define CTTUPLE(skb, member)						\
({									\
	goto fallback;							\
	0;								\
})
#endif

static u32 flow_get_nfct_src(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP):
		return ntohl(CTTUPLE(skb, src.u3.ip));
	case htons(ETH_P_IPV6):
		return ntohl(CTTUPLE(skb, src.u3.ip6[3]));
	}
fallback:
	return flow_get_src(skb);
}

static u32 flow_get_nfct_dst(struct sk_buff *skb)
{
	switch (skb->protocol) {
	case htons(ETH_P_IP):
		return ntohl(CTTUPLE(skb, dst.u3.ip));
	case htons(ETH_P_IPV6):
		return ntohl(CTTUPLE(skb, dst.u3.ip6[3]));
	}
fallback:
	return flow_get_dst(skb);
}

static u32 flow_get_nfct_proto_src(struct sk_buff *skb)
{
	return ntohs(CTTUPLE(skb, src.u.all));
fallback:
	return flow_get_proto_src(skb);
}

static u32 flow_get_nfct_proto_dst(struct sk_buff *skb)
{
	return ntohs(CTTUPLE(skb, dst.u.all));
fallback:
	return flow_get_proto_dst(skb);
}

static u32 flow_get_rtclassid(const struct sk_buff *skb)
{
#ifdef CONFIG_NET_CLS_ROUTE
	if (skb_dst(skb))
		return skb_dst(skb)->tclassid;
#endif
	return 0;
}

static u32 flow_get_skuid(const struct sk_buff *skb)
{
	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
		return skb->sk->sk_socket->file->f_cred->fsuid;
	return 0;
}

static u32 flow_get_skgid(const struct sk_buff *skb)
{
	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
		return skb->sk->sk_socket->file->f_cred->fsgid;
	return 0;
}

static u32 flow_get_vlan_tag(const struct sk_buff *skb)
{
	u16 uninitialized_var(tag);

	if (vlan_get_tag(skb, &tag) < 0)
		return 0;
	return tag & VLAN_VID_MASK;
}

static u32 flow_get_rxhash(struct sk_buff *skb)
{
	return skb_get_rxhash(skb);
}

static u32 flow_key_get(struct sk_buff *skb, int key)
{
	switch (key) {
	case FLOW_KEY_SRC:
		return flow_get_src(skb);
	case FLOW_KEY_DST:
		return flow_get_dst(skb);
	case FLOW_KEY_PROTO:
		return flow_get_proto(skb);
	case FLOW_KEY_PROTO_SRC:
		return flow_get_proto_src(skb);
	case FLOW_KEY_PROTO_DST:
		return flow_get_proto_dst(skb);
	case FLOW_KEY_IIF:
		return flow_get_iif(skb);
	case FLOW_KEY_PRIORITY:
		return flow_get_priority(skb);
	case FLOW_KEY_MARK:
		return flow_get_mark(skb);
	case FLOW_KEY_NFCT:
		return flow_get_nfct(skb);
	case FLOW_KEY_NFCT_SRC:
		return flow_get_nfct_src(skb);
	case FLOW_KEY_NFCT_DST:
		return flow_get_nfct_dst(skb);
	case FLOW_KEY_NFCT_PROTO_SRC:
		return flow_get_nfct_proto_src(skb);
	case FLOW_KEY_NFCT_PROTO_DST:
		return flow_get_nfct_proto_dst(skb);
	case FLOW_KEY_RTCLASSID:
		return flow_get_rtclassid(skb);
	case FLOW_KEY_SKUID:
		return flow_get_skuid(skb);
	case FLOW_KEY_SKGID:
		return flow_get_skgid(skb);
	case FLOW_KEY_VLAN_TAG:
		return flow_get_vlan_tag(skb);
	case FLOW_KEY_RXHASH:
		return flow_get_rxhash(skb);
	default:
		WARN_ON(1);
		return 0;
	}
}

static int flow_classify(struct sk_buff *skb, struct tcf_proto *tp,
			 struct tcf_result *res)
{
	struct flow_head *head = tp->root;
	struct flow_filter *f;
	u32 keymask;
	u32 classid;
	unsigned int n, key;
	int r;

	list_for_each_entry(f, &head->filters, list) {
		u32 keys[f->nkeys];

		if (!tcf_em_tree_match(skb, &f->ematches, NULL))
			continue;

		keymask = f->keymask;

		for (n = 0; n < f->nkeys; n++) {
			key = ffs(keymask) - 1;
			keymask &= ~(1 << key);
			keys[n] = flow_key_get(skb, key);
		}

		if (f->mode == FLOW_MODE_HASH)
			classid = jhash2(keys, f->nkeys, f->hashrnd);
		else {
			classid = keys[0];
			classid = (classid & f->mask) ^ f->xor;
			classid = (classid >> f->rshift) + f->addend;
		}

		if (f->divisor)
			classid %= f->divisor;

		res->class   = 0;
		res->classid = TC_H_MAKE(f->baseclass, f->baseclass + classid);

		r = tcf_exts_exec(skb, &f->exts, res);
		if (r < 0)
			continue;
		return r;
	}
	return -1;
}

static void flow_perturbation(unsigned long arg)
{
	struct flow_filter *f = (struct flow_filter *)arg;

	get_random_bytes(&f->hashrnd, 4);
	if (f->perturb_period)
		mod_timer(&f->perturb_timer, jiffies + f->perturb_period);
}

static const struct nla_policy flow_policy[TCA_FLOW_MAX + 1] = {
	[TCA_FLOW_KEYS]		= { .type = NLA_U32 },
	[TCA_FLOW_MODE]		= { .type = NLA_U32 },
	[TCA_FLOW_BASECLASS]	= { .type = NLA_U32 },
	[TCA_FLOW_RSHIFT]	= { .type = NLA_U32 },
	[TCA_FLOW_ADDEND]	= { .type = NLA_U32 },
	[TCA_FLOW_MASK]		= { .type = NLA_U32 },
	[TCA_FLOW_XOR]		= { .type = NLA_U32 },
	[TCA_FLOW_DIVISOR]	= { .type = NLA_U32 },
	[TCA_FLOW_ACT]		= { .type = NLA_NESTED },
	[TCA_FLOW_POLICE]	= { .type = NLA_NESTED },
	[TCA_FLOW_EMATCHES]	= { .type = NLA_NESTED },
	[TCA_FLOW_PERTURB]	= { .type = NLA_U32 },
};

static int flow_change(struct tcf_proto *tp, unsigned long base,
		       u32 handle, struct nlattr **tca,
		       unsigned long *arg)
{
	struct flow_head *head = tp->root;
	struct flow_filter *f;
	struct nlattr *opt = tca[TCA_OPTIONS];
	struct nlattr *tb[TCA_FLOW_MAX + 1];
	struct tcf_exts e;
	struct tcf_ematch_tree t;
	unsigned int nkeys = 0;
	unsigned int perturb_period = 0;
	u32 baseclass = 0;
	u32 keymask = 0;
	u32 mode;
	int err;

	if (opt == NULL)
		return -EINVAL;

	err = nla_parse_nested(tb, TCA_FLOW_MAX, opt, flow_policy);
	if (err < 0)
		return err;

	if (tb[TCA_FLOW_BASECLASS]) {
		baseclass = nla_get_u32(tb[TCA_FLOW_BASECLASS]);
		if (TC_H_MIN(baseclass) == 0)
			return -EINVAL;
	}

	if (tb[TCA_FLOW_KEYS]) {
		keymask = nla_get_u32(tb[TCA_FLOW_KEYS]);

		nkeys = hweight32(keymask);
		if (nkeys == 0)
			return -EINVAL;

		if (fls(keymask) - 1 > FLOW_KEY_MAX)
			return -EOPNOTSUPP;
	}

	err = tcf_exts_validate(tp, tb, tca[TCA_RATE], &e, &flow_ext_map);
	if (err < 0)
		return err;

	err = tcf_em_tree_validate(tp, tb[TCA_FLOW_EMATCHES], &t);
	if (err < 0)
		goto err1;

	f = (struct flow_filter *)*arg;
	if (f != NULL) {
		err = -EINVAL;
		if (f->handle != handle && handle)
			goto err2;

		mode = f->mode;
		if (tb[TCA_FLOW_MODE])
			mode = nla_get_u32(tb[TCA_FLOW_MODE]);
		if (mode != FLOW_MODE_HASH && nkeys > 1)
			goto err2;

		if (mode == FLOW_MODE_HASH)
			perturb_period = f->perturb_period;
		if (tb[TCA_FLOW_PERTURB]) {
			if (mode != FLOW_MODE_HASH)
				goto err2;
			perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
		}
	} else {
		err = -EINVAL;
		if (!handle)
			goto err2;
		if (!tb[TCA_FLOW_KEYS])
			goto err2;

		mode = FLOW_MODE_MAP;
		if (tb[TCA_FLOW_MODE])
			mode = nla_get_u32(tb[TCA_FLOW_MODE]);
		if (mode != FLOW_MODE_HASH && nkeys > 1)
			goto err2;

		if (tb[TCA_FLOW_PERTURB]) {
			if (mode != FLOW_MODE_HASH)
				goto err2;
			perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
		}

		if (TC_H_MAJ(baseclass) == 0)
			baseclass = TC_H_MAKE(tp->q->handle, baseclass);
		if (TC_H_MIN(baseclass) == 0)
			baseclass = TC_H_MAKE(baseclass, 1);

		err = -ENOBUFS;
		f = kzalloc(sizeof(*f), GFP_KERNEL);
		if (f == NULL)
			goto err2;

		f->handle = handle;
		f->mask	  = ~0U;

		get_random_bytes(&f->hashrnd, 4);
		f->perturb_timer.function = flow_perturbation;
		f->perturb_timer.data = (unsigned long)f;
		init_timer_deferrable(&f->perturb_timer);
	}

	tcf_exts_change(tp, &f->exts, &e);
	tcf_em_tree_change(tp, &f->ematches, &t);

	tcf_tree_lock(tp);

	if (tb[TCA_FLOW_KEYS]) {
		f->keymask = keymask;
		f->nkeys   = nkeys;
	}

	f->mode = mode;

	if (tb[TCA_FLOW_MASK])
		f->mask = nla_get_u32(tb[TCA_FLOW_MASK]);
	if (tb[TCA_FLOW_XOR])
		f->xor = nla_get_u32(tb[TCA_FLOW_XOR]);
	if (tb[TCA_FLOW_RSHIFT])
		f->rshift = nla_get_u32(tb[TCA_FLOW_RSHIFT]);
	if (tb[TCA_FLOW_ADDEND])
		f->addend = nla_get_u32(tb[TCA_FLOW_ADDEND]);

	if (tb[TCA_FLOW_DIVISOR])
		f->divisor = nla_get_u32(tb[TCA_FLOW_DIVISOR]);
	if (baseclass)
		f->baseclass = baseclass;

	f->perturb_period = perturb_period;
	del_timer(&f->perturb_timer);
	if (perturb_period)
		mod_timer(&f->perturb_timer, jiffies + perturb_period);

	if (*arg == 0)
		list_add_tail(&f->list, &head->filters);

	tcf_tree_unlock(tp);

	*arg = (unsigned long)f;
	return 0;

err2:
	tcf_em_tree_destroy(tp, &t);
err1:
	tcf_exts_destroy(tp, &e);
	return err;
}

static void flow_destroy_filter(struct tcf_proto *tp, struct flow_filter *f)
{
	del_timer_sync(&f->perturb_timer);
	tcf_exts_destroy(tp, &f->exts);
	tcf_em_tree_destroy(tp, &f->ematches);
	kfree(f);
}

static int flow_delete(struct tcf_proto *tp, unsigned long arg)
{
	struct flow_filter *f = (struct flow_filter *)arg;

	tcf_tree_lock(tp);
	list_del(&f->list);
	tcf_tree_unlock(tp);
	flow_destroy_filter(tp, f);
	return 0;
}

static int flow_init(struct tcf_proto *tp)
{
	struct flow_head *head;

	head = kzalloc(sizeof(*head), GFP_KERNEL);
	if (head == NULL)
		return -ENOBUFS;
	INIT_LIST_HEAD(&head->filters);
	tp->root = head;
	return 0;
}

static void flow_destroy(struct tcf_proto *tp)
{
	struct flow_head *head = tp->root;
	struct flow_filter *f, *next;

	list_for_each_entry_safe(f, next, &head->filters, list) {
		list_del(&f->list);
		flow_destroy_filter(tp, f);
	}
	kfree(head);
}

static unsigned long flow_get(struct tcf_proto *tp, u32 handle)
{
	struct flow_head *head = tp->root;
	struct flow_filter *f;

	list_for_each_entry(f, &head->filters, list)
		if (f->handle == handle)
			return (unsigned long)f;
	return 0;
}

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

static int flow_dump(struct tcf_proto *tp, unsigned long fh,
		     struct sk_buff *skb, struct tcmsg *t)
{
	struct flow_filter *f = (struct flow_filter *)fh;
	struct nlattr *nest;

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

	t->tcm_handle = f->handle;

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

	NLA_PUT_U32(skb, TCA_FLOW_KEYS, f->keymask);
	NLA_PUT_U32(skb, TCA_FLOW_MODE, f->mode);

	if (f->mask != ~0 || f->xor != 0) {
		NLA_PUT_U32(skb, TCA_FLOW_MASK, f->mask);
		NLA_PUT_U32(skb, TCA_FLOW_XOR, f->xor);
	}
	if (f->rshift)
		NLA_PUT_U32(skb, TCA_FLOW_RSHIFT, f->rshift);
	if (f->addend)
		NLA_PUT_U32(skb, TCA_FLOW_ADDEND, f->addend);

	if (f->divisor)
		NLA_PUT_U32(skb, TCA_FLOW_DIVISOR, f->divisor);
	if (f->baseclass)
		NLA_PUT_U32(skb, TCA_FLOW_BASECLASS, f->baseclass);

	if (f->perturb_period)
		NLA_PUT_U32(skb, TCA_FLOW_PERTURB, f->perturb_period / HZ);

	if (tcf_exts_dump(skb, &f->exts, &flow_ext_map) < 0)
		goto nla_put_failure;
#ifdef CONFIG_NET_EMATCH
	if (f->ematches.hdr.nmatches &&
	    tcf_em_tree_dump(skb, &f->ematches, TCA_FLOW_EMATCHES) < 0)
		goto nla_put_failure;
#endif
	nla_nest_end(skb, nest);

	if (tcf_exts_dump_stats(skb, &f->exts, &flow_ext_map) < 0)
		goto nla_put_failure;

	return skb->len;

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

static void flow_walk(struct tcf_proto *tp, struct tcf_walker *arg)
{
	struct flow_head *head = tp->root;
	struct flow_filter *f;

	list_for_each_entry(f, &head->filters, list) {
		if (arg->count < arg->skip)
			goto skip;
		if (arg->fn(tp, (unsigned long)f, arg) < 0) {
			arg->stop = 1;
			break;
		}
skip:
		arg->count++;
	}
}

static struct tcf_proto_ops cls_flow_ops __read_mostly = {
	.kind		= "flow",
	.classify	= flow_classify,
	.init		= flow_init,
	.destroy	= flow_destroy,
	.change		= flow_change,
	.delete		= flow_delete,
	.get		= flow_get,
	.put		= flow_put,
	.dump		= flow_dump,
	.walk		= flow_walk,
	.owner		= THIS_MODULE,
};

static int __init cls_flow_init(void)
{
	return register_tcf_proto_ops(&cls_flow_ops);
}

static void __exit cls_flow_exit(void)
{
	unregister_tcf_proto_ops(&cls_flow_ops);
}

module_init(cls_flow_init);
module_exit(cls_flow_exit);

MODULE_LICENSE("GPL");
MODULE_AUTHOR("Patrick McHardy <kaber@trash.net>");
MODULE_DESCRIPTION("TC flow classifier");
Commit	Line	Data
e5dfb815 PM	1	/*
	2	* net/sched/cls_flow.c Generic flow classifier
	3	*
	4	* Copyright (c) 2007, 2008 Patrick McHardy <kaber@trash.net>
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public License
	8	* as published by the Free Software Foundation; either version 2
	9	* of the License, or (at your option) any later version.
	10	*/
	11
	12	#include <linux/kernel.h>
	13	#include <linux/init.h>
	14	#include <linux/list.h>
	15	#include <linux/jhash.h>
	16	#include <linux/random.h>
	17	#include <linux/pkt_cls.h>
	18	#include <linux/skbuff.h>
	19	#include <linux/in.h>
	20	#include <linux/ip.h>
	21	#include <linux/ipv6.h>
9ec13810	22	#include <linux/if_vlan.h>
5a0e3ad6	23	#include <linux/slab.h>
e5dfb815 PM	24
	25	#include <net/pkt_cls.h>
	26	#include <net/ip.h>
	27	#include <net/route.h>
	28	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
	29	#include <net/netfilter/nf_conntrack.h>
	30	#endif
	31
	32	struct flow_head {
	33	struct list_head filters;
	34	};
	35
	36	struct flow_filter {
	37	struct list_head list;
	38	struct tcf_exts exts;
	39	struct tcf_ematch_tree ematches;
72d9794f PM	40	struct timer_list perturb_timer;
72d9794f PM	41	u32 perturb_period;
e5dfb815 PM	42	u32 handle;
	43
	44	u32 nkeys;
	45	u32 keymask;
	46	u32 mode;
	47	u32 mask;
	48	u32 xor;
	49	u32 rshift;
	50	u32 addend;
	51	u32 divisor;
	52	u32 baseclass;
72d9794f	53	u32 hashrnd;
e5dfb815 PM	54	};
e5dfb815 PM	55
e5dfb815 PM	56	static const struct tcf_ext_map flow_ext_map = {
	57	.action = TCA_FLOW_ACT,
	58	.police = TCA_FLOW_POLICE,
	59	};
	60
	61	static inline u32 addr_fold(void *addr)
	62	{
	63	unsigned long a = (unsigned long)addr;
	64
	65	return (a & 0xFFFFFFFF) ^ (BITS_PER_LONG > 32 ? a >> 32 : 0);
	66	}
	67
4b95c3d4	68	static u32 flow_get_src(struct sk_buff *skb)
e5dfb815 PM	69	{
e5dfb815 PM	70	switch (skb->protocol) {
60678040	71	case htons(ETH_P_IP):
4b95c3d4 CG	72	if (pskb_network_may_pull(skb, sizeof(struct iphdr)))
	73	return ntohl(ip_hdr(skb)->saddr);
	74	break;
60678040	75	case htons(ETH_P_IPV6):
4b95c3d4 CG	76	if (pskb_network_may_pull(skb, sizeof(struct ipv6hdr)))
	77	return ntohl(ipv6_hdr(skb)->saddr.s6_addr32[3]);
	78	break;
e5dfb815	79	}
4b95c3d4 CG	80
4b95c3d4 CG	81	return addr_fold(skb->sk);
e5dfb815 PM	82	}
e5dfb815 PM	83
4b95c3d4	84	static u32 flow_get_dst(struct sk_buff *skb)
e5dfb815 PM	85	{
e5dfb815 PM	86	switch (skb->protocol) {
60678040	87	case htons(ETH_P_IP):
4b95c3d4 CG	88	if (pskb_network_may_pull(skb, sizeof(struct iphdr)))
	89	return ntohl(ip_hdr(skb)->daddr);
	90	break;
60678040	91	case htons(ETH_P_IPV6):
4b95c3d4 CG	92	if (pskb_network_may_pull(skb, sizeof(struct ipv6hdr)))
	93	return ntohl(ipv6_hdr(skb)->daddr.s6_addr32[3]);
	94	break;
e5dfb815	95	}
4b95c3d4 CG	96
4b95c3d4 CG	97	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
e5dfb815 PM	98	}
e5dfb815 PM	99
4b95c3d4	100	static u32 flow_get_proto(struct sk_buff *skb)
e5dfb815 PM	101	{
e5dfb815 PM	102	switch (skb->protocol) {
60678040	103	case htons(ETH_P_IP):
4b95c3d4 CG	104	return pskb_network_may_pull(skb, sizeof(struct iphdr)) ?
4b95c3d4 CG	105	ip_hdr(skb)->protocol : 0;
60678040	106	case htons(ETH_P_IPV6):
4b95c3d4 CG	107	return pskb_network_may_pull(skb, sizeof(struct ipv6hdr)) ?
4b95c3d4 CG	108	ipv6_hdr(skb)->nexthdr : 0;
e5dfb815 PM	109	default:
	110	return 0;
	111	}
	112	}
	113
4b95c3d4	114	static u32 flow_get_proto_src(struct sk_buff *skb)
e5dfb815	115	{
e5dfb815	116	switch (skb->protocol) {
60678040	117	case htons(ETH_P_IP): {
4b95c3d4	118	struct iphdr *iph;
78d3307e	119	int poff;
e5dfb815	120
4b95c3d4 CG	121	if (!pskb_network_may_pull(skb, sizeof(*iph)))
	122	break;
	123	iph = ip_hdr(skb);
78d3307e CG	124	if (iph->frag_off & htons(IP_MF\|IP_OFFSET))
	125	break;
	126	poff = proto_ports_offset(iph->protocol);
	127	if (poff >= 0 &&
	128	pskb_network_may_pull(skb, iph->ihl * 4 + 2 + poff)) {
	129	iph = ip_hdr(skb);
	130	return ntohs((__be16 )((void )iph + iph->ihl 4 +
	131	poff));
	132	}
e5dfb815 PM	133	break;
e5dfb815 PM	134	}
60678040	135	case htons(ETH_P_IPV6): {
4b95c3d4	136	struct ipv6hdr *iph;
78d3307e	137	int poff;
e5dfb815	138
78d3307e	139	if (!pskb_network_may_pull(skb, sizeof(*iph)))
4b95c3d4 CG	140	break;
4b95c3d4 CG	141	iph = ipv6_hdr(skb);
78d3307e CG	142	poff = proto_ports_offset(iph->nexthdr);
	143	if (poff >= 0 &&
	144	pskb_network_may_pull(skb, sizeof(*iph) + poff + 2)) {
	145	iph = ipv6_hdr(skb);
	146	return ntohs((__be16 )((void )iph + sizeof(iph) +
	147	poff));
	148	}
e5dfb815 PM	149	break;
e5dfb815 PM	150	}
e5dfb815 PM	151	}
e5dfb815 PM	152
4b95c3d4	153	return addr_fold(skb->sk);
e5dfb815 PM	154	}
e5dfb815 PM	155
4b95c3d4	156	static u32 flow_get_proto_dst(struct sk_buff *skb)
e5dfb815	157	{
e5dfb815	158	switch (skb->protocol) {
60678040	159	case htons(ETH_P_IP): {
4b95c3d4	160	struct iphdr *iph;
78d3307e	161	int poff;
e5dfb815	162
4b95c3d4 CG	163	if (!pskb_network_may_pull(skb, sizeof(*iph)))
	164	break;
	165	iph = ip_hdr(skb);
78d3307e CG	166	if (iph->frag_off & htons(IP_MF\|IP_OFFSET))
	167	break;
	168	poff = proto_ports_offset(iph->protocol);
	169	if (poff >= 0 &&
	170	pskb_network_may_pull(skb, iph->ihl * 4 + 4 + poff)) {
	171	iph = ip_hdr(skb);
	172	return ntohs((__be16 )((void )iph + iph->ihl 4 +
	173	2 + poff));
	174	}
e5dfb815 PM	175	break;
e5dfb815 PM	176	}
60678040	177	case htons(ETH_P_IPV6): {
4b95c3d4	178	struct ipv6hdr *iph;
78d3307e	179	int poff;
e5dfb815	180
78d3307e	181	if (!pskb_network_may_pull(skb, sizeof(*iph)))
4b95c3d4 CG	182	break;
4b95c3d4 CG	183	iph = ipv6_hdr(skb);
78d3307e CG	184	poff = proto_ports_offset(iph->nexthdr);
	185	if (poff >= 0 &&
	186	pskb_network_may_pull(skb, sizeof(*iph) + poff + 4)) {
	187	iph = ipv6_hdr(skb);
	188	return ntohs((__be16 )((void )iph + sizeof(iph) +
	189	poff + 2));
	190	}
e5dfb815 PM	191	break;
e5dfb815 PM	192	}
e5dfb815 PM	193	}
e5dfb815 PM	194
4b95c3d4	195	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
e5dfb815 PM	196	}
	197
	198	static u32 flow_get_iif(const struct sk_buff *skb)
	199	{
8964be4a	200	return skb->skb_iif;
e5dfb815 PM	201	}
	202
	203	static u32 flow_get_priority(const struct sk_buff *skb)
	204	{
	205	return skb->priority;
	206	}
	207
	208	static u32 flow_get_mark(const struct sk_buff *skb)
	209	{
	210	return skb->mark;
	211	}
	212
	213	static u32 flow_get_nfct(const struct sk_buff *skb)
	214	{
	215	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
	216	return addr_fold(skb->nfct);
	217	#else
	218	return 0;
	219	#endif
	220	}
	221
	222	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
	223	#define CTTUPLE(skb, member) \
	224	({ \
	225	enum ip_conntrack_info ctinfo; \
	226	struct nf_conn *ct = nf_ct_get(skb, &ctinfo); \
	227	if (ct == NULL) \
	228	goto fallback; \
	229	ct->tuplehash[CTINFO2DIR(ctinfo)].tuple.member; \
	230	})
	231	#else
	232	#define CTTUPLE(skb, member) \
	233	({ \
	234	goto fallback; \
	235	0; \
	236	})
	237	#endif
	238
4b95c3d4	239	static u32 flow_get_nfct_src(struct sk_buff *skb)
e5dfb815 PM	240	{
e5dfb815 PM	241	switch (skb->protocol) {
60678040	242	case htons(ETH_P_IP):
e5dfb815	243	return ntohl(CTTUPLE(skb, src.u3.ip));
60678040	244	case htons(ETH_P_IPV6):
e5dfb815 PM	245	return ntohl(CTTUPLE(skb, src.u3.ip6[3]));
	246	}
	247	fallback:
	248	return flow_get_src(skb);
	249	}
	250
4b95c3d4	251	static u32 flow_get_nfct_dst(struct sk_buff *skb)
e5dfb815 PM	252	{
e5dfb815 PM	253	switch (skb->protocol) {
60678040	254	case htons(ETH_P_IP):
e5dfb815	255	return ntohl(CTTUPLE(skb, dst.u3.ip));
60678040	256	case htons(ETH_P_IPV6):
e5dfb815 PM	257	return ntohl(CTTUPLE(skb, dst.u3.ip6[3]));
	258	}
	259	fallback:
	260	return flow_get_dst(skb);
	261	}
	262
4b95c3d4	263	static u32 flow_get_nfct_proto_src(struct sk_buff *skb)
e5dfb815 PM	264	{
	265	return ntohs(CTTUPLE(skb, src.u.all));
	266	fallback:
	267	return flow_get_proto_src(skb);
	268	}
	269
4b95c3d4	270	static u32 flow_get_nfct_proto_dst(struct sk_buff *skb)
e5dfb815 PM	271	{
	272	return ntohs(CTTUPLE(skb, dst.u.all));
	273	fallback:
	274	return flow_get_proto_dst(skb);
	275	}
	276
	277	static u32 flow_get_rtclassid(const struct sk_buff *skb)
	278	{
	279	#ifdef CONFIG_NET_CLS_ROUTE
adf30907 ED	280	if (skb_dst(skb))
adf30907 ED	281	return skb_dst(skb)->tclassid;
e5dfb815 PM	282	#endif
	283	return 0;
	284	}
	285
	286	static u32 flow_get_skuid(const struct sk_buff *skb)
	287	{
	288	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
d76b0d9b	289	return skb->sk->sk_socket->file->f_cred->fsuid;
e5dfb815 PM	290	return 0;
	291	}
	292
	293	static u32 flow_get_skgid(const struct sk_buff *skb)
	294	{
	295	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
d76b0d9b	296	return skb->sk->sk_socket->file->f_cred->fsgid;
e5dfb815 PM	297	return 0;
	298	}
	299
9ec13810 PM	300	static u32 flow_get_vlan_tag(const struct sk_buff *skb)
	301	{
	302	u16 uninitialized_var(tag);
	303
	304	if (vlan_get_tag(skb, &tag) < 0)
	305	return 0;
	306	return tag & VLAN_VID_MASK;
	307	}
	308
739a91ef CG	309	static u32 flow_get_rxhash(struct sk_buff *skb)
	310	{
	311	return skb_get_rxhash(skb);
	312	}
	313
4b95c3d4	314	static u32 flow_key_get(struct sk_buff *skb, int key)
e5dfb815 PM	315	{
	316	switch (key) {
	317	case FLOW_KEY_SRC:
	318	return flow_get_src(skb);
	319	case FLOW_KEY_DST:
	320	return flow_get_dst(skb);
	321	case FLOW_KEY_PROTO:
	322	return flow_get_proto(skb);
	323	case FLOW_KEY_PROTO_SRC:
	324	return flow_get_proto_src(skb);
	325	case FLOW_KEY_PROTO_DST:
	326	return flow_get_proto_dst(skb);
	327	case FLOW_KEY_IIF:
	328	return flow_get_iif(skb);
	329	case FLOW_KEY_PRIORITY:
	330	return flow_get_priority(skb);
	331	case FLOW_KEY_MARK:
	332	return flow_get_mark(skb);
	333	case FLOW_KEY_NFCT:
	334	return flow_get_nfct(skb);
	335	case FLOW_KEY_NFCT_SRC:
	336	return flow_get_nfct_src(skb);
	337	case FLOW_KEY_NFCT_DST:
	338	return flow_get_nfct_dst(skb);
	339	case FLOW_KEY_NFCT_PROTO_SRC:
	340	return flow_get_nfct_proto_src(skb);
	341	case FLOW_KEY_NFCT_PROTO_DST:
	342	return flow_get_nfct_proto_dst(skb);
	343	case FLOW_KEY_RTCLASSID:
	344	return flow_get_rtclassid(skb);
	345	case FLOW_KEY_SKUID:
	346	return flow_get_skuid(skb);
	347	case FLOW_KEY_SKGID:
	348	return flow_get_skgid(skb);
9ec13810 PM	349	case FLOW_KEY_VLAN_TAG:
9ec13810 PM	350	return flow_get_vlan_tag(skb);
739a91ef CG	351	case FLOW_KEY_RXHASH:
739a91ef CG	352	return flow_get_rxhash(skb);
e5dfb815 PM	353	default:
	354	WARN_ON(1);
	355	return 0;
	356	}
	357	}
	358
	359	static int flow_classify(struct sk_buff skb, struct tcf_proto tp,
	360	struct tcf_result *res)
	361	{
	362	struct flow_head *head = tp->root;
	363	struct flow_filter *f;
	364	u32 keymask;
	365	u32 classid;
	366	unsigned int n, key;
	367	int r;
	368
	369	list_for_each_entry(f, &head->filters, list) {
	370	u32 keys[f->nkeys];
	371
	372	if (!tcf_em_tree_match(skb, &f->ematches, NULL))
	373	continue;
	374
	375	keymask = f->keymask;
	376
	377	for (n = 0; n < f->nkeys; n++) {
	378	key = ffs(keymask) - 1;
	379	keymask &= ~(1 << key);
	380	keys[n] = flow_key_get(skb, key);
	381	}
	382
	383	if (f->mode == FLOW_MODE_HASH)
72d9794f	384	classid = jhash2(keys, f->nkeys, f->hashrnd);
e5dfb815 PM	385	else {
	386	classid = keys[0];
	387	classid = (classid & f->mask) ^ f->xor;
	388	classid = (classid >> f->rshift) + f->addend;
	389	}
	390
	391	if (f->divisor)
	392	classid %= f->divisor;
	393
	394	res->class = 0;
	395	res->classid = TC_H_MAKE(f->baseclass, f->baseclass + classid);
	396
	397	r = tcf_exts_exec(skb, &f->exts, res);
	398	if (r < 0)
	399	continue;
	400	return r;
	401	}
	402	return -1;
	403	}
	404
72d9794f PM	405	static void flow_perturbation(unsigned long arg)
	406	{
	407	struct flow_filter f = (struct flow_filter )arg;
	408
	409	get_random_bytes(&f->hashrnd, 4);
	410	if (f->perturb_period)
	411	mod_timer(&f->perturb_timer, jiffies + f->perturb_period);
	412	}
	413
e5dfb815 PM	414	static const struct nla_policy flow_policy[TCA_FLOW_MAX + 1] = {
	415	[TCA_FLOW_KEYS] = { .type = NLA_U32 },
	416	[TCA_FLOW_MODE] = { .type = NLA_U32 },
	417	[TCA_FLOW_BASECLASS] = { .type = NLA_U32 },
	418	[TCA_FLOW_RSHIFT] = { .type = NLA_U32 },
	419	[TCA_FLOW_ADDEND] = { .type = NLA_U32 },
	420	[TCA_FLOW_MASK] = { .type = NLA_U32 },
	421	[TCA_FLOW_XOR] = { .type = NLA_U32 },
	422	[TCA_FLOW_DIVISOR] = { .type = NLA_U32 },
	423	[TCA_FLOW_ACT] = { .type = NLA_NESTED },
	424	[TCA_FLOW_POLICE] = { .type = NLA_NESTED },
	425	[TCA_FLOW_EMATCHES] = { .type = NLA_NESTED },
72d9794f	426	[TCA_FLOW_PERTURB] = { .type = NLA_U32 },
e5dfb815 PM	427	};
	428
	429	static int flow_change(struct tcf_proto *tp, unsigned long base,
	430	u32 handle, struct nlattr **tca,
	431	unsigned long *arg)
	432	{
	433	struct flow_head *head = tp->root;
	434	struct flow_filter *f;
	435	struct nlattr *opt = tca[TCA_OPTIONS];
	436	struct nlattr *tb[TCA_FLOW_MAX + 1];
	437	struct tcf_exts e;
	438	struct tcf_ematch_tree t;
	439	unsigned int nkeys = 0;
72d9794f	440	unsigned int perturb_period = 0;
e5dfb815 PM	441	u32 baseclass = 0;
	442	u32 keymask = 0;
	443	u32 mode;
	444	int err;
	445
	446	if (opt == NULL)
	447	return -EINVAL;
	448
	449	err = nla_parse_nested(tb, TCA_FLOW_MAX, opt, flow_policy);
	450	if (err < 0)
	451	return err;
	452
	453	if (tb[TCA_FLOW_BASECLASS]) {
	454	baseclass = nla_get_u32(tb[TCA_FLOW_BASECLASS]);
	455	if (TC_H_MIN(baseclass) == 0)
	456	return -EINVAL;
	457	}
	458
	459	if (tb[TCA_FLOW_KEYS]) {
	460	keymask = nla_get_u32(tb[TCA_FLOW_KEYS]);
e5dfb815 PM	461
	462	nkeys = hweight32(keymask);
	463	if (nkeys == 0)
	464	return -EINVAL;
4f250491 PM	465
	466	if (fls(keymask) - 1 > FLOW_KEY_MAX)
	467	return -EOPNOTSUPP;
e5dfb815 PM	468	}
	469
	470	err = tcf_exts_validate(tp, tb, tca[TCA_RATE], &e, &flow_ext_map);
	471	if (err < 0)
	472	return err;
	473
	474	err = tcf_em_tree_validate(tp, tb[TCA_FLOW_EMATCHES], &t);
	475	if (err < 0)
	476	goto err1;
	477
	478	f = (struct flow_filter )arg;
	479	if (f != NULL) {
	480	err = -EINVAL;
	481	if (f->handle != handle && handle)
	482	goto err2;
	483
	484	mode = f->mode;
	485	if (tb[TCA_FLOW_MODE])
	486	mode = nla_get_u32(tb[TCA_FLOW_MODE]);
	487	if (mode != FLOW_MODE_HASH && nkeys > 1)
	488	goto err2;
72d9794f PM	489
	490	if (mode == FLOW_MODE_HASH)
	491	perturb_period = f->perturb_period;
	492	if (tb[TCA_FLOW_PERTURB]) {
	493	if (mode != FLOW_MODE_HASH)
	494	goto err2;
	495	perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
	496	}
e5dfb815 PM	497	} else {
	498	err = -EINVAL;
	499	if (!handle)
	500	goto err2;
	501	if (!tb[TCA_FLOW_KEYS])
	502	goto err2;
	503
	504	mode = FLOW_MODE_MAP;
	505	if (tb[TCA_FLOW_MODE])
	506	mode = nla_get_u32(tb[TCA_FLOW_MODE]);
	507	if (mode != FLOW_MODE_HASH && nkeys > 1)
	508	goto err2;
	509
72d9794f PM	510	if (tb[TCA_FLOW_PERTURB]) {
	511	if (mode != FLOW_MODE_HASH)
	512	goto err2;
	513	perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
	514	}
	515
e5dfb815 PM	516	if (TC_H_MAJ(baseclass) == 0)
	517	baseclass = TC_H_MAKE(tp->q->handle, baseclass);
	518	if (TC_H_MIN(baseclass) == 0)
	519	baseclass = TC_H_MAKE(baseclass, 1);
	520
	521	err = -ENOBUFS;
	522	f = kzalloc(sizeof(*f), GFP_KERNEL);
	523	if (f == NULL)
	524	goto err2;
	525
	526	f->handle = handle;
	527	f->mask = ~0U;
72d9794f PM	528
	529	get_random_bytes(&f->hashrnd, 4);
	530	f->perturb_timer.function = flow_perturbation;
	531	f->perturb_timer.data = (unsigned long)f;
	532	init_timer_deferrable(&f->perturb_timer);
e5dfb815 PM	533	}
	534
	535	tcf_exts_change(tp, &f->exts, &e);
	536	tcf_em_tree_change(tp, &f->ematches, &t);
	537
	538	tcf_tree_lock(tp);
	539
	540	if (tb[TCA_FLOW_KEYS]) {
	541	f->keymask = keymask;
	542	f->nkeys = nkeys;
	543	}
	544
	545	f->mode = mode;
	546
	547	if (tb[TCA_FLOW_MASK])
	548	f->mask = nla_get_u32(tb[TCA_FLOW_MASK]);
	549	if (tb[TCA_FLOW_XOR])
	550	f->xor = nla_get_u32(tb[TCA_FLOW_XOR]);
	551	if (tb[TCA_FLOW_RSHIFT])
	552	f->rshift = nla_get_u32(tb[TCA_FLOW_RSHIFT]);
	553	if (tb[TCA_FLOW_ADDEND])
	554	f->addend = nla_get_u32(tb[TCA_FLOW_ADDEND]);
	555
	556	if (tb[TCA_FLOW_DIVISOR])
	557	f->divisor = nla_get_u32(tb[TCA_FLOW_DIVISOR]);
	558	if (baseclass)
	559	f->baseclass = baseclass;
	560
72d9794f PM	561	f->perturb_period = perturb_period;
	562	del_timer(&f->perturb_timer);
	563	if (perturb_period)
	564	mod_timer(&f->perturb_timer, jiffies + perturb_period);
	565
e5dfb815 PM	566	if (*arg == 0)
	567	list_add_tail(&f->list, &head->filters);
	568
	569	tcf_tree_unlock(tp);
	570
	571	*arg = (unsigned long)f;
	572	return 0;
	573
	574	err2:
	575	tcf_em_tree_destroy(tp, &t);
	576	err1:
	577	tcf_exts_destroy(tp, &e);
	578	return err;
	579	}
	580
	581	static void flow_destroy_filter(struct tcf_proto tp, struct flow_filter f)
	582	{
72d9794f	583	del_timer_sync(&f->perturb_timer);
e5dfb815 PM	584	tcf_exts_destroy(tp, &f->exts);
	585	tcf_em_tree_destroy(tp, &f->ematches);
	586	kfree(f);
	587	}
	588
	589	static int flow_delete(struct tcf_proto *tp, unsigned long arg)
	590	{
	591	struct flow_filter f = (struct flow_filter )arg;
	592
	593	tcf_tree_lock(tp);
	594	list_del(&f->list);
	595	tcf_tree_unlock(tp);
	596	flow_destroy_filter(tp, f);
	597	return 0;
	598	}
	599
	600	static int flow_init(struct tcf_proto *tp)
	601	{
	602	struct flow_head *head;
	603
e5dfb815 PM	604	head = kzalloc(sizeof(*head), GFP_KERNEL);
	605	if (head == NULL)
	606	return -ENOBUFS;
	607	INIT_LIST_HEAD(&head->filters);
	608	tp->root = head;
	609	return 0;
	610	}
	611
	612	static void flow_destroy(struct tcf_proto *tp)
	613	{
	614	struct flow_head *head = tp->root;
	615	struct flow_filter f, next;
	616
	617	list_for_each_entry_safe(f, next, &head->filters, list) {
	618	list_del(&f->list);
	619	flow_destroy_filter(tp, f);
	620	}
	621	kfree(head);
	622	}
	623
	624	static unsigned long flow_get(struct tcf_proto *tp, u32 handle)
	625	{
	626	struct flow_head *head = tp->root;
	627	struct flow_filter *f;
	628
	629	list_for_each_entry(f, &head->filters, list)
	630	if (f->handle == handle)
	631	return (unsigned long)f;
	632	return 0;
	633	}
	634
	635	static void flow_put(struct tcf_proto *tp, unsigned long f)
	636	{
e5dfb815 PM	637	}
	638
	639	static int flow_dump(struct tcf_proto *tp, unsigned long fh,
	640	struct sk_buff skb, struct tcmsg t)
	641	{
	642	struct flow_filter f = (struct flow_filter )fh;
	643	struct nlattr *nest;
	644
	645	if (f == NULL)
	646	return skb->len;
	647
	648	t->tcm_handle = f->handle;
	649
	650	nest = nla_nest_start(skb, TCA_OPTIONS);
	651	if (nest == NULL)
	652	goto nla_put_failure;
	653
	654	NLA_PUT_U32(skb, TCA_FLOW_KEYS, f->keymask);
	655	NLA_PUT_U32(skb, TCA_FLOW_MODE, f->mode);
	656
	657	if (f->mask != ~0 \|\| f->xor != 0) {
	658	NLA_PUT_U32(skb, TCA_FLOW_MASK, f->mask);
	659	NLA_PUT_U32(skb, TCA_FLOW_XOR, f->xor);
	660	}
	661	if (f->rshift)
	662	NLA_PUT_U32(skb, TCA_FLOW_RSHIFT, f->rshift);
	663	if (f->addend)
	664	NLA_PUT_U32(skb, TCA_FLOW_ADDEND, f->addend);
	665
	666	if (f->divisor)
	667	NLA_PUT_U32(skb, TCA_FLOW_DIVISOR, f->divisor);
	668	if (f->baseclass)
	669	NLA_PUT_U32(skb, TCA_FLOW_BASECLASS, f->baseclass);
	670
72d9794f PM	671	if (f->perturb_period)
	672	NLA_PUT_U32(skb, TCA_FLOW_PERTURB, f->perturb_period / HZ);
	673
e5dfb815 PM	674	if (tcf_exts_dump(skb, &f->exts, &flow_ext_map) < 0)
e5dfb815 PM	675	goto nla_put_failure;
0aead543	676	#ifdef CONFIG_NET_EMATCH
e5dfb815 PM	677	if (f->ematches.hdr.nmatches &&
	678	tcf_em_tree_dump(skb, &f->ematches, TCA_FLOW_EMATCHES) < 0)
	679	goto nla_put_failure;
0aead543	680	#endif
e5dfb815 PM	681	nla_nest_end(skb, nest);
	682
	683	if (tcf_exts_dump_stats(skb, &f->exts, &flow_ext_map) < 0)
	684	goto nla_put_failure;
	685
	686	return skb->len;
	687
	688	nla_put_failure:
	689	nlmsg_trim(skb, nest);
	690	return -1;
	691	}
	692
	693	static void flow_walk(struct tcf_proto tp, struct tcf_walker arg)
	694	{
	695	struct flow_head *head = tp->root;
	696	struct flow_filter *f;
	697
	698	list_for_each_entry(f, &head->filters, list) {
	699	if (arg->count < arg->skip)
	700	goto skip;
	701	if (arg->fn(tp, (unsigned long)f, arg) < 0) {
	702	arg->stop = 1;
	703	break;
	704	}
	705	skip:
	706	arg->count++;
	707	}
	708	}
	709
	710	static struct tcf_proto_ops cls_flow_ops __read_mostly = {
	711	.kind = "flow",
	712	.classify = flow_classify,
	713	.init = flow_init,
	714	.destroy = flow_destroy,
	715	.change = flow_change,
	716	.delete = flow_delete,
	717	.get = flow_get,
	718	.put = flow_put,
	719	.dump = flow_dump,
	720	.walk = flow_walk,
	721	.owner = THIS_MODULE,
	722	};
	723
	724	static int __init cls_flow_init(void)
	725	{
	726	return register_tcf_proto_ops(&cls_flow_ops);
	727	}
	728
	729	static void __exit cls_flow_exit(void)
	730	{
	731	unregister_tcf_proto_ops(&cls_flow_ops);
	732	}
	733
	734	module_init(cls_flow_init);
	735	module_exit(cls_flow_exit);
	736
	737	MODULE_LICENSE("GPL");
	738	MODULE_AUTHOR("Patrick McHardy <kaber@trash.net>");
	739	MODULE_DESCRIPTION("TC flow classifier");