[mirror_ubuntu-hirsute-kernel.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(const struct sk_buff *skb, int nhoff)
{
	__be32 *data = NULL, hdata;

	switch (skb->protocol) {
	case htons(ETH_P_IP):
		data = skb_header_pointer(skb,
					  nhoff + offsetof(struct iphdr,
							   saddr),
					  4, &hdata);
		break;
	case htons(ETH_P_IPV6):
		data = skb_header_pointer(skb,
					 nhoff + offsetof(struct ipv6hdr,
							  saddr.s6_addr32[3]),
					 4, &hdata);
		break;
	}

	if (data)
		return ntohl(*data);
	return addr_fold(skb->sk);
}

static u32 flow_get_dst(const struct sk_buff *skb, int nhoff)
{
	__be32 *data = NULL, hdata;

	switch (skb->protocol) {
	case htons(ETH_P_IP):
		data = skb_header_pointer(skb,
					  nhoff + offsetof(struct iphdr,
							   daddr),
					  4, &hdata);
		break;
	case htons(ETH_P_IPV6):
		data = skb_header_pointer(skb,
					 nhoff + offsetof(struct ipv6hdr,
							  daddr.s6_addr32[3]),
					 4, &hdata);
		break;
	}

	if (data)
		return ntohl(*data);
	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
}

static u32 flow_get_proto(const struct sk_buff *skb, int nhoff)
{
	__u8 *data = NULL, hdata;

	switch (skb->protocol) {
	case htons(ETH_P_IP):
		data = skb_header_pointer(skb,
					  nhoff + offsetof(struct iphdr,
							   protocol),
					  1, &hdata);
		break;
	case htons(ETH_P_IPV6):
		data = skb_header_pointer(skb,
					 nhoff + offsetof(struct ipv6hdr,
							  nexthdr),
					 1, &hdata);
		break;
	}
	if (data)
		return *data;
	return 0;
}

/* helper function to get either src or dst port */
static __be16 *flow_get_proto_common(const struct sk_buff *skb, int nhoff,
				     __be16 *_port, int dst)
{
	__be16 *port = NULL;
	int poff;

	switch (skb->protocol) {
	case htons(ETH_P_IP): {
		struct iphdr *iph, _iph;

		iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph);
		if (!iph)
			break;
		if (ip_is_fragment(iph))
			break;
		poff = proto_ports_offset(iph->protocol);
		if (poff >= 0)
			port = skb_header_pointer(skb,
					nhoff + iph->ihl * 4 + poff + dst,
					sizeof(*_port), _port);
		break;
	}
	case htons(ETH_P_IPV6): {
		struct ipv6hdr *iph, _iph;

		iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph);
		if (!iph)
			break;
		poff = proto_ports_offset(iph->nexthdr);
		if (poff >= 0)
			port = skb_header_pointer(skb,
					nhoff + sizeof(*iph) + poff + dst,
					sizeof(*_port), _port);
		break;
	}
	}

	return port;
}

static u32 flow_get_proto_src(const struct sk_buff *skb, int nhoff)
{
	__be16 _port, *port = flow_get_proto_common(skb, nhoff, &_port, 0);

	if (port)
		return ntohs(*port);

	return addr_fold(skb->sk);
}

static u32 flow_get_proto_dst(const struct sk_buff *skb, int nhoff)
{
	__be16 _port, *port = flow_get_proto_common(skb, nhoff, &_port, 2);

	if (port)
		return ntohs(*port);

	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;					\
	const 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(const struct sk_buff *skb, int nhoff)
{
	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, nhoff);
}

static u32 flow_get_nfct_dst(const struct sk_buff *skb, int nhoff)
{
	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, nhoff);
}

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

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

static u32 flow_get_rtclassid(const struct sk_buff *skb)
{
#ifdef CONFIG_IP_ROUTE_CLASSID
	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)
{
	int nhoff = skb_network_offset(skb);

	switch (key) {
	case FLOW_KEY_SRC:
		return flow_get_src(skb, nhoff);
	case FLOW_KEY_DST:
		return flow_get_dst(skb, nhoff);
	case FLOW_KEY_PROTO:
		return flow_get_proto(skb, nhoff);
	case FLOW_KEY_PROTO_SRC:
		return flow_get_proto_src(skb, nhoff);
	case FLOW_KEY_PROTO_DST:
		return flow_get_proto_dst(skb, nhoff);
	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, nhoff);
	case FLOW_KEY_NFCT_DST:
		return flow_get_nfct_dst(skb, nhoff);
	case FLOW_KEY_NFCT_PROTO_SRC:
		return flow_get_nfct_proto_src(skb, nhoff);
	case FLOW_KEY_NFCT_PROTO_DST:
		return flow_get_nfct_proto_dst(skb, nhoff);
	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, const 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
859c2012	68	static u32 flow_get_src(const struct sk_buff *skb, int nhoff)
e5dfb815	69	{
859c2012 ED	70	__be32 *data = NULL, hdata;
859c2012 ED	71
e5dfb815	72	switch (skb->protocol) {
60678040	73	case htons(ETH_P_IP):
859c2012 ED	74	data = skb_header_pointer(skb,
	75	nhoff + offsetof(struct iphdr,
	76	saddr),
	77	4, &hdata);
4b95c3d4	78	break;
60678040	79	case htons(ETH_P_IPV6):
859c2012 ED	80	data = skb_header_pointer(skb,
	81	nhoff + offsetof(struct ipv6hdr,
	82	saddr.s6_addr32[3]),
	83	4, &hdata);
4b95c3d4	84	break;
e5dfb815	85	}
4b95c3d4	86
859c2012 ED	87	if (data)
859c2012 ED	88	return ntohl(*data);
4b95c3d4	89	return addr_fold(skb->sk);
e5dfb815 PM	90	}
e5dfb815 PM	91
859c2012	92	static u32 flow_get_dst(const struct sk_buff *skb, int nhoff)
e5dfb815	93	{
859c2012 ED	94	__be32 *data = NULL, hdata;
859c2012 ED	95
e5dfb815	96	switch (skb->protocol) {
60678040	97	case htons(ETH_P_IP):
859c2012 ED	98	data = skb_header_pointer(skb,
	99	nhoff + offsetof(struct iphdr,
	100	daddr),
	101	4, &hdata);
4b95c3d4	102	break;
60678040	103	case htons(ETH_P_IPV6):
859c2012 ED	104	data = skb_header_pointer(skb,
	105	nhoff + offsetof(struct ipv6hdr,
	106	daddr.s6_addr32[3]),
	107	4, &hdata);
4b95c3d4	108	break;
e5dfb815	109	}
4b95c3d4	110
859c2012 ED	111	if (data)
859c2012 ED	112	return ntohl(*data);
4b95c3d4	113	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
e5dfb815 PM	114	}
e5dfb815 PM	115
859c2012	116	static u32 flow_get_proto(const struct sk_buff *skb, int nhoff)
e5dfb815	117	{
859c2012 ED	118	__u8 *data = NULL, hdata;
859c2012 ED	119
e5dfb815	120	switch (skb->protocol) {
60678040	121	case htons(ETH_P_IP):
859c2012 ED	122	data = skb_header_pointer(skb,
	123	nhoff + offsetof(struct iphdr,
	124	protocol),
	125	1, &hdata);
	126	break;
60678040	127	case htons(ETH_P_IPV6):
859c2012 ED	128	data = skb_header_pointer(skb,
	129	nhoff + offsetof(struct ipv6hdr,
	130	nexthdr),
	131	1, &hdata);
	132	break;
e5dfb815	133	}
859c2012 ED	134	if (data)
	135	return *data;
	136	return 0;
e5dfb815 PM	137	}
e5dfb815 PM	138
859c2012 ED	139	/* helper function to get either src or dst port */
	140	static __be16 flow_get_proto_common(const struct sk_buff skb, int nhoff,
	141	__be16 *_port, int dst)
e5dfb815	142	{
859c2012 ED	143	__be16 *port = NULL;
	144	int poff;
	145
e5dfb815	146	switch (skb->protocol) {
60678040	147	case htons(ETH_P_IP): {
859c2012	148	struct iphdr *iph, _iph;
e5dfb815	149
859c2012 ED	150	iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph);
859c2012 ED	151	if (!iph)
4b95c3d4	152	break;
56f8a75c	153	if (ip_is_fragment(iph))
78d3307e CG	154	break;
78d3307e CG	155	poff = proto_ports_offset(iph->protocol);
859c2012 ED	156	if (poff >= 0)
	157	port = skb_header_pointer(skb,
	158	nhoff + iph->ihl * 4 + poff + dst,
	159	sizeof(*_port), _port);
e5dfb815 PM	160	break;
e5dfb815 PM	161	}
60678040	162	case htons(ETH_P_IPV6): {
859c2012	163	struct ipv6hdr *iph, _iph;
e5dfb815	164
859c2012 ED	165	iph = skb_header_pointer(skb, nhoff, sizeof(_iph), &_iph);
859c2012 ED	166	if (!iph)
4b95c3d4	167	break;
78d3307e	168	poff = proto_ports_offset(iph->nexthdr);
859c2012 ED	169	if (poff >= 0)
	170	port = skb_header_pointer(skb,
	171	nhoff + sizeof(*iph) + poff + dst,
	172	sizeof(*_port), _port);
e5dfb815 PM	173	break;
e5dfb815 PM	174	}
e5dfb815 PM	175	}
e5dfb815 PM	176
859c2012	177	return port;
e5dfb815 PM	178	}
e5dfb815 PM	179
859c2012	180	static u32 flow_get_proto_src(const struct sk_buff *skb, int nhoff)
e5dfb815	181	{
859c2012	182	__be16 _port, *port = flow_get_proto_common(skb, nhoff, &_port, 0);
e5dfb815	183
859c2012 ED	184	if (port)
859c2012 ED	185	return ntohs(*port);
e5dfb815	186
859c2012 ED	187	return addr_fold(skb->sk);
	188	}
	189
	190	static u32 flow_get_proto_dst(const struct sk_buff *skb, int nhoff)
	191	{
	192	__be16 _port, *port = flow_get_proto_common(skb, nhoff, &_port, 2);
	193
	194	if (port)
	195	return ntohs(*port);
e5dfb815	196
4b95c3d4	197	return addr_fold(skb_dst(skb)) ^ (__force u16)skb->protocol;
e5dfb815 PM	198	}
	199
	200	static u32 flow_get_iif(const struct sk_buff *skb)
	201	{
8964be4a	202	return skb->skb_iif;
e5dfb815 PM	203	}
	204
	205	static u32 flow_get_priority(const struct sk_buff *skb)
	206	{
	207	return skb->priority;
	208	}
	209
	210	static u32 flow_get_mark(const struct sk_buff *skb)
	211	{
	212	return skb->mark;
	213	}
	214
	215	static u32 flow_get_nfct(const struct sk_buff *skb)
	216	{
	217	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
	218	return addr_fold(skb->nfct);
	219	#else
	220	return 0;
	221	#endif
	222	}
	223
	224	#if defined(CONFIG_NF_CONNTRACK) \|\| defined(CONFIG_NF_CONNTRACK_MODULE)
	225	#define CTTUPLE(skb, member) \
	226	({ \
	227	enum ip_conntrack_info ctinfo; \
859c2012	228	const struct nf_conn *ct = nf_ct_get(skb, &ctinfo); \
e5dfb815 PM	229	if (ct == NULL) \
	230	goto fallback; \
	231	ct->tuplehash[CTINFO2DIR(ctinfo)].tuple.member; \
	232	})
	233	#else
	234	#define CTTUPLE(skb, member) \
	235	({ \
	236	goto fallback; \
	237	0; \
	238	})
	239	#endif
	240
859c2012	241	static u32 flow_get_nfct_src(const struct sk_buff *skb, int nhoff)
e5dfb815 PM	242	{
e5dfb815 PM	243	switch (skb->protocol) {
60678040	244	case htons(ETH_P_IP):
e5dfb815	245	return ntohl(CTTUPLE(skb, src.u3.ip));
60678040	246	case htons(ETH_P_IPV6):
e5dfb815 PM	247	return ntohl(CTTUPLE(skb, src.u3.ip6[3]));
	248	}
	249	fallback:
859c2012	250	return flow_get_src(skb, nhoff);
e5dfb815 PM	251	}
e5dfb815 PM	252
859c2012	253	static u32 flow_get_nfct_dst(const struct sk_buff *skb, int nhoff)
e5dfb815 PM	254	{
e5dfb815 PM	255	switch (skb->protocol) {
60678040	256	case htons(ETH_P_IP):
e5dfb815	257	return ntohl(CTTUPLE(skb, dst.u3.ip));
60678040	258	case htons(ETH_P_IPV6):
e5dfb815 PM	259	return ntohl(CTTUPLE(skb, dst.u3.ip6[3]));
	260	}
	261	fallback:
859c2012	262	return flow_get_dst(skb, nhoff);
e5dfb815 PM	263	}
e5dfb815 PM	264
859c2012	265	static u32 flow_get_nfct_proto_src(const struct sk_buff *skb, int nhoff)
e5dfb815 PM	266	{
	267	return ntohs(CTTUPLE(skb, src.u.all));
	268	fallback:
859c2012	269	return flow_get_proto_src(skb, nhoff);
e5dfb815 PM	270	}
e5dfb815 PM	271
859c2012	272	static u32 flow_get_nfct_proto_dst(const struct sk_buff *skb, int nhoff)
e5dfb815 PM	273	{
	274	return ntohs(CTTUPLE(skb, dst.u.all));
	275	fallback:
859c2012	276	return flow_get_proto_dst(skb, nhoff);
e5dfb815 PM	277	}
	278
	279	static u32 flow_get_rtclassid(const struct sk_buff *skb)
	280	{
c7066f70	281	#ifdef CONFIG_IP_ROUTE_CLASSID
adf30907 ED	282	if (skb_dst(skb))
adf30907 ED	283	return skb_dst(skb)->tclassid;
e5dfb815 PM	284	#endif
	285	return 0;
	286	}
	287
	288	static u32 flow_get_skuid(const struct sk_buff *skb)
	289	{
	290	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
d76b0d9b	291	return skb->sk->sk_socket->file->f_cred->fsuid;
e5dfb815 PM	292	return 0;
	293	}
	294
	295	static u32 flow_get_skgid(const struct sk_buff *skb)
	296	{
	297	if (skb->sk && skb->sk->sk_socket && skb->sk->sk_socket->file)
d76b0d9b	298	return skb->sk->sk_socket->file->f_cred->fsgid;
e5dfb815 PM	299	return 0;
	300	}
	301
9ec13810 PM	302	static u32 flow_get_vlan_tag(const struct sk_buff *skb)
	303	{
	304	u16 uninitialized_var(tag);
	305
	306	if (vlan_get_tag(skb, &tag) < 0)
	307	return 0;
	308	return tag & VLAN_VID_MASK;
	309	}
	310
739a91ef CG	311	static u32 flow_get_rxhash(struct sk_buff *skb)
	312	{
	313	return skb_get_rxhash(skb);
	314	}
	315
4b95c3d4	316	static u32 flow_key_get(struct sk_buff *skb, int key)
e5dfb815	317	{
859c2012 ED	318	int nhoff = skb_network_offset(skb);
859c2012 ED	319
e5dfb815 PM	320	switch (key) {
e5dfb815 PM	321	case FLOW_KEY_SRC:
859c2012	322	return flow_get_src(skb, nhoff);
e5dfb815	323	case FLOW_KEY_DST:
859c2012	324	return flow_get_dst(skb, nhoff);
e5dfb815	325	case FLOW_KEY_PROTO:
859c2012	326	return flow_get_proto(skb, nhoff);
e5dfb815	327	case FLOW_KEY_PROTO_SRC:
859c2012	328	return flow_get_proto_src(skb, nhoff);
e5dfb815	329	case FLOW_KEY_PROTO_DST:
859c2012	330	return flow_get_proto_dst(skb, nhoff);
e5dfb815 PM	331	case FLOW_KEY_IIF:
	332	return flow_get_iif(skb);
	333	case FLOW_KEY_PRIORITY:
	334	return flow_get_priority(skb);
	335	case FLOW_KEY_MARK:
	336	return flow_get_mark(skb);
	337	case FLOW_KEY_NFCT:
	338	return flow_get_nfct(skb);
	339	case FLOW_KEY_NFCT_SRC:
859c2012	340	return flow_get_nfct_src(skb, nhoff);
e5dfb815	341	case FLOW_KEY_NFCT_DST:
859c2012	342	return flow_get_nfct_dst(skb, nhoff);
e5dfb815	343	case FLOW_KEY_NFCT_PROTO_SRC:
859c2012	344	return flow_get_nfct_proto_src(skb, nhoff);
e5dfb815	345	case FLOW_KEY_NFCT_PROTO_DST:
859c2012	346	return flow_get_nfct_proto_dst(skb, nhoff);
e5dfb815 PM	347	case FLOW_KEY_RTCLASSID:
	348	return flow_get_rtclassid(skb);
	349	case FLOW_KEY_SKUID:
	350	return flow_get_skuid(skb);
	351	case FLOW_KEY_SKGID:
	352	return flow_get_skgid(skb);
9ec13810 PM	353	case FLOW_KEY_VLAN_TAG:
9ec13810 PM	354	return flow_get_vlan_tag(skb);
739a91ef CG	355	case FLOW_KEY_RXHASH:
739a91ef CG	356	return flow_get_rxhash(skb);
e5dfb815 PM	357	default:
	358	WARN_ON(1);
	359	return 0;
	360	}
	361	}
	362
dc7f9f6e	363	static int flow_classify(struct sk_buff skb, const struct tcf_proto tp,
e5dfb815 PM	364	struct tcf_result *res)
	365	{
	366	struct flow_head *head = tp->root;
	367	struct flow_filter *f;
	368	u32 keymask;
	369	u32 classid;
	370	unsigned int n, key;
	371	int r;
	372
	373	list_for_each_entry(f, &head->filters, list) {
	374	u32 keys[f->nkeys];
	375
	376	if (!tcf_em_tree_match(skb, &f->ematches, NULL))
	377	continue;
	378
	379	keymask = f->keymask;
	380
	381	for (n = 0; n < f->nkeys; n++) {
	382	key = ffs(keymask) - 1;
	383	keymask &= ~(1 << key);
	384	keys[n] = flow_key_get(skb, key);
	385	}
	386
	387	if (f->mode == FLOW_MODE_HASH)
72d9794f	388	classid = jhash2(keys, f->nkeys, f->hashrnd);
e5dfb815 PM	389	else {
	390	classid = keys[0];
	391	classid = (classid & f->mask) ^ f->xor;
	392	classid = (classid >> f->rshift) + f->addend;
	393	}
	394
	395	if (f->divisor)
	396	classid %= f->divisor;
	397
	398	res->class = 0;
	399	res->classid = TC_H_MAKE(f->baseclass, f->baseclass + classid);
	400
	401	r = tcf_exts_exec(skb, &f->exts, res);
	402	if (r < 0)
	403	continue;
	404	return r;
	405	}
	406	return -1;
	407	}
	408
72d9794f PM	409	static void flow_perturbation(unsigned long arg)
	410	{
	411	struct flow_filter f = (struct flow_filter )arg;
	412
	413	get_random_bytes(&f->hashrnd, 4);
	414	if (f->perturb_period)
	415	mod_timer(&f->perturb_timer, jiffies + f->perturb_period);
	416	}
	417
e5dfb815 PM	418	static const struct nla_policy flow_policy[TCA_FLOW_MAX + 1] = {
	419	[TCA_FLOW_KEYS] = { .type = NLA_U32 },
	420	[TCA_FLOW_MODE] = { .type = NLA_U32 },
	421	[TCA_FLOW_BASECLASS] = { .type = NLA_U32 },
	422	[TCA_FLOW_RSHIFT] = { .type = NLA_U32 },
	423	[TCA_FLOW_ADDEND] = { .type = NLA_U32 },
	424	[TCA_FLOW_MASK] = { .type = NLA_U32 },
	425	[TCA_FLOW_XOR] = { .type = NLA_U32 },
	426	[TCA_FLOW_DIVISOR] = { .type = NLA_U32 },
	427	[TCA_FLOW_ACT] = { .type = NLA_NESTED },
	428	[TCA_FLOW_POLICE] = { .type = NLA_NESTED },
	429	[TCA_FLOW_EMATCHES] = { .type = NLA_NESTED },
72d9794f	430	[TCA_FLOW_PERTURB] = { .type = NLA_U32 },
e5dfb815 PM	431	};
	432
	433	static int flow_change(struct tcf_proto *tp, unsigned long base,
	434	u32 handle, struct nlattr **tca,
	435	unsigned long *arg)
	436	{
	437	struct flow_head *head = tp->root;
	438	struct flow_filter *f;
	439	struct nlattr *opt = tca[TCA_OPTIONS];
	440	struct nlattr *tb[TCA_FLOW_MAX + 1];
	441	struct tcf_exts e;
	442	struct tcf_ematch_tree t;
	443	unsigned int nkeys = 0;
72d9794f	444	unsigned int perturb_period = 0;
e5dfb815 PM	445	u32 baseclass = 0;
	446	u32 keymask = 0;
	447	u32 mode;
	448	int err;
	449
	450	if (opt == NULL)
	451	return -EINVAL;
	452
	453	err = nla_parse_nested(tb, TCA_FLOW_MAX, opt, flow_policy);
	454	if (err < 0)
	455	return err;
	456
	457	if (tb[TCA_FLOW_BASECLASS]) {
	458	baseclass = nla_get_u32(tb[TCA_FLOW_BASECLASS]);
	459	if (TC_H_MIN(baseclass) == 0)
	460	return -EINVAL;
	461	}
	462
	463	if (tb[TCA_FLOW_KEYS]) {
	464	keymask = nla_get_u32(tb[TCA_FLOW_KEYS]);
e5dfb815 PM	465
	466	nkeys = hweight32(keymask);
	467	if (nkeys == 0)
	468	return -EINVAL;
4f250491 PM	469
	470	if (fls(keymask) - 1 > FLOW_KEY_MAX)
	471	return -EOPNOTSUPP;
e5dfb815 PM	472	}
	473
	474	err = tcf_exts_validate(tp, tb, tca[TCA_RATE], &e, &flow_ext_map);
	475	if (err < 0)
	476	return err;
	477
	478	err = tcf_em_tree_validate(tp, tb[TCA_FLOW_EMATCHES], &t);
	479	if (err < 0)
	480	goto err1;
	481
	482	f = (struct flow_filter )arg;
	483	if (f != NULL) {
	484	err = -EINVAL;
	485	if (f->handle != handle && handle)
	486	goto err2;
	487
	488	mode = f->mode;
	489	if (tb[TCA_FLOW_MODE])
	490	mode = nla_get_u32(tb[TCA_FLOW_MODE]);
	491	if (mode != FLOW_MODE_HASH && nkeys > 1)
	492	goto err2;
72d9794f PM	493
	494	if (mode == FLOW_MODE_HASH)
	495	perturb_period = f->perturb_period;
	496	if (tb[TCA_FLOW_PERTURB]) {
	497	if (mode != FLOW_MODE_HASH)
	498	goto err2;
	499	perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
	500	}
e5dfb815 PM	501	} else {
	502	err = -EINVAL;
	503	if (!handle)
	504	goto err2;
	505	if (!tb[TCA_FLOW_KEYS])
	506	goto err2;
	507
	508	mode = FLOW_MODE_MAP;
	509	if (tb[TCA_FLOW_MODE])
	510	mode = nla_get_u32(tb[TCA_FLOW_MODE]);
	511	if (mode != FLOW_MODE_HASH && nkeys > 1)
	512	goto err2;
	513
72d9794f PM	514	if (tb[TCA_FLOW_PERTURB]) {
	515	if (mode != FLOW_MODE_HASH)
	516	goto err2;
	517	perturb_period = nla_get_u32(tb[TCA_FLOW_PERTURB]) * HZ;
	518	}
	519
e5dfb815 PM	520	if (TC_H_MAJ(baseclass) == 0)
	521	baseclass = TC_H_MAKE(tp->q->handle, baseclass);
	522	if (TC_H_MIN(baseclass) == 0)
	523	baseclass = TC_H_MAKE(baseclass, 1);
	524
	525	err = -ENOBUFS;
	526	f = kzalloc(sizeof(*f), GFP_KERNEL);
	527	if (f == NULL)
	528	goto err2;
	529
	530	f->handle = handle;
	531	f->mask = ~0U;
72d9794f PM	532
	533	get_random_bytes(&f->hashrnd, 4);
	534	f->perturb_timer.function = flow_perturbation;
	535	f->perturb_timer.data = (unsigned long)f;
	536	init_timer_deferrable(&f->perturb_timer);
e5dfb815 PM	537	}
	538
	539	tcf_exts_change(tp, &f->exts, &e);
	540	tcf_em_tree_change(tp, &f->ematches, &t);
	541
	542	tcf_tree_lock(tp);
	543
	544	if (tb[TCA_FLOW_KEYS]) {
	545	f->keymask = keymask;
	546	f->nkeys = nkeys;
	547	}
	548
	549	f->mode = mode;
	550
	551	if (tb[TCA_FLOW_MASK])
	552	f->mask = nla_get_u32(tb[TCA_FLOW_MASK]);
	553	if (tb[TCA_FLOW_XOR])
	554	f->xor = nla_get_u32(tb[TCA_FLOW_XOR]);
	555	if (tb[TCA_FLOW_RSHIFT])
	556	f->rshift = nla_get_u32(tb[TCA_FLOW_RSHIFT]);
	557	if (tb[TCA_FLOW_ADDEND])
	558	f->addend = nla_get_u32(tb[TCA_FLOW_ADDEND]);
	559
	560	if (tb[TCA_FLOW_DIVISOR])
	561	f->divisor = nla_get_u32(tb[TCA_FLOW_DIVISOR]);
	562	if (baseclass)
	563	f->baseclass = baseclass;
	564
72d9794f PM	565	f->perturb_period = perturb_period;
	566	del_timer(&f->perturb_timer);
	567	if (perturb_period)
	568	mod_timer(&f->perturb_timer, jiffies + perturb_period);
	569
e5dfb815 PM	570	if (*arg == 0)
	571	list_add_tail(&f->list, &head->filters);
	572
	573	tcf_tree_unlock(tp);
	574
	575	*arg = (unsigned long)f;
	576	return 0;
	577
	578	err2:
	579	tcf_em_tree_destroy(tp, &t);
	580	err1:
	581	tcf_exts_destroy(tp, &e);
	582	return err;
	583	}
	584
	585	static void flow_destroy_filter(struct tcf_proto tp, struct flow_filter f)
	586	{
72d9794f	587	del_timer_sync(&f->perturb_timer);
e5dfb815 PM	588	tcf_exts_destroy(tp, &f->exts);
	589	tcf_em_tree_destroy(tp, &f->ematches);
	590	kfree(f);
	591	}
	592
	593	static int flow_delete(struct tcf_proto *tp, unsigned long arg)
	594	{
	595	struct flow_filter f = (struct flow_filter )arg;
	596
	597	tcf_tree_lock(tp);
	598	list_del(&f->list);
	599	tcf_tree_unlock(tp);
	600	flow_destroy_filter(tp, f);
	601	return 0;
	602	}
	603
	604	static int flow_init(struct tcf_proto *tp)
	605	{
	606	struct flow_head *head;
	607
e5dfb815 PM	608	head = kzalloc(sizeof(*head), GFP_KERNEL);
	609	if (head == NULL)
	610	return -ENOBUFS;
	611	INIT_LIST_HEAD(&head->filters);
	612	tp->root = head;
	613	return 0;
	614	}
	615
	616	static void flow_destroy(struct tcf_proto *tp)
	617	{
	618	struct flow_head *head = tp->root;
	619	struct flow_filter f, next;
	620
	621	list_for_each_entry_safe(f, next, &head->filters, list) {
	622	list_del(&f->list);
	623	flow_destroy_filter(tp, f);
	624	}
	625	kfree(head);
	626	}
	627
	628	static unsigned long flow_get(struct tcf_proto *tp, u32 handle)
	629	{
	630	struct flow_head *head = tp->root;
	631	struct flow_filter *f;
	632
	633	list_for_each_entry(f, &head->filters, list)
	634	if (f->handle == handle)
	635	return (unsigned long)f;
	636	return 0;
	637	}
	638
	639	static void flow_put(struct tcf_proto *tp, unsigned long f)
	640	{
e5dfb815 PM	641	}
	642
	643	static int flow_dump(struct tcf_proto *tp, unsigned long fh,
	644	struct sk_buff skb, struct tcmsg t)
	645	{
	646	struct flow_filter f = (struct flow_filter )fh;
	647	struct nlattr *nest;
	648
	649	if (f == NULL)
	650	return skb->len;
	651
	652	t->tcm_handle = f->handle;
	653
	654	nest = nla_nest_start(skb, TCA_OPTIONS);
	655	if (nest == NULL)
	656	goto nla_put_failure;
	657
	658	NLA_PUT_U32(skb, TCA_FLOW_KEYS, f->keymask);
	659	NLA_PUT_U32(skb, TCA_FLOW_MODE, f->mode);
	660
	661	if (f->mask != ~0 \|\| f->xor != 0) {
	662	NLA_PUT_U32(skb, TCA_FLOW_MASK, f->mask);
	663	NLA_PUT_U32(skb, TCA_FLOW_XOR, f->xor);
	664	}
	665	if (f->rshift)
	666	NLA_PUT_U32(skb, TCA_FLOW_RSHIFT, f->rshift);
	667	if (f->addend)
	668	NLA_PUT_U32(skb, TCA_FLOW_ADDEND, f->addend);
	669
	670	if (f->divisor)
	671	NLA_PUT_U32(skb, TCA_FLOW_DIVISOR, f->divisor);
	672	if (f->baseclass)
	673	NLA_PUT_U32(skb, TCA_FLOW_BASECLASS, f->baseclass);
	674
72d9794f PM	675	if (f->perturb_period)
	676	NLA_PUT_U32(skb, TCA_FLOW_PERTURB, f->perturb_period / HZ);
	677
e5dfb815 PM	678	if (tcf_exts_dump(skb, &f->exts, &flow_ext_map) < 0)
e5dfb815 PM	679	goto nla_put_failure;
0aead543	680	#ifdef CONFIG_NET_EMATCH
e5dfb815 PM	681	if (f->ematches.hdr.nmatches &&
	682	tcf_em_tree_dump(skb, &f->ematches, TCA_FLOW_EMATCHES) < 0)
	683	goto nla_put_failure;
0aead543	684	#endif
e5dfb815 PM	685	nla_nest_end(skb, nest);
	686
	687	if (tcf_exts_dump_stats(skb, &f->exts, &flow_ext_map) < 0)
	688	goto nla_put_failure;
	689
	690	return skb->len;
	691
	692	nla_put_failure:
	693	nlmsg_trim(skb, nest);
	694	return -1;
	695	}
	696
	697	static void flow_walk(struct tcf_proto tp, struct tcf_walker arg)
	698	{
	699	struct flow_head *head = tp->root;
	700	struct flow_filter *f;
	701
	702	list_for_each_entry(f, &head->filters, list) {
	703	if (arg->count < arg->skip)
	704	goto skip;
	705	if (arg->fn(tp, (unsigned long)f, arg) < 0) {
	706	arg->stop = 1;
	707	break;
	708	}
	709	skip:
	710	arg->count++;
	711	}
	712	}
	713
	714	static struct tcf_proto_ops cls_flow_ops __read_mostly = {
	715	.kind = "flow",
	716	.classify = flow_classify,
	717	.init = flow_init,
	718	.destroy = flow_destroy,
	719	.change = flow_change,
	720	.delete = flow_delete,
	721	.get = flow_get,
	722	.put = flow_put,
	723	.dump = flow_dump,
	724	.walk = flow_walk,
	725	.owner = THIS_MODULE,
	726	};
	727
	728	static int __init cls_flow_init(void)
	729	{
	730	return register_tcf_proto_ops(&cls_flow_ops);
	731	}
	732
	733	static void __exit cls_flow_exit(void)
	734	{
	735	unregister_tcf_proto_ops(&cls_flow_ops);
	736	}
	737
	738	module_init(cls_flow_init);
	739	module_exit(cls_flow_exit);
	740
	741	MODULE_LICENSE("GPL");
	742	MODULE_AUTHOR("Patrick McHardy <kaber@trash.net>");
	743	MODULE_DESCRIPTION("TC flow classifier");