[mirror_ovs.git] / datapath / actions.c

/*
 * Copyright (c) 2007-2012 Nicira, Inc.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of version 2 of the GNU General Public
 * License as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
 * 02110-1301, USA
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/skbuff.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/openvswitch.h>
#include <linux/tcp.h>
#include <linux/udp.h>
#include <linux/in6.h>
#include <linux/if_arp.h>
#include <linux/if_vlan.h>
#include <net/ip.h>
#include <net/ipv6.h>
#include <net/checksum.h>
#include <net/dsfield.h>

#include "checksum.h"
#include "datapath.h"
#include "vlan.h"
#include "vport.h"

static int do_execute_actions(struct datapath *dp, struct sk_buff *skb,
			      const struct nlattr *attr, int len, bool keep_skb);

static int make_writable(struct sk_buff *skb, int write_len)
{
	if (!skb_cloned(skb) || skb_clone_writable(skb, write_len))
		return 0;

	return pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
}

/* remove VLAN header from packet and update csum accordingly. */
static int __pop_vlan_tci(struct sk_buff *skb, __be16 *current_tci)
{
	struct vlan_hdr *vhdr;
	int err;

	err = make_writable(skb, VLAN_ETH_HLEN);
	if (unlikely(err))
		return err;

	if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
		skb->csum = csum_sub(skb->csum, csum_partial(skb->data
					+ (2 * ETH_ALEN), VLAN_HLEN, 0));

	vhdr = (struct vlan_hdr *)(skb->data + ETH_HLEN);
	*current_tci = vhdr->h_vlan_TCI;

	memmove(skb->data + VLAN_HLEN, skb->data, 2 * ETH_ALEN);
	__skb_pull(skb, VLAN_HLEN);

	vlan_set_encap_proto(skb, vhdr);
	skb->mac_header += VLAN_HLEN;
	skb_reset_mac_len(skb);

	return 0;
}

static int pop_vlan(struct sk_buff *skb)
{
	__be16 tci;
	int err;

	if (likely(vlan_tx_tag_present(skb))) {
		vlan_set_tci(skb, 0);
	} else {
		if (unlikely(skb->protocol != htons(ETH_P_8021Q) ||
			     skb->len < VLAN_ETH_HLEN))
			return 0;

		err = __pop_vlan_tci(skb, &tci);
		if (err)
			return err;
	}
	/* move next vlan tag to hw accel tag */
	if (likely(skb->protocol != htons(ETH_P_8021Q) ||
		   skb->len < VLAN_ETH_HLEN))
		return 0;

	err = __pop_vlan_tci(skb, &tci);
	if (unlikely(err))
		return err;

	__vlan_hwaccel_put_tag(skb, ntohs(tci));
	return 0;
}

static int push_vlan(struct sk_buff *skb, const struct ovs_action_push_vlan *vlan)
{
	if (unlikely(vlan_tx_tag_present(skb))) {
		u16 current_tag;

		/* push down current VLAN tag */
		current_tag = vlan_tx_tag_get(skb);

		if (!__vlan_put_tag(skb, current_tag))
			return -ENOMEM;

		if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
			skb->csum = csum_add(skb->csum, csum_partial(skb->data
					+ (2 * ETH_ALEN), VLAN_HLEN, 0));

	}
	__vlan_hwaccel_put_tag(skb, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT);
	return 0;
}

static int set_eth_addr(struct sk_buff *skb,
			const struct ovs_key_ethernet *eth_key)
{
	int err;
	err = make_writable(skb, ETH_HLEN);
	if (unlikely(err))
		return err;

	memcpy(eth_hdr(skb)->h_source, eth_key->eth_src, ETH_ALEN);
	memcpy(eth_hdr(skb)->h_dest, eth_key->eth_dst, ETH_ALEN);

	return 0;
}

static void set_ip_addr(struct sk_buff *skb, struct iphdr *nh,
				__be32 *addr, __be32 new_addr)
{
	int transport_len = skb->len - skb_transport_offset(skb);

	if (nh->protocol == IPPROTO_TCP) {
		if (likely(transport_len >= sizeof(struct tcphdr)))
			inet_proto_csum_replace4(&tcp_hdr(skb)->check, skb,
						 *addr, new_addr, 1);
	} else if (nh->protocol == IPPROTO_UDP) {
		if (likely(transport_len >= sizeof(struct udphdr))) {
			struct udphdr *uh = udp_hdr(skb);

			if (uh->check ||
			    get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
				inet_proto_csum_replace4(&uh->check, skb,
							 *addr, new_addr, 1);
				if (!uh->check)
					uh->check = CSUM_MANGLED_0;
			}
		}
	}

	csum_replace4(&nh->check, *addr, new_addr);
	skb_clear_rxhash(skb);
	*addr = new_addr;
}

static void update_ipv6_checksum(struct sk_buff *skb, u8 l4_proto,
				 __be32 addr[4], const __be32 new_addr[4])
{
	int transport_len = skb->len - skb_transport_offset(skb);

	if (l4_proto == IPPROTO_TCP) {
		if (likely(transport_len >= sizeof(struct tcphdr)))
			inet_proto_csum_replace16(&tcp_hdr(skb)->check, skb,
						  addr, new_addr, 1);
	} else if (l4_proto == IPPROTO_UDP) {
		if (likely(transport_len >= sizeof(struct udphdr))) {
			struct udphdr *uh = udp_hdr(skb);

			if (uh->check ||
			    get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
				inet_proto_csum_replace16(&uh->check, skb,
							  addr, new_addr, 1);
				if (!uh->check)
					uh->check = CSUM_MANGLED_0;
			}
		}
	}
}

static void set_ipv6_addr(struct sk_buff *skb, u8 l4_proto,
			  __be32 addr[4], const __be32 new_addr[4],
			  bool recalculate_csum)
{
	if (recalculate_csum)
		update_ipv6_checksum(skb, l4_proto, addr, new_addr);

	skb_clear_rxhash(skb);
	memcpy(addr, new_addr, sizeof(__be32[4]));
}

static void set_ipv6_tc(struct ipv6hdr *nh, u8 tc)
{
	nh->priority = tc >> 4;
	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0x0F) | ((tc & 0x0F) << 4);
}

static void set_ipv6_fl(struct ipv6hdr *nh, u32 fl)
{
	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0xF0) | (fl & 0x000F0000) >> 16;
	nh->flow_lbl[1] = (fl & 0x0000FF00) >> 8;
	nh->flow_lbl[2] = fl & 0x000000FF;
}

static void set_ip_ttl(struct sk_buff *skb, struct iphdr *nh, u8 new_ttl)
{
	csum_replace2(&nh->check, htons(nh->ttl << 8), htons(new_ttl << 8));
	nh->ttl = new_ttl;
}

static int set_ipv4(struct sk_buff *skb, const struct ovs_key_ipv4 *ipv4_key)
{
	struct iphdr *nh;
	int err;

	err = make_writable(skb, skb_network_offset(skb) +
				 sizeof(struct iphdr));
	if (unlikely(err))
		return err;

	nh = ip_hdr(skb);

	if (ipv4_key->ipv4_src != nh->saddr)
		set_ip_addr(skb, nh, &nh->saddr, ipv4_key->ipv4_src);

	if (ipv4_key->ipv4_dst != nh->daddr)
		set_ip_addr(skb, nh, &nh->daddr, ipv4_key->ipv4_dst);

	if (ipv4_key->ipv4_tos != nh->tos)
		ipv4_change_dsfield(nh, 0, ipv4_key->ipv4_tos);

	if (ipv4_key->ipv4_ttl != nh->ttl)
		set_ip_ttl(skb, nh, ipv4_key->ipv4_ttl);

	return 0;
}

static int set_ipv6(struct sk_buff *skb, const struct ovs_key_ipv6 *ipv6_key)
{
	struct ipv6hdr *nh;
	int err;
	__be32 *saddr;
	__be32 *daddr;

	err = make_writable(skb, skb_network_offset(skb) +
			    sizeof(struct ipv6hdr));
	if (unlikely(err))
		return err;

	nh = ipv6_hdr(skb);
	saddr = (__be32 *)&nh->saddr;
	daddr = (__be32 *)&nh->daddr;

	if (memcmp(ipv6_key->ipv6_src, saddr, sizeof(ipv6_key->ipv6_src)))
		set_ipv6_addr(skb, ipv6_key->ipv6_proto, saddr,
			      ipv6_key->ipv6_src, true);

	if (memcmp(ipv6_key->ipv6_dst, daddr, sizeof(ipv6_key->ipv6_dst))) {
		unsigned int offset = 0;
		int flags = OVS_IP6T_FH_F_SKIP_RH;
		bool recalc_csum = true;

		if (ipv6_ext_hdr(nh->nexthdr))
			recalc_csum = ipv6_find_hdr(skb, &offset,
						    NEXTHDR_ROUTING, NULL,
						    &flags) != NEXTHDR_ROUTING;

		set_ipv6_addr(skb, ipv6_key->ipv6_proto, daddr,
			      ipv6_key->ipv6_dst, recalc_csum);
	}

	set_ipv6_tc(nh, ipv6_key->ipv6_tclass);
	set_ipv6_fl(nh, ntohl(ipv6_key->ipv6_label));
	nh->hop_limit = ipv6_key->ipv6_hlimit;

	return 0;
}

/* Must follow make_writable() since that can move the skb data. */
static void set_tp_port(struct sk_buff *skb, __be16 *port,
			 __be16 new_port, __sum16 *check)
{
	inet_proto_csum_replace2(check, skb, *port, new_port, 0);
	*port = new_port;
	skb_clear_rxhash(skb);
}

static void set_udp_port(struct sk_buff *skb, __be16 *port, __be16 new_port)
{
	struct udphdr *uh = udp_hdr(skb);

	if (uh->check && get_ip_summed(skb) != OVS_CSUM_PARTIAL) {
		set_tp_port(skb, port, new_port, &uh->check);

		if (!uh->check)
			uh->check = CSUM_MANGLED_0;
	} else {
		*port = new_port;
		skb_clear_rxhash(skb);
	}
}

static int set_udp(struct sk_buff *skb, const struct ovs_key_udp *udp_port_key)
{
	struct udphdr *uh;
	int err;

	err = make_writable(skb, skb_transport_offset(skb) +
				 sizeof(struct udphdr));
	if (unlikely(err))
		return err;

	uh = udp_hdr(skb);
	if (udp_port_key->udp_src != uh->source)
		set_udp_port(skb, &uh->source, udp_port_key->udp_src);

	if (udp_port_key->udp_dst != uh->dest)
		set_udp_port(skb, &uh->dest, udp_port_key->udp_dst);

	return 0;
}

static int set_tcp(struct sk_buff *skb, const struct ovs_key_tcp *tcp_port_key)
{
	struct tcphdr *th;
	int err;

	err = make_writable(skb, skb_transport_offset(skb) +
				 sizeof(struct tcphdr));
	if (unlikely(err))
		return err;

	th = tcp_hdr(skb);
	if (tcp_port_key->tcp_src != th->source)
		set_tp_port(skb, &th->source, tcp_port_key->tcp_src, &th->check);

	if (tcp_port_key->tcp_dst != th->dest)
		set_tp_port(skb, &th->dest, tcp_port_key->tcp_dst, &th->check);

	return 0;
}

static int do_output(struct datapath *dp, struct sk_buff *skb, int out_port)
{
	struct vport *vport;

	if (unlikely(!skb))
		return -ENOMEM;

	vport = ovs_vport_rcu(dp, out_port);
	if (unlikely(!vport)) {
		kfree_skb(skb);
		return -ENODEV;
	}

	ovs_vport_send(vport, skb);
	return 0;
}

static int output_userspace(struct datapath *dp, struct sk_buff *skb,
			    const struct nlattr *attr)
{
	struct dp_upcall_info upcall;
	const struct nlattr *a;
	int rem;

	upcall.cmd = OVS_PACKET_CMD_ACTION;
	upcall.key = &OVS_CB(skb)->flow->key;
	upcall.userdata = NULL;
	upcall.portid = 0;

	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
		 a = nla_next(a, &rem)) {
		switch (nla_type(a)) {
		case OVS_USERSPACE_ATTR_USERDATA:
			upcall.userdata = a;
			break;

		case OVS_USERSPACE_ATTR_PID:
			upcall.portid = nla_get_u32(a);
			break;
		}
	}

	return ovs_dp_upcall(dp, skb, &upcall);
}

static int sample(struct datapath *dp, struct sk_buff *skb,
		  const struct nlattr *attr)
{
	const struct nlattr *acts_list = NULL;
	const struct nlattr *a;
	int rem;

	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
		 a = nla_next(a, &rem)) {
		switch (nla_type(a)) {
		case OVS_SAMPLE_ATTR_PROBABILITY:
			if (net_random() >= nla_get_u32(a))
				return 0;
			break;

		case OVS_SAMPLE_ATTR_ACTIONS:
			acts_list = a;
			break;
		}
	}

	return do_execute_actions(dp, skb, nla_data(acts_list),
				  nla_len(acts_list), true);
}

static int execute_set_action(struct sk_buff *skb,
				 const struct nlattr *nested_attr)
{
	int err = 0;

	switch (nla_type(nested_attr)) {
	case OVS_KEY_ATTR_PRIORITY:
		skb->priority = nla_get_u32(nested_attr);
		break;

	case OVS_KEY_ATTR_SKB_MARK:
		skb_set_mark(skb, nla_get_u32(nested_attr));
		break;

	case OVS_KEY_ATTR_IPV4_TUNNEL:
		OVS_CB(skb)->tun_key = nla_data(nested_attr);
		break;

	case OVS_KEY_ATTR_ETHERNET:
		err = set_eth_addr(skb, nla_data(nested_attr));
		break;

	case OVS_KEY_ATTR_IPV4:
		err = set_ipv4(skb, nla_data(nested_attr));
		break;

	case OVS_KEY_ATTR_IPV6:
		err = set_ipv6(skb, nla_data(nested_attr));
		break;

	case OVS_KEY_ATTR_TCP:
		err = set_tcp(skb, nla_data(nested_attr));
		break;

	case OVS_KEY_ATTR_UDP:
		err = set_udp(skb, nla_data(nested_attr));
		break;
	}

	return err;
}

/* Execute a list of actions against 'skb'. */
static int do_execute_actions(struct datapath *dp, struct sk_buff *skb,
			const struct nlattr *attr, int len, bool keep_skb)
{
	/* Every output action needs a separate clone of 'skb', but the common
	 * case is just a single output action, so that doing a clone and
	 * then freeing the original skbuff is wasteful.  So the following code
	 * is slightly obscure just to avoid that. */
	int prev_port = -1;
	const struct nlattr *a;
	int rem;

	for (a = attr, rem = len; rem > 0;
	     a = nla_next(a, &rem)) {
		int err = 0;

		if (prev_port != -1) {
			do_output(dp, skb_clone(skb, GFP_ATOMIC), prev_port);
			prev_port = -1;
		}

		switch (nla_type(a)) {
		case OVS_ACTION_ATTR_OUTPUT:
			prev_port = nla_get_u32(a);
			break;

		case OVS_ACTION_ATTR_USERSPACE:
			output_userspace(dp, skb, a);
			break;

		case OVS_ACTION_ATTR_PUSH_VLAN:
			err = push_vlan(skb, nla_data(a));
			if (unlikely(err)) /* skb already freed. */
				return err;
			break;

		case OVS_ACTION_ATTR_POP_VLAN:
			err = pop_vlan(skb);
			break;

		case OVS_ACTION_ATTR_SET:
			err = execute_set_action(skb, nla_data(a));
			break;

		case OVS_ACTION_ATTR_SAMPLE:
			err = sample(dp, skb, a);
			break;
		}

		if (unlikely(err)) {
			kfree_skb(skb);
			return err;
		}
	}

	if (prev_port != -1) {
		if (keep_skb)
			skb = skb_clone(skb, GFP_ATOMIC);

		do_output(dp, skb, prev_port);
	} else if (!keep_skb)
		consume_skb(skb);

	return 0;
}

/* We limit the number of times that we pass into execute_actions()
 * to avoid blowing out the stack in the event that we have a loop. */
#define MAX_LOOPS 5

struct loop_counter {
	u8 count;		/* Count. */
	bool looping;		/* Loop detected? */
};

static DEFINE_PER_CPU(struct loop_counter, loop_counters);

static int loop_suppress(struct datapath *dp, struct sw_flow_actions *actions)
{
	if (net_ratelimit())
		pr_warn("%s: flow looped %d times, dropping\n",
				ovs_dp_name(dp), MAX_LOOPS);
	actions->actions_len = 0;
	return -ELOOP;
}

/* Execute a list of actions against 'skb'. */
int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb)
{
	struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts);
	struct loop_counter *loop;
	int error;

	/* Check whether we've looped too much. */
	loop = &__get_cpu_var(loop_counters);
	if (unlikely(++loop->count > MAX_LOOPS))
		loop->looping = true;
	if (unlikely(loop->looping)) {
		error = loop_suppress(dp, acts);
		kfree_skb(skb);
		goto out_loop;
	}

	OVS_CB(skb)->tun_key = NULL;
	error = do_execute_actions(dp, skb, acts->actions,
					 acts->actions_len, false);

	/* Check whether sub-actions looped too much. */
	if (unlikely(loop->looping))
		error = loop_suppress(dp, acts);

out_loop:
	/* Decrement loop counter. */
	if (!--loop->count)
		loop->looping = false;

	return error;
}
Commit	Line	Data
064af421	1	/*
e0edde6f	2	* Copyright (c) 2007-2012 Nicira, Inc.
a14bc59f	3	*
a9a29d22 JG	4	* This program is free software; you can redistribute it and/or
	5	* modify it under the terms of version 2 of the GNU General Public
	6	* License as published by the Free Software Foundation.
	7	*
	8	* This program is distributed in the hope that it will be useful, but
	9	* WITHOUT ANY WARRANTY; without even the implied warranty of
	10	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	11	* General Public License for more details.
	12	*
	13	* You should have received a copy of the GNU General Public License
	14	* along with this program; if not, write to the Free Software
	15	* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
	16	* 02110-1301, USA
064af421 BP	17	*/
064af421 BP	18
e9141eec PS	19	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
e9141eec PS	20
064af421 BP	21	#include <linux/skbuff.h>
	22	#include <linux/in.h>
	23	#include <linux/ip.h>
077257b8	24	#include <linux/openvswitch.h>
064af421 BP	25	#include <linux/tcp.h>
	26	#include <linux/udp.h>
	27	#include <linux/in6.h>
401eeb92	28	#include <linux/if_arp.h>
064af421 BP	29	#include <linux/if_vlan.h>
064af421 BP	30	#include <net/ip.h>
bc7a5acd	31	#include <net/ipv6.h>
064af421	32	#include <net/checksum.h>
530180fd	33	#include <net/dsfield.h>
f2459fe7	34
dd8d6b8c	35	#include "checksum.h"
f2459fe7	36	#include "datapath.h"
6ce39213	37	#include "vlan.h"
f2459fe7	38	#include "vport.h"
064af421	39
6ff686f2	40	static int do_execute_actions(struct datapath dp, struct sk_buff skb,
85c9de19	41	const struct nlattr *attr, int len, bool keep_skb);
871dfe07	42
10db8b20	43	static int make_writable(struct sk_buff *skb, int write_len)
064af421	44	{
10db8b20 JG	45	if (!skb_cloned(skb) \|\| skb_clone_writable(skb, write_len))
10db8b20 JG	46	return 0;
0cd8a05e	47
10db8b20	48	return pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
064af421 BP	49	}
064af421 BP	50
b2492cb7	51	/* remove VLAN header from packet and update csum accordingly. */
d9065a90	52	static int __pop_vlan_tci(struct sk_buff skb, __be16 current_tci)
064af421	53	{
ceb176fd	54	struct vlan_hdr *vhdr;
10db8b20	55	int err;
064af421	56
10db8b20 JG	57	err = make_writable(skb, VLAN_ETH_HLEN);
	58	if (unlikely(err))
	59	return err;
6ce39213	60
dd8d6b8c	61	if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
635c9298	62	skb->csum = csum_sub(skb->csum, csum_partial(skb->data
af9d14a8	63	+ (2 * ETH_ALEN), VLAN_HLEN, 0));
635c9298	64
ceb176fd PS	65	vhdr = (struct vlan_hdr *)(skb->data + ETH_HLEN);
ceb176fd PS	66	*current_tci = vhdr->h_vlan_TCI;
d9065a90	67
6ce39213	68	memmove(skb->data + VLAN_HLEN, skb->data, 2 * ETH_ALEN);
ceb176fd	69	__skb_pull(skb, VLAN_HLEN);
064af421	70
ceb176fd	71	vlan_set_encap_proto(skb, vhdr);
064af421	72	skb->mac_header += VLAN_HLEN;
ceb176fd	73	skb_reset_mac_len(skb);
064af421	74
10db8b20	75	return 0;
064af421 BP	76	}
064af421 BP	77
d9065a90	78	static int pop_vlan(struct sk_buff *skb)
064af421	79	{
d9065a90 PS	80	__be16 tci;
d9065a90 PS	81	int err;
10db8b20	82
d9065a90 PS	83	if (likely(vlan_tx_tag_present(skb))) {
	84	vlan_set_tci(skb, 0);
	85	} else {
	86	if (unlikely(skb->protocol != htons(ETH_P_8021Q) \|\|
fb516ed8	87	skb->len < VLAN_ETH_HLEN))
10db8b20	88	return 0;
064af421	89
d9065a90 PS	90	err = __pop_vlan_tci(skb, &tci);
d9065a90 PS	91	if (err)
10db8b20	92	return err;
064af421	93	}
d9065a90 PS	94	/* move next vlan tag to hw accel tag */
d9065a90 PS	95	if (likely(skb->protocol != htons(ETH_P_8021Q) \|\|
fb516ed8	96	skb->len < VLAN_ETH_HLEN))
d9065a90 PS	97	return 0;
	98
	99	err = __pop_vlan_tci(skb, &tci);
	100	if (unlikely(err))
	101	return err;
064af421	102
10db8b20	103	__vlan_hwaccel_put_tag(skb, ntohs(tci));
d9065a90 PS	104	return 0;
	105	}
	106
fea393b1	107	static int push_vlan(struct sk_buff skb, const struct ovs_action_push_vlan vlan)
d9065a90 PS	108	{
	109	if (unlikely(vlan_tx_tag_present(skb))) {
	110	u16 current_tag;
	111
	112	/* push down current VLAN tag */
	113	current_tag = vlan_tx_tag_get(skb);
	114
	115	if (!__vlan_put_tag(skb, current_tag))
	116	return -ENOMEM;
10db8b20	117
d9065a90 PS	118	if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
d9065a90 PS	119	skb->csum = csum_add(skb->csum, csum_partial(skb->data
af9d14a8	120	+ (2 * ETH_ALEN), VLAN_HLEN, 0));
d9065a90 PS	121
d9065a90 PS	122	}
8ddc056d	123	__vlan_hwaccel_put_tag(skb, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT);
10db8b20	124	return 0;
064af421 BP	125	}
064af421 BP	126
4edb9ae9 PS	127	static int set_eth_addr(struct sk_buff *skb,
4edb9ae9 PS	128	const struct ovs_key_ethernet *eth_key)
ca78c6b6	129	{
4edb9ae9 PS	130	int err;
	131	err = make_writable(skb, ETH_HLEN);
	132	if (unlikely(err))
	133	return err;
	134
39640c1b PS	135	memcpy(eth_hdr(skb)->h_source, eth_key->eth_src, ETH_ALEN);
39640c1b PS	136	memcpy(eth_hdr(skb)->h_dest, eth_key->eth_dst, ETH_ALEN);
4edb9ae9 PS	137
4edb9ae9 PS	138	return 0;
ca78c6b6 BP	139	}
ca78c6b6 BP	140
4edb9ae9 PS	141	static void set_ip_addr(struct sk_buff skb, struct iphdr nh,
4edb9ae9 PS	142	__be32 *addr, __be32 new_addr)
ca78c6b6 BP	143	{
ca78c6b6 BP	144	int transport_len = skb->len - skb_transport_offset(skb);
4edb9ae9 PS	145
4edb9ae9 PS	146	if (nh->protocol == IPPROTO_TCP) {
ca78c6b6	147	if (likely(transport_len >= sizeof(struct tcphdr)))
4edb9ae9 PS	148	inet_proto_csum_replace4(&tcp_hdr(skb)->check, skb,
	149	*addr, new_addr, 1);
	150	} else if (nh->protocol == IPPROTO_UDP) {
55ce87bc JG	151	if (likely(transport_len >= sizeof(struct udphdr))) {
	152	struct udphdr *uh = udp_hdr(skb);
	153
	154	if (uh->check \|\|
	155	get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
	156	inet_proto_csum_replace4(&uh->check, skb,
	157	*addr, new_addr, 1);
	158	if (!uh->check)
	159	uh->check = CSUM_MANGLED_0;
	160	}
	161	}
ca78c6b6	162	}
4edb9ae9 PS	163
	164	csum_replace4(&nh->check, *addr, new_addr);
	165	skb_clear_rxhash(skb);
	166	*addr = new_addr;
ca78c6b6 BP	167	}
ca78c6b6 BP	168
bc7a5acd AA	169	static void update_ipv6_checksum(struct sk_buff *skb, u8 l4_proto,
	170	__be32 addr[4], const __be32 new_addr[4])
	171	{
	172	int transport_len = skb->len - skb_transport_offset(skb);
	173
	174	if (l4_proto == IPPROTO_TCP) {
	175	if (likely(transport_len >= sizeof(struct tcphdr)))
	176	inet_proto_csum_replace16(&tcp_hdr(skb)->check, skb,
	177	addr, new_addr, 1);
	178	} else if (l4_proto == IPPROTO_UDP) {
	179	if (likely(transport_len >= sizeof(struct udphdr))) {
	180	struct udphdr *uh = udp_hdr(skb);
	181
	182	if (uh->check \|\|
	183	get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
	184	inet_proto_csum_replace16(&uh->check, skb,
	185	addr, new_addr, 1);
	186	if (!uh->check)
	187	uh->check = CSUM_MANGLED_0;
	188	}
	189	}
	190	}
	191	}
	192
	193	static void set_ipv6_addr(struct sk_buff *skb, u8 l4_proto,
	194	__be32 addr[4], const __be32 new_addr[4],
	195	bool recalculate_csum)
	196	{
	197	if (recalculate_csum)
	198	update_ipv6_checksum(skb, l4_proto, addr, new_addr);
	199
	200	skb_clear_rxhash(skb);
	201	memcpy(addr, new_addr, sizeof(__be32[4]));
	202	}
	203
	204	static void set_ipv6_tc(struct ipv6hdr *nh, u8 tc)
	205	{
	206	nh->priority = tc >> 4;
	207	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0x0F) \| ((tc & 0x0F) << 4);
	208	}
	209
	210	static void set_ipv6_fl(struct ipv6hdr *nh, u32 fl)
	211	{
	212	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0xF0) \| (fl & 0x000F0000) >> 16;
	213	nh->flow_lbl[1] = (fl & 0x0000FF00) >> 8;
	214	nh->flow_lbl[2] = fl & 0x000000FF;
	215	}
	216
a61680c6 JP	217	static void set_ip_ttl(struct sk_buff skb, struct iphdr nh, u8 new_ttl)
	218	{
	219	csum_replace2(&nh->check, htons(nh->ttl << 8), htons(new_ttl << 8));
	220	nh->ttl = new_ttl;
	221	}
	222
4edb9ae9	223	static int set_ipv4(struct sk_buff skb, const struct ovs_key_ipv4 ipv4_key)
064af421	224	{
ca78c6b6	225	struct iphdr *nh;
10db8b20	226	int err;
ca78c6b6	227
10db8b20 JG	228	err = make_writable(skb, skb_network_offset(skb) +
	229	sizeof(struct iphdr));
	230	if (unlikely(err))
	231	return err;
ca78c6b6 BP	232
ca78c6b6 BP	233	nh = ip_hdr(skb);
ca78c6b6	234
4edb9ae9 PS	235	if (ipv4_key->ipv4_src != nh->saddr)
4edb9ae9 PS	236	set_ip_addr(skb, nh, &nh->saddr, ipv4_key->ipv4_src);
ca78c6b6	237
4edb9ae9 PS	238	if (ipv4_key->ipv4_dst != nh->daddr)
4edb9ae9 PS	239	set_ip_addr(skb, nh, &nh->daddr, ipv4_key->ipv4_dst);
a4a26436	240
4edb9ae9	241	if (ipv4_key->ipv4_tos != nh->tos)
530180fd	242	ipv4_change_dsfield(nh, 0, ipv4_key->ipv4_tos);
ca78c6b6	243
a61680c6 JP	244	if (ipv4_key->ipv4_ttl != nh->ttl)
	245	set_ip_ttl(skb, nh, ipv4_key->ipv4_ttl);
	246
10db8b20	247	return 0;
064af421 BP	248	}
064af421 BP	249
bc7a5acd AA	250	static int set_ipv6(struct sk_buff skb, const struct ovs_key_ipv6 ipv6_key)
	251	{
	252	struct ipv6hdr *nh;
	253	int err;
	254	__be32 *saddr;
	255	__be32 *daddr;
	256
	257	err = make_writable(skb, skb_network_offset(skb) +
	258	sizeof(struct ipv6hdr));
	259	if (unlikely(err))
	260	return err;
	261
	262	nh = ipv6_hdr(skb);
	263	saddr = (__be32 *)&nh->saddr;
	264	daddr = (__be32 *)&nh->daddr;
	265
	266	if (memcmp(ipv6_key->ipv6_src, saddr, sizeof(ipv6_key->ipv6_src)))
	267	set_ipv6_addr(skb, ipv6_key->ipv6_proto, saddr,
	268	ipv6_key->ipv6_src, true);
	269
	270	if (memcmp(ipv6_key->ipv6_dst, daddr, sizeof(ipv6_key->ipv6_dst))) {
	271	unsigned int offset = 0;
	272	int flags = OVS_IP6T_FH_F_SKIP_RH;
	273	bool recalc_csum = true;
	274
	275	if (ipv6_ext_hdr(nh->nexthdr))
	276	recalc_csum = ipv6_find_hdr(skb, &offset,
	277	NEXTHDR_ROUTING, NULL,
	278	&flags) != NEXTHDR_ROUTING;
	279
	280	set_ipv6_addr(skb, ipv6_key->ipv6_proto, daddr,
	281	ipv6_key->ipv6_dst, recalc_csum);
	282	}
	283
	284	set_ipv6_tc(nh, ipv6_key->ipv6_tclass);
	285	set_ipv6_fl(nh, ntohl(ipv6_key->ipv6_label));
	286	nh->hop_limit = ipv6_key->ipv6_hlimit;
	287
	288	return 0;
	289	}
	290
4edb9ae9 PS	291	/* Must follow make_writable() since that can move the skb data. */
	292	static void set_tp_port(struct sk_buff skb, __be16 port,
	293	__be16 new_port, __sum16 *check)
959a2ecd	294	{
4edb9ae9 PS	295	inet_proto_csum_replace2(check, skb, *port, new_port, 0);
	296	*port = new_port;
	297	skb_clear_rxhash(skb);
	298	}
10db8b20	299
55ce87bc JG	300	static void set_udp_port(struct sk_buff skb, __be16 port, __be16 new_port)
	301	{
	302	struct udphdr *uh = udp_hdr(skb);
	303
	304	if (uh->check && get_ip_summed(skb) != OVS_CSUM_PARTIAL) {
	305	set_tp_port(skb, port, new_port, &uh->check);
	306
	307	if (!uh->check)
	308	uh->check = CSUM_MANGLED_0;
	309	} else {
	310	*port = new_port;
	311	skb_clear_rxhash(skb);
	312	}
	313	}
	314
	315	static int set_udp(struct sk_buff skb, const struct ovs_key_udp udp_port_key)
4edb9ae9 PS	316	{
	317	struct udphdr *uh;
	318	int err;
10db8b20	319
4edb9ae9 PS	320	err = make_writable(skb, skb_transport_offset(skb) +
4edb9ae9 PS	321	sizeof(struct udphdr));
10db8b20 JG	322	if (unlikely(err))
	323	return err;
	324
4edb9ae9 PS	325	uh = udp_hdr(skb);
4edb9ae9 PS	326	if (udp_port_key->udp_src != uh->source)
55ce87bc	327	set_udp_port(skb, &uh->source, udp_port_key->udp_src);
4edb9ae9 PS	328
4edb9ae9 PS	329	if (udp_port_key->udp_dst != uh->dest)
55ce87bc	330	set_udp_port(skb, &uh->dest, udp_port_key->udp_dst);
10db8b20 JG	331
10db8b20 JG	332	return 0;
959a2ecd JP	333	}
959a2ecd JP	334
55ce87bc	335	static int set_tcp(struct sk_buff skb, const struct ovs_key_tcp tcp_port_key)
064af421	336	{
4edb9ae9	337	struct tcphdr *th;
10db8b20	338	int err;
064af421	339
10db8b20 JG	340	err = make_writable(skb, skb_transport_offset(skb) +
	341	sizeof(struct tcphdr));
	342	if (unlikely(err))
	343	return err;
ca78c6b6	344
4edb9ae9 PS	345	th = tcp_hdr(skb);
	346	if (tcp_port_key->tcp_src != th->source)
	347	set_tp_port(skb, &th->source, tcp_port_key->tcp_src, &th->check);
064af421	348
4edb9ae9 PS	349	if (tcp_port_key->tcp_dst != th->dest)
4edb9ae9 PS	350	set_tp_port(skb, &th->dest, tcp_port_key->tcp_dst, &th->check);
ca78c6b6	351
10db8b20	352	return 0;
064af421 BP	353	}
064af421 BP	354
f15c8639	355	static int do_output(struct datapath dp, struct sk_buff skb, int out_port)
064af421	356	{
f15c8639	357	struct vport *vport;
064af421	358
f15c8639 JG	359	if (unlikely(!skb))
f15c8639 JG	360	return -ENOMEM;
064af421	361
95b1d73a	362	vport = ovs_vport_rcu(dp, out_port);
f15c8639 JG	363	if (unlikely(!vport)) {
	364	kfree_skb(skb);
	365	return -ENODEV;
	366	}
064af421	367
850b6b3b	368	ovs_vport_send(vport, skb);
f15c8639	369	return 0;
064af421 BP	370	}
064af421 BP	371
98403001 BP	372	static int output_userspace(struct datapath dp, struct sk_buff skb,
98403001 BP	373	const struct nlattr *attr)
064af421	374	{
856081f6	375	struct dp_upcall_info upcall;
98403001 BP	376	const struct nlattr *a;
98403001 BP	377	int rem;
856081f6	378
df2c07f4	379	upcall.cmd = OVS_PACKET_CMD_ACTION;
a4af2475	380	upcall.key = &OVS_CB(skb)->flow->key;
98403001	381	upcall.userdata = NULL;
28aea917	382	upcall.portid = 0;
98403001 BP	383
	384	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
	385	a = nla_next(a, &rem)) {
	386	switch (nla_type(a)) {
	387	case OVS_USERSPACE_ATTR_USERDATA:
	388	upcall.userdata = a;
	389	break;
	390
	391	case OVS_USERSPACE_ATTR_PID:
28aea917	392	upcall.portid = nla_get_u32(a);
98403001 BP	393	break;
	394	}
	395	}
	396
850b6b3b	397	return ovs_dp_upcall(dp, skb, &upcall);
064af421 BP	398	}
064af421 BP	399
6ff686f2	400	static int sample(struct datapath dp, struct sk_buff skb,
85c9de19	401	const struct nlattr *attr)
6ff686f2 PS	402	{
	403	const struct nlattr *acts_list = NULL;
	404	const struct nlattr *a;
	405	int rem;
	406
	407	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
	408	a = nla_next(a, &rem)) {
	409	switch (nla_type(a)) {
	410	case OVS_SAMPLE_ATTR_PROBABILITY:
	411	if (net_random() >= nla_get_u32(a))
	412	return 0;
	413	break;
	414
	415	case OVS_SAMPLE_ATTR_ACTIONS:
	416	acts_list = a;
	417	break;
	418	}
	419	}
	420
	421	return do_execute_actions(dp, skb, nla_data(acts_list),
85c9de19	422	nla_len(acts_list), true);
6ff686f2 PS	423	}
6ff686f2 PS	424
4edb9ae9	425	static int execute_set_action(struct sk_buff *skb,
85c9de19	426	const struct nlattr *nested_attr)
4edb9ae9	427	{
15c39847	428	int err = 0;
4edb9ae9 PS	429
4edb9ae9 PS	430	switch (nla_type(nested_attr)) {
abff858b PS	431	case OVS_KEY_ATTR_PRIORITY:
	432	skb->priority = nla_get_u32(nested_attr);
	433	break;
	434
72e8bf28 AA	435	case OVS_KEY_ATTR_SKB_MARK:
	436	skb_set_mark(skb, nla_get_u32(nested_attr));
	437	break;
	438
356af50b KM	439	case OVS_KEY_ATTR_IPV4_TUNNEL:
356af50b KM	440	OVS_CB(skb)->tun_key = nla_data(nested_attr);
4edb9ae9 PS	441	break;
	442
	443	case OVS_KEY_ATTR_ETHERNET:
	444	err = set_eth_addr(skb, nla_data(nested_attr));
	445	break;
	446
	447	case OVS_KEY_ATTR_IPV4:
	448	err = set_ipv4(skb, nla_data(nested_attr));
	449	break;
	450
bc7a5acd AA	451	case OVS_KEY_ATTR_IPV6:
	452	err = set_ipv6(skb, nla_data(nested_attr));
	453	break;
	454
4edb9ae9	455	case OVS_KEY_ATTR_TCP:
55ce87bc	456	err = set_tcp(skb, nla_data(nested_attr));
4edb9ae9 PS	457	break;
	458
	459	case OVS_KEY_ATTR_UDP:
55ce87bc	460	err = set_udp(skb, nla_data(nested_attr));
4edb9ae9 PS	461	break;
4edb9ae9 PS	462	}
15c39847	463
4edb9ae9 PS	464	return err;
	465	}
	466
064af421	467	/* Execute a list of actions against 'skb'. */
871dfe07	468	static int do_execute_actions(struct datapath dp, struct sk_buff skb,
85c9de19	469	const struct nlattr *attr, int len, bool keep_skb)
064af421 BP	470	{
	471	/* Every output action needs a separate clone of 'skb', but the common
	472	* case is just a single output action, so that doing a clone and
	473	* then freeing the original skbuff is wasteful. So the following code
	474	* is slightly obscure just to avoid that. */
	475	int prev_port = -1;
cdee00fd	476	const struct nlattr *a;
10db8b20	477	int rem;
72b06300	478
6ff686f2	479	for (a = attr, rem = len; rem > 0;
a4af2475	480	a = nla_next(a, &rem)) {
10db8b20 JG	481	int err = 0;
10db8b20 JG	482
064af421	483	if (prev_port != -1) {
7956695a	484	do_output(dp, skb_clone(skb, GFP_ATOMIC), prev_port);
064af421 BP	485	prev_port = -1;
	486	}
	487
cdee00fd	488	switch (nla_type(a)) {
df2c07f4	489	case OVS_ACTION_ATTR_OUTPUT:
cdee00fd	490	prev_port = nla_get_u32(a);
064af421 BP	491	break;
064af421 BP	492
df2c07f4	493	case OVS_ACTION_ATTR_USERSPACE:
98403001	494	output_userspace(dp, skb, a);
064af421 BP	495	break;
064af421 BP	496
fea393b1 BP	497	case OVS_ACTION_ATTR_PUSH_VLAN:
fea393b1 BP	498	err = push_vlan(skb, nla_data(a));
4edb9ae9	499	if (unlikely(err)) /* skb already freed. */
d9065a90	500	return err;
064af421 BP	501	break;
064af421 BP	502
fea393b1	503	case OVS_ACTION_ATTR_POP_VLAN:
d9065a90	504	err = pop_vlan(skb);
064af421 BP	505	break;
064af421 BP	506
4edb9ae9	507	case OVS_ACTION_ATTR_SET:
85c9de19	508	err = execute_set_action(skb, nla_data(a));
064af421	509	break;
c1c9c9c4	510
6ff686f2	511	case OVS_ACTION_ATTR_SAMPLE:
85c9de19	512	err = sample(dp, skb, a);
6ff686f2	513	break;
6ff686f2	514	}
15c39847	515
10db8b20 JG	516	if (unlikely(err)) {
	517	kfree_skb(skb);
	518	return err;
	519	}
064af421	520	}
6c222e55	521
6ff686f2 PS	522	if (prev_port != -1) {
	523	if (keep_skb)
	524	skb = skb_clone(skb, GFP_ATOMIC);
	525
064af421	526	do_output(dp, skb, prev_port);
6ff686f2	527	} else if (!keep_skb)
5b95ab0e	528	consume_skb(skb);
10db8b20	529
a5225dd6	530	return 0;
064af421	531	}
871dfe07	532
e9141eec PS	533	/* We limit the number of times that we pass into execute_actions()
	534	* to avoid blowing out the stack in the event that we have a loop. */
	535	#define MAX_LOOPS 5
	536
	537	struct loop_counter {
	538	u8 count; /* Count. */
	539	bool looping; /* Loop detected? */
	540	};
	541
	542	static DEFINE_PER_CPU(struct loop_counter, loop_counters);
	543
	544	static int loop_suppress(struct datapath dp, struct sw_flow_actions actions)
	545	{
	546	if (net_ratelimit())
	547	pr_warn("%s: flow looped %d times, dropping\n",
850b6b3b	548	ovs_dp_name(dp), MAX_LOOPS);
e9141eec PS	549	actions->actions_len = 0;
	550	return -ELOOP;
	551	}
	552
871dfe07	553	/* Execute a list of actions against 'skb'. */
850b6b3b	554	int ovs_execute_actions(struct datapath dp, struct sk_buff skb)
871dfe07	555	{
a4af2475 BP	556	struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts);
	557	struct loop_counter *loop;
	558	int error;
	559
	560	/* Check whether we've looped too much. */
e9141eec	561	loop = &__get_cpu_var(loop_counters);
a4af2475 BP	562	if (unlikely(++loop->count > MAX_LOOPS))
	563	loop->looping = true;
	564	if (unlikely(loop->looping)) {
	565	error = loop_suppress(dp, acts);
	566	kfree_skb(skb);
	567	goto out_loop;
	568	}
871dfe07	569
356af50b	570	OVS_CB(skb)->tun_key = NULL;
6ff686f2	571	error = do_execute_actions(dp, skb, acts->actions,
85c9de19	572	acts->actions_len, false);
a4af2475 BP	573
	574	/* Check whether sub-actions looped too much. */
	575	if (unlikely(loop->looping))
	576	error = loop_suppress(dp, acts);
	577
	578	out_loop:
	579	/* Decrement loop counter. */
	580	if (!--loop->count)
	581	loop->looping = false;
871dfe07	582
a4af2475	583	return error;
871dfe07	584	}