[people/arne_f/kernel.git] / net / sched / act_sample.c

/*
 * net/sched/act_sample.c - Packet sampling tc action
 * Copyright (c) 2017 Yotam Gigi <yotamg@mellanox.com>
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

#include <linux/types.h>
#include <linux/kernel.h>
#include <linux/string.h>
#include <linux/errno.h>
#include <linux/skbuff.h>
#include <linux/rtnetlink.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/gfp.h>
#include <net/net_namespace.h>
#include <net/netlink.h>
#include <net/pkt_sched.h>
#include <linux/tc_act/tc_sample.h>
#include <net/tc_act/tc_sample.h>
#include <net/psample.h>

#include <linux/if_arp.h>

static unsigned int sample_net_id;
static struct tc_action_ops act_sample_ops;

static const struct nla_policy sample_policy[TCA_SAMPLE_MAX + 1] = {
	[TCA_SAMPLE_PARMS]		= { .len = sizeof(struct tc_sample) },
	[TCA_SAMPLE_RATE]		= { .type = NLA_U32 },
	[TCA_SAMPLE_TRUNC_SIZE]		= { .type = NLA_U32 },
	[TCA_SAMPLE_PSAMPLE_GROUP]	= { .type = NLA_U32 },
};

static int tcf_sample_init(struct net *net, struct nlattr *nla,
			   struct nlattr *est, struct tc_action **a, int ovr,
			   int bind)
{
	struct tc_action_net *tn = net_generic(net, sample_net_id);
	struct nlattr *tb[TCA_SAMPLE_MAX + 1];
	struct psample_group *psample_group;
	struct tc_sample *parm;
	struct tcf_sample *s;
	bool exists = false;
	int ret;

	if (!nla)
		return -EINVAL;
	ret = nla_parse_nested(tb, TCA_SAMPLE_MAX, nla, sample_policy, NULL);
	if (ret < 0)
		return ret;
	if (!tb[TCA_SAMPLE_PARMS] || !tb[TCA_SAMPLE_RATE] ||
	    !tb[TCA_SAMPLE_PSAMPLE_GROUP])
		return -EINVAL;

	parm = nla_data(tb[TCA_SAMPLE_PARMS]);

	exists = tcf_idr_check(tn, parm->index, a, bind);
	if (exists && bind)
		return 0;

	if (!exists) {
		ret = tcf_idr_create(tn, parm->index, est, a,
				     &act_sample_ops, bind, false);
		if (ret)
			return ret;
		ret = ACT_P_CREATED;
	} else {
		tcf_idr_release(*a, bind);
		if (!ovr)
			return -EEXIST;
	}
	s = to_sample(*a);

	s->tcf_action = parm->action;
	s->rate = nla_get_u32(tb[TCA_SAMPLE_RATE]);
	s->psample_group_num = nla_get_u32(tb[TCA_SAMPLE_PSAMPLE_GROUP]);
	psample_group = psample_group_get(net, s->psample_group_num);
	if (!psample_group) {
		if (ret == ACT_P_CREATED)
			tcf_idr_release(*a, bind);
		return -ENOMEM;
	}
	RCU_INIT_POINTER(s->psample_group, psample_group);

	if (tb[TCA_SAMPLE_TRUNC_SIZE]) {
		s->truncate = true;
		s->trunc_size = nla_get_u32(tb[TCA_SAMPLE_TRUNC_SIZE]);
	}

	if (ret == ACT_P_CREATED)
		tcf_idr_insert(tn, *a);
	return ret;
}

static void tcf_sample_cleanup_rcu(struct rcu_head *rcu)
{
	struct tcf_sample *s = container_of(rcu, struct tcf_sample, rcu);
	struct psample_group *psample_group;

	psample_group = rcu_dereference_protected(s->psample_group, 1);
	RCU_INIT_POINTER(s->psample_group, NULL);
	psample_group_put(psample_group);
}

static void tcf_sample_cleanup(struct tc_action *a, int bind)
{
	struct tcf_sample *s = to_sample(a);

	call_rcu(&s->rcu, tcf_sample_cleanup_rcu);
}

static bool tcf_sample_dev_ok_push(struct net_device *dev)
{
	switch (dev->type) {
	case ARPHRD_TUNNEL:
	case ARPHRD_TUNNEL6:
	case ARPHRD_SIT:
	case ARPHRD_IPGRE:
	case ARPHRD_VOID:
	case ARPHRD_NONE:
		return false;
	default:
		return true;
	}
}

static int tcf_sample_act(struct sk_buff *skb, const struct tc_action *a,
			  struct tcf_result *res)
{
	struct tcf_sample *s = to_sample(a);
	struct psample_group *psample_group;
	int retval;
	int size;
	int iif;
	int oif;

	tcf_lastuse_update(&s->tcf_tm);
	bstats_cpu_update(this_cpu_ptr(s->common.cpu_bstats), skb);
	retval = READ_ONCE(s->tcf_action);

	rcu_read_lock();
	psample_group = rcu_dereference(s->psample_group);

	/* randomly sample packets according to rate */
	if (psample_group && (prandom_u32() % s->rate == 0)) {
		if (!skb_at_tc_ingress(skb)) {
			iif = skb->skb_iif;
			oif = skb->dev->ifindex;
		} else {
			iif = skb->dev->ifindex;
			oif = 0;
		}

		/* on ingress, the mac header gets popped, so push it back */
		if (skb_at_tc_ingress(skb) && tcf_sample_dev_ok_push(skb->dev))
			skb_push(skb, skb->mac_len);

		size = s->truncate ? s->trunc_size : skb->len;
		psample_sample_packet(psample_group, skb, size, iif, oif,
				      s->rate);

		if (skb_at_tc_ingress(skb) && tcf_sample_dev_ok_push(skb->dev))
			skb_pull(skb, skb->mac_len);
	}

	rcu_read_unlock();
	return retval;
}

static int tcf_sample_dump(struct sk_buff *skb, struct tc_action *a,
			   int bind, int ref)
{
	unsigned char *b = skb_tail_pointer(skb);
	struct tcf_sample *s = to_sample(a);
	struct tc_sample opt = {
		.index      = s->tcf_index,
		.action     = s->tcf_action,
		.refcnt     = s->tcf_refcnt - ref,
		.bindcnt    = s->tcf_bindcnt - bind,
	};
	struct tcf_t t;

	if (nla_put(skb, TCA_SAMPLE_PARMS, sizeof(opt), &opt))
		goto nla_put_failure;

	tcf_tm_dump(&t, &s->tcf_tm);
	if (nla_put_64bit(skb, TCA_SAMPLE_TM, sizeof(t), &t, TCA_SAMPLE_PAD))
		goto nla_put_failure;

	if (nla_put_u32(skb, TCA_SAMPLE_RATE, s->rate))
		goto nla_put_failure;

	if (s->truncate)
		if (nla_put_u32(skb, TCA_SAMPLE_TRUNC_SIZE, s->trunc_size))
			goto nla_put_failure;

	if (nla_put_u32(skb, TCA_SAMPLE_PSAMPLE_GROUP, s->psample_group_num))
		goto nla_put_failure;
	return skb->len;

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

static int tcf_sample_walker(struct net *net, struct sk_buff *skb,
			     struct netlink_callback *cb, int type,
			     const struct tc_action_ops *ops)
{
	struct tc_action_net *tn = net_generic(net, sample_net_id);

	return tcf_generic_walker(tn, skb, cb, type, ops);
}

static int tcf_sample_search(struct net *net, struct tc_action **a, u32 index)
{
	struct tc_action_net *tn = net_generic(net, sample_net_id);

	return tcf_idr_search(tn, a, index);
}

static struct tc_action_ops act_sample_ops = {
	.kind	  = "sample",
	.type	  = TCA_ACT_SAMPLE,
	.owner	  = THIS_MODULE,
	.act	  = tcf_sample_act,
	.dump	  = tcf_sample_dump,
	.init	  = tcf_sample_init,
	.cleanup  = tcf_sample_cleanup,
	.walk	  = tcf_sample_walker,
	.lookup	  = tcf_sample_search,
	.size	  = sizeof(struct tcf_sample),
};

static __net_init int sample_init_net(struct net *net)
{
	struct tc_action_net *tn = net_generic(net, sample_net_id);

	return tc_action_net_init(tn, &act_sample_ops);
}

static void __net_exit sample_exit_net(struct net *net)
{
	struct tc_action_net *tn = net_generic(net, sample_net_id);

	tc_action_net_exit(tn);
}

static struct pernet_operations sample_net_ops = {
	.init = sample_init_net,
	.exit = sample_exit_net,
	.id   = &sample_net_id,
	.size = sizeof(struct tc_action_net),
};

static int __init sample_init_module(void)
{
	return tcf_register_action(&act_sample_ops, &sample_net_ops);
}

static void __exit sample_cleanup_module(void)
{
	rcu_barrier();
	tcf_unregister_action(&act_sample_ops, &sample_net_ops);
}

module_init(sample_init_module);
module_exit(sample_cleanup_module);

MODULE_AUTHOR("Yotam Gigi <yotam.gi@gmail.com>");
MODULE_DESCRIPTION("Packet sampling action");
MODULE_LICENSE("GPL v2");
Commit	Line	Data
5c5670fa YG	1	/*
	2	* net/sched/act_sample.c - Packet sampling tc action
	3	* Copyright (c) 2017 Yotam Gigi <yotamg@mellanox.com>
	4	*
	5	* This program is free software; you can redistribute it and/or modify
	6	* it under the terms of the GNU General Public License version 2 as
	7	* published by the Free Software Foundation.
	8	*/
	9
	10	#include <linux/types.h>
	11	#include <linux/kernel.h>
	12	#include <linux/string.h>
	13	#include <linux/errno.h>
	14	#include <linux/skbuff.h>
	15	#include <linux/rtnetlink.h>
	16	#include <linux/module.h>
	17	#include <linux/init.h>
	18	#include <linux/gfp.h>
	19	#include <net/net_namespace.h>
	20	#include <net/netlink.h>
	21	#include <net/pkt_sched.h>
	22	#include <linux/tc_act/tc_sample.h>
	23	#include <net/tc_act/tc_sample.h>
	24	#include <net/psample.h>
	25
	26	#include <linux/if_arp.h>
	27
5c5670fa YG	28	static unsigned int sample_net_id;
	29	static struct tc_action_ops act_sample_ops;
	30
	31	static const struct nla_policy sample_policy[TCA_SAMPLE_MAX + 1] = {
	32	[TCA_SAMPLE_PARMS] = { .len = sizeof(struct tc_sample) },
	33	[TCA_SAMPLE_RATE] = { .type = NLA_U32 },
	34	[TCA_SAMPLE_TRUNC_SIZE] = { .type = NLA_U32 },
	35	[TCA_SAMPLE_PSAMPLE_GROUP] = { .type = NLA_U32 },
	36	};
	37
	38	static int tcf_sample_init(struct net net, struct nlattr nla,
	39	struct nlattr est, struct tc_action *a, int ovr,
	40	int bind)
	41	{
	42	struct tc_action_net *tn = net_generic(net, sample_net_id);
	43	struct nlattr *tb[TCA_SAMPLE_MAX + 1];
	44	struct psample_group *psample_group;
	45	struct tc_sample *parm;
	46	struct tcf_sample *s;
	47	bool exists = false;
	48	int ret;
	49
	50	if (!nla)
	51	return -EINVAL;
fceb6435	52	ret = nla_parse_nested(tb, TCA_SAMPLE_MAX, nla, sample_policy, NULL);
5c5670fa YG	53	if (ret < 0)
	54	return ret;
	55	if (!tb[TCA_SAMPLE_PARMS] \|\| !tb[TCA_SAMPLE_RATE] \|\|
	56	!tb[TCA_SAMPLE_PSAMPLE_GROUP])
	57	return -EINVAL;
	58
	59	parm = nla_data(tb[TCA_SAMPLE_PARMS]);
	60
65a206c0	61	exists = tcf_idr_check(tn, parm->index, a, bind);
5c5670fa YG	62	if (exists && bind)
	63	return 0;
	64
	65	if (!exists) {
65a206c0 CM	66	ret = tcf_idr_create(tn, parm->index, est, a,
65a206c0 CM	67	&act_sample_ops, bind, false);
5c5670fa YG	68	if (ret)
	69	return ret;
	70	ret = ACT_P_CREATED;
	71	} else {
65a206c0	72	tcf_idr_release(*a, bind);
5c5670fa YG	73	if (!ovr)
	74	return -EEXIST;
	75	}
	76	s = to_sample(*a);
	77
5c5670fa YG	78	s->tcf_action = parm->action;
	79	s->rate = nla_get_u32(tb[TCA_SAMPLE_RATE]);
	80	s->psample_group_num = nla_get_u32(tb[TCA_SAMPLE_PSAMPLE_GROUP]);
	81	psample_group = psample_group_get(net, s->psample_group_num);
cadb9c9f YG	82	if (!psample_group) {
cadb9c9f YG	83	if (ret == ACT_P_CREATED)
65a206c0	84	tcf_idr_release(*a, bind);
5c5670fa	85	return -ENOMEM;
cadb9c9f	86	}
5c5670fa YG	87	RCU_INIT_POINTER(s->psample_group, psample_group);
	88
	89	if (tb[TCA_SAMPLE_TRUNC_SIZE]) {
	90	s->truncate = true;
	91	s->trunc_size = nla_get_u32(tb[TCA_SAMPLE_TRUNC_SIZE]);
	92	}
	93
	94	if (ret == ACT_P_CREATED)
65a206c0	95	tcf_idr_insert(tn, *a);
5c5670fa YG	96	return ret;
	97	}
	98
	99	static void tcf_sample_cleanup_rcu(struct rcu_head *rcu)
	100	{
	101	struct tcf_sample *s = container_of(rcu, struct tcf_sample, rcu);
	102	struct psample_group *psample_group;
	103
	104	psample_group = rcu_dereference_protected(s->psample_group, 1);
	105	RCU_INIT_POINTER(s->psample_group, NULL);
	106	psample_group_put(psample_group);
	107	}
	108
	109	static void tcf_sample_cleanup(struct tc_action *a, int bind)
	110	{
	111	struct tcf_sample *s = to_sample(a);
	112
	113	call_rcu(&s->rcu, tcf_sample_cleanup_rcu);
	114	}
	115
	116	static bool tcf_sample_dev_ok_push(struct net_device *dev)
	117	{
	118	switch (dev->type) {
	119	case ARPHRD_TUNNEL:
	120	case ARPHRD_TUNNEL6:
	121	case ARPHRD_SIT:
	122	case ARPHRD_IPGRE:
	123	case ARPHRD_VOID:
	124	case ARPHRD_NONE:
	125	return false;
	126	default:
	127	return true;
	128	}
	129	}
	130
	131	static int tcf_sample_act(struct sk_buff skb, const struct tc_action a,
	132	struct tcf_result *res)
	133	{
	134	struct tcf_sample *s = to_sample(a);
	135	struct psample_group *psample_group;
	136	int retval;
	137	int size;
	138	int iif;
	139	int oif;
	140
	141	tcf_lastuse_update(&s->tcf_tm);
	142	bstats_cpu_update(this_cpu_ptr(s->common.cpu_bstats), skb);
	143	retval = READ_ONCE(s->tcf_action);
	144
	145	rcu_read_lock();
	146	psample_group = rcu_dereference(s->psample_group);
	147
	148	/* randomly sample packets according to rate */
	149	if (psample_group && (prandom_u32() % s->rate == 0)) {
	150	if (!skb_at_tc_ingress(skb)) {
	151	iif = skb->skb_iif;
	152	oif = skb->dev->ifindex;
	153	} else {
	154	iif = skb->dev->ifindex;
	155	oif = 0;
	156	}
	157
	158	/* on ingress, the mac header gets popped, so push it back */
	159	if (skb_at_tc_ingress(skb) && tcf_sample_dev_ok_push(skb->dev))
160	skb_push(skb, skb->mac_len);
161
162	size = s->truncate ? s->trunc_size : skb->len;
163	psample_sample_packet(psample_group, skb, size, iif, oif,
164	s->rate);
165
166	if (skb_at_tc_ingress(skb) && tcf_sample_dev_ok_push(skb->dev))
167	skb_pull(skb, skb->mac_len);
168	}
169
170	rcu_read_unlock();
171	return retval;
172	}
173
174	static int tcf_sample_dump(struct sk_buff skb, struct tc_action a,
175	int bind, int ref)
176	{
177	unsigned char *b = skb_tail_pointer(skb);
178	struct tcf_sample *s = to_sample(a);
179	struct tc_sample opt = {
180	.index = s->tcf_index,
181	.action = s->tcf_action,
182	.refcnt = s->tcf_refcnt - ref,
183	.bindcnt = s->tcf_bindcnt - bind,
184	};
185	struct tcf_t t;
186
187	if (nla_put(skb, TCA_SAMPLE_PARMS, sizeof(opt), &opt))
188	goto nla_put_failure;
189
190	tcf_tm_dump(&t, &s->tcf_tm);
191	if (nla_put_64bit(skb, TCA_SAMPLE_TM, sizeof(t), &t, TCA_SAMPLE_PAD))
192	goto nla_put_failure;
193
194	if (nla_put_u32(skb, TCA_SAMPLE_RATE, s->rate))
195	goto nla_put_failure;
196
197	if (s->truncate)
198	if (nla_put_u32(skb, TCA_SAMPLE_TRUNC_SIZE, s->trunc_size))
199	goto nla_put_failure;
200
201	if (nla_put_u32(skb, TCA_SAMPLE_PSAMPLE_GROUP, s->psample_group_num))
202	goto nla_put_failure;
203	return skb->len;
204
205	nla_put_failure:
206	nlmsg_trim(skb, b);
207	return -1;
208	}
209
210	static int tcf_sample_walker(struct net net, struct sk_buff skb,
211	struct netlink_callback *cb, int type,
212	const struct tc_action_ops *ops)
213	{
214	struct tc_action_net *tn = net_generic(net, sample_net_id);
215
216	return tcf_generic_walker(tn, skb, cb, type, ops);
217	}
218
219	static int tcf_sample_search(struct net net, struct tc_action *a, u32 index)
220	{
221	struct tc_action_net *tn = net_generic(net, sample_net_id);
222
65a206c0	223	return tcf_idr_search(tn, a, index);
5c5670fa YG	224	}
	225
	226	static struct tc_action_ops act_sample_ops = {
	227	.kind = "sample",
	228	.type = TCA_ACT_SAMPLE,
	229	.owner = THIS_MODULE,
	230	.act = tcf_sample_act,
	231	.dump = tcf_sample_dump,
	232	.init = tcf_sample_init,
	233	.cleanup = tcf_sample_cleanup,
	234	.walk = tcf_sample_walker,
	235	.lookup = tcf_sample_search,
	236	.size = sizeof(struct tcf_sample),
	237	};
	238
	239	static __net_init int sample_init_net(struct net *net)
	240	{
	241	struct tc_action_net *tn = net_generic(net, sample_net_id);
	242
c7e460ce	243	return tc_action_net_init(tn, &act_sample_ops);
5c5670fa YG	244	}
	245
	246	static void __net_exit sample_exit_net(struct net *net)
	247	{
	248	struct tc_action_net *tn = net_generic(net, sample_net_id);
	249
	250	tc_action_net_exit(tn);
	251	}
	252
	253	static struct pernet_operations sample_net_ops = {
	254	.init = sample_init_net,
	255	.exit = sample_exit_net,
	256	.id = &sample_net_id,
	257	.size = sizeof(struct tc_action_net),
	258	};
	259
	260	static int __init sample_init_module(void)
	261	{
	262	return tcf_register_action(&act_sample_ops, &sample_net_ops);
	263	}
	264
	265	static void __exit sample_cleanup_module(void)
	266	{
46e235c1	267	rcu_barrier();
5c5670fa YG	268	tcf_unregister_action(&act_sample_ops, &sample_net_ops);
	269	}
	270
	271	module_init(sample_init_module);
	272	module_exit(sample_cleanup_module);
	273
f1fd20c3	274	MODULE_AUTHOR("Yotam Gigi <yotam.gi@gmail.com>");
5c5670fa YG	275	MODULE_DESCRIPTION("Packet sampling action");
5c5670fa YG	276	MODULE_LICENSE("GPL v2");