[deliverable/linux.git] / net / openvswitch / actions.c

/*
 * Copyright (c) 2007-2013 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/sctp.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 <net/sctp/checksum.h>

#include "datapath.h"
#include "vport.h"

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

static int make_writable(struct sk_buff *skb, int write_len)
{
	if (!pskb_may_pull(skb, write_len))
		return -ENOMEM;

	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 (skb->ip_summed == CHECKSUM_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;
	if (skb_network_offset(skb) < ETH_HLEN)
		skb_set_network_header(skb, ETH_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))) {
		skb->vlan_tci = 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, htons(ETH_P_8021Q), 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, skb->vlan_proto, current_tag))
			return -ENOMEM;

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

	}
	__vlan_hwaccel_put_tag(skb, vlan->vlan_tpid, 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;

	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);

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

	ovs_skb_postpush_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);

	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 || skb->ip_summed == CHECKSUM_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_hash(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 || skb->ip_summed == CHECKSUM_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_hash(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 = IP6_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_hash(skb);
}

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

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

		if (!uh->check)
			uh->check = CSUM_MANGLED_0;
	} else {
		*port = new_port;
		skb_clear_hash(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 set_sctp(struct sk_buff *skb,
		     const struct ovs_key_sctp *sctp_port_key)
{
	struct sctphdr *sh;
	int err;
	unsigned int sctphoff = skb_transport_offset(skb);

	err = make_writable(skb, sctphoff + sizeof(struct sctphdr));
	if (unlikely(err))
		return err;

	sh = sctp_hdr(skb);
	if (sctp_port_key->sctp_src != sh->source ||
	    sctp_port_key->sctp_dst != sh->dest) {
		__le32 old_correct_csum, new_csum, old_csum;

		old_csum = sh->checksum;
		old_correct_csum = sctp_compute_cksum(skb, sctphoff);

		sh->source = sctp_port_key->sctp_src;
		sh->dest = sctp_port_key->sctp_dst;

		new_csum = sctp_compute_cksum(skb, sctphoff);

		/* Carry any checksum errors through. */
		sh->checksum = old_csum ^ old_correct_csum ^ new_csum;

		skb_clear_hash(skb);
	}

	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,
			    struct sw_flow_key *key, const struct nlattr *attr)
{
	struct dp_upcall_info upcall;
	const struct nlattr *a;
	int rem;

	upcall.cmd = OVS_PACKET_CMD_ACTION;
	upcall.key = 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 bool last_action(const struct nlattr *a, int rem)
{
	return a->nla_len == rem;
}

static int sample(struct datapath *dp, struct sk_buff *skb,
		  struct sw_flow_key *key, const struct nlattr *attr)
{
	const struct nlattr *acts_list = NULL;
	const struct nlattr *a;
	struct sk_buff *sample_skb;
	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 (prandom_u32() >= nla_get_u32(a))
				return 0;
			break;

		case OVS_SAMPLE_ATTR_ACTIONS:
			acts_list = a;
			break;
		}
	}

	rem = nla_len(acts_list);
	a = nla_data(acts_list);

	/* Actions list is either empty or only contains a single user-space
	 * action, the latter being a special case as it is the only known
	 * usage of the sample action.
	 * In these special cases don't clone the skb as there are no
	 * side-effects in the nested actions.
	 * Otherwise, clone in case the nested actions have side effects.
	 */
	if (likely(rem == 0 || (nla_type(a) == OVS_ACTION_ATTR_USERSPACE &&
				last_action(a, rem)))) {
		sample_skb = skb;
		skb_get(skb);
	} else {
		sample_skb = skb_clone(skb, GFP_ATOMIC);
		if (!sample_skb) /* Skip sample action when out of memory. */
			return 0;
	}

	/* Note that do_execute_actions() never consumes skb.
	 * In the case where skb has been cloned above it is the clone that
	 * is consumed.  Otherwise the skb_get(skb) call prevents
	 * consumption by do_execute_actions(). Thus, it is safe to simply
	 * return the error code and let the caller (also
	 * do_execute_actions()) free skb on error.
	 */
	return do_execute_actions(dp, sample_skb, key, a, rem);
}

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->mark = nla_get_u32(nested_attr);
		break;

	case OVS_KEY_ATTR_IPV4_TUNNEL:
		OVS_CB(skb)->egress_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;

	case OVS_KEY_ATTR_SCTP:
		err = set_sctp(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,
			      struct sw_flow_key *key,
			      const struct nlattr *attr, int len)
{
	/* 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, key, 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, key, a);
			if (unlikely(err)) /* skb already freed. */
				return err;
			break;
		}

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

	if (prev_port != -1)
		do_output(dp, skb, prev_port);
	else
		consume_skb(skb);

	return 0;
}

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

	return do_execute_actions(dp, skb, key,
				  acts->actions, acts->actions_len);
}
Commit	Line	Data
ccb1352e	1	/*
03f0d916	2	* Copyright (c) 2007-2013 Nicira, Inc.
ccb1352e JG	3	*
	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
	17	*/
	18
	19	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
	20
	21	#include <linux/skbuff.h>
	22	#include <linux/in.h>
	23	#include <linux/ip.h>
	24	#include <linux/openvswitch.h>
a175a723	25	#include <linux/sctp.h>
ccb1352e JG	26	#include <linux/tcp.h>
	27	#include <linux/udp.h>
	28	#include <linux/in6.h>
	29	#include <linux/if_arp.h>
	30	#include <linux/if_vlan.h>
	31	#include <net/ip.h>
3fdbd1ce	32	#include <net/ipv6.h>
ccb1352e JG	33	#include <net/checksum.h>
ccb1352e JG	34	#include <net/dsfield.h>
a175a723	35	#include <net/sctp/checksum.h>
ccb1352e JG	36
	37	#include "datapath.h"
	38	#include "vport.h"
	39
	40	static int do_execute_actions(struct datapath dp, struct sk_buff skb,
2ff3e4e4	41	struct sw_flow_key *key,
651887b0	42	const struct nlattr *attr, int len);
ccb1352e JG	43
	44	static int make_writable(struct sk_buff *skb, int write_len)
	45	{
2ba5af42 JB	46	if (!pskb_may_pull(skb, write_len))
	47	return -ENOMEM;
	48
ccb1352e JG	49	if (!skb_cloned(skb) \|\| skb_clone_writable(skb, write_len))
	50	return 0;
	51
	52	return pskb_expand_head(skb, 0, 0, GFP_ATOMIC);
	53	}
	54
39855b5b	55	/* remove VLAN header from packet and update csum accordingly. */
ccb1352e JG	56	static int __pop_vlan_tci(struct sk_buff skb, __be16 current_tci)
	57	{
	58	struct vlan_hdr *vhdr;
	59	int err;
	60
	61	err = make_writable(skb, VLAN_ETH_HLEN);
	62	if (unlikely(err))
	63	return err;
	64
	65	if (skb->ip_summed == CHECKSUM_COMPLETE)
	66	skb->csum = csum_sub(skb->csum, csum_partial(skb->data
7b024082	67	+ (2 * ETH_ALEN), VLAN_HLEN, 0));
ccb1352e JG	68
	69	vhdr = (struct vlan_hdr *)(skb->data + ETH_HLEN);
	70	*current_tci = vhdr->h_vlan_TCI;
	71
	72	memmove(skb->data + VLAN_HLEN, skb->data, 2 * ETH_ALEN);
	73	__skb_pull(skb, VLAN_HLEN);
	74
	75	vlan_set_encap_proto(skb, vhdr);
	76	skb->mac_header += VLAN_HLEN;
2ba5af42 JB	77	if (skb_network_offset(skb) < ETH_HLEN)
2ba5af42 JB	78	skb_set_network_header(skb, ETH_HLEN);
ccb1352e JG	79	skb_reset_mac_len(skb);
	80
	81	return 0;
	82	}
	83
	84	static int pop_vlan(struct sk_buff *skb)
	85	{
	86	__be16 tci;
	87	int err;
	88
	89	if (likely(vlan_tx_tag_present(skb))) {
	90	skb->vlan_tci = 0;
	91	} else {
	92	if (unlikely(skb->protocol != htons(ETH_P_8021Q) \|\|
	93	skb->len < VLAN_ETH_HLEN))
	94	return 0;
	95
	96	err = __pop_vlan_tci(skb, &tci);
	97	if (err)
	98	return err;
	99	}
	100	/* move next vlan tag to hw accel tag */
	101	if (likely(skb->protocol != htons(ETH_P_8021Q) \|\|
	102	skb->len < VLAN_ETH_HLEN))
	103	return 0;
	104
	105	err = __pop_vlan_tci(skb, &tci);
	106	if (unlikely(err))
	107	return err;
	108
86a9bad3	109	__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ntohs(tci));
ccb1352e JG	110	return 0;
	111	}
	112
	113	static int push_vlan(struct sk_buff skb, const struct ovs_action_push_vlan vlan)
	114	{
	115	if (unlikely(vlan_tx_tag_present(skb))) {
	116	u16 current_tag;
	117
	118	/* push down current VLAN tag */
	119	current_tag = vlan_tx_tag_get(skb);
	120
86a9bad3	121	if (!__vlan_put_tag(skb, skb->vlan_proto, current_tag))
ccb1352e JG	122	return -ENOMEM;
	123
	124	if (skb->ip_summed == CHECKSUM_COMPLETE)
	125	skb->csum = csum_add(skb->csum, csum_partial(skb->data
7b024082	126	+ (2 * ETH_ALEN), VLAN_HLEN, 0));
ccb1352e JG	127
ccb1352e JG	128	}
86a9bad3	129	__vlan_hwaccel_put_tag(skb, vlan->vlan_tpid, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT);
ccb1352e JG	130	return 0;
	131	}
	132
	133	static int set_eth_addr(struct sk_buff *skb,
	134	const struct ovs_key_ethernet *eth_key)
	135	{
	136	int err;
	137	err = make_writable(skb, ETH_HLEN);
	138	if (unlikely(err))
	139	return err;
	140
b34df5e8 PS	141	skb_postpull_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);
b34df5e8 PS	142
8c63ff09 JP	143	ether_addr_copy(eth_hdr(skb)->h_source, eth_key->eth_src);
8c63ff09 JP	144	ether_addr_copy(eth_hdr(skb)->h_dest, eth_key->eth_dst);
ccb1352e	145
b34df5e8 PS	146	ovs_skb_postpush_rcsum(skb, eth_hdr(skb), ETH_ALEN * 2);
b34df5e8 PS	147
ccb1352e JG	148	return 0;
	149	}
	150
	151	static void set_ip_addr(struct sk_buff skb, struct iphdr nh,
	152	__be32 *addr, __be32 new_addr)
	153	{
	154	int transport_len = skb->len - skb_transport_offset(skb);
	155
	156	if (nh->protocol == IPPROTO_TCP) {
	157	if (likely(transport_len >= sizeof(struct tcphdr)))
	158	inet_proto_csum_replace4(&tcp_hdr(skb)->check, skb,
	159	*addr, new_addr, 1);
	160	} else if (nh->protocol == IPPROTO_UDP) {
81e5d41d JG	161	if (likely(transport_len >= sizeof(struct udphdr))) {
	162	struct udphdr *uh = udp_hdr(skb);
	163
	164	if (uh->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	165	inet_proto_csum_replace4(&uh->check, skb,
	166	*addr, new_addr, 1);
	167	if (!uh->check)
	168	uh->check = CSUM_MANGLED_0;
	169	}
	170	}
ccb1352e JG	171	}
	172
	173	csum_replace4(&nh->check, *addr, new_addr);
7539fadc	174	skb_clear_hash(skb);
ccb1352e JG	175	*addr = new_addr;
	176	}
	177
3fdbd1ce AA	178	static void update_ipv6_checksum(struct sk_buff *skb, u8 l4_proto,
	179	__be32 addr[4], const __be32 new_addr[4])
	180	{
	181	int transport_len = skb->len - skb_transport_offset(skb);
	182
	183	if (l4_proto == IPPROTO_TCP) {
	184	if (likely(transport_len >= sizeof(struct tcphdr)))
	185	inet_proto_csum_replace16(&tcp_hdr(skb)->check, skb,
	186	addr, new_addr, 1);
	187	} else if (l4_proto == IPPROTO_UDP) {
	188	if (likely(transport_len >= sizeof(struct udphdr))) {
	189	struct udphdr *uh = udp_hdr(skb);
	190
	191	if (uh->check \|\| skb->ip_summed == CHECKSUM_PARTIAL) {
	192	inet_proto_csum_replace16(&uh->check, skb,
	193	addr, new_addr, 1);
	194	if (!uh->check)
	195	uh->check = CSUM_MANGLED_0;
	196	}
	197	}
	198	}
	199	}
	200
	201	static void set_ipv6_addr(struct sk_buff *skb, u8 l4_proto,
	202	__be32 addr[4], const __be32 new_addr[4],
	203	bool recalculate_csum)
	204	{
	205	if (recalculate_csum)
	206	update_ipv6_checksum(skb, l4_proto, addr, new_addr);
	207
7539fadc	208	skb_clear_hash(skb);
3fdbd1ce AA	209	memcpy(addr, new_addr, sizeof(__be32[4]));
	210	}
	211
	212	static void set_ipv6_tc(struct ipv6hdr *nh, u8 tc)
	213	{
	214	nh->priority = tc >> 4;
	215	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0x0F) \| ((tc & 0x0F) << 4);
	216	}
	217
	218	static void set_ipv6_fl(struct ipv6hdr *nh, u32 fl)
	219	{
	220	nh->flow_lbl[0] = (nh->flow_lbl[0] & 0xF0) \| (fl & 0x000F0000) >> 16;
	221	nh->flow_lbl[1] = (fl & 0x0000FF00) >> 8;
	222	nh->flow_lbl[2] = fl & 0x000000FF;
	223	}
	224
ccb1352e JG	225	static void set_ip_ttl(struct sk_buff skb, struct iphdr nh, u8 new_ttl)
	226	{
	227	csum_replace2(&nh->check, htons(nh->ttl << 8), htons(new_ttl << 8));
	228	nh->ttl = new_ttl;
	229	}
	230
	231	static int set_ipv4(struct sk_buff skb, const struct ovs_key_ipv4 ipv4_key)
	232	{
	233	struct iphdr *nh;
	234	int err;
	235
	236	err = make_writable(skb, skb_network_offset(skb) +
	237	sizeof(struct iphdr));
	238	if (unlikely(err))
	239	return err;
	240
	241	nh = ip_hdr(skb);
	242
	243	if (ipv4_key->ipv4_src != nh->saddr)
	244	set_ip_addr(skb, nh, &nh->saddr, ipv4_key->ipv4_src);
	245
	246	if (ipv4_key->ipv4_dst != nh->daddr)
	247	set_ip_addr(skb, nh, &nh->daddr, ipv4_key->ipv4_dst);
	248
	249	if (ipv4_key->ipv4_tos != nh->tos)
	250	ipv4_change_dsfield(nh, 0, ipv4_key->ipv4_tos);
	251
	252	if (ipv4_key->ipv4_ttl != nh->ttl)
	253	set_ip_ttl(skb, nh, ipv4_key->ipv4_ttl);
	254
	255	return 0;
	256	}
	257
3fdbd1ce AA	258	static int set_ipv6(struct sk_buff skb, const struct ovs_key_ipv6 ipv6_key)
	259	{
	260	struct ipv6hdr *nh;
	261	int err;
	262	__be32 *saddr;
	263	__be32 *daddr;
	264
	265	err = make_writable(skb, skb_network_offset(skb) +
	266	sizeof(struct ipv6hdr));
	267	if (unlikely(err))
	268	return err;
	269
	270	nh = ipv6_hdr(skb);
	271	saddr = (__be32 *)&nh->saddr;
	272	daddr = (__be32 *)&nh->daddr;
	273
	274	if (memcmp(ipv6_key->ipv6_src, saddr, sizeof(ipv6_key->ipv6_src)))
	275	set_ipv6_addr(skb, ipv6_key->ipv6_proto, saddr,
	276	ipv6_key->ipv6_src, true);
	277
	278	if (memcmp(ipv6_key->ipv6_dst, daddr, sizeof(ipv6_key->ipv6_dst))) {
	279	unsigned int offset = 0;
	280	int flags = IP6_FH_F_SKIP_RH;
	281	bool recalc_csum = true;
	282
	283	if (ipv6_ext_hdr(nh->nexthdr))
	284	recalc_csum = ipv6_find_hdr(skb, &offset,
	285	NEXTHDR_ROUTING, NULL,
	286	&flags) != NEXTHDR_ROUTING;
	287
	288	set_ipv6_addr(skb, ipv6_key->ipv6_proto, daddr,
	289	ipv6_key->ipv6_dst, recalc_csum);
	290	}
	291
	292	set_ipv6_tc(nh, ipv6_key->ipv6_tclass);
	293	set_ipv6_fl(nh, ntohl(ipv6_key->ipv6_label));
	294	nh->hop_limit = ipv6_key->ipv6_hlimit;
	295
	296	return 0;
	297	}
	298
ccb1352e JG	299	/* Must follow make_writable() since that can move the skb data. */
	300	static void set_tp_port(struct sk_buff skb, __be16 port,
	301	__be16 new_port, __sum16 *check)
	302	{
	303	inet_proto_csum_replace2(check, skb, *port, new_port, 0);
	304	*port = new_port;
7539fadc	305	skb_clear_hash(skb);
ccb1352e JG	306	}
ccb1352e JG	307
81e5d41d JG	308	static void set_udp_port(struct sk_buff skb, __be16 port, __be16 new_port)
	309	{
	310	struct udphdr *uh = udp_hdr(skb);
	311
	312	if (uh->check && skb->ip_summed != CHECKSUM_PARTIAL) {
	313	set_tp_port(skb, port, new_port, &uh->check);
	314
	315	if (!uh->check)
	316	uh->check = CSUM_MANGLED_0;
	317	} else {
	318	*port = new_port;
7539fadc	319	skb_clear_hash(skb);
81e5d41d JG	320	}
	321	}
	322
	323	static int set_udp(struct sk_buff skb, const struct ovs_key_udp udp_port_key)
ccb1352e JG	324	{
	325	struct udphdr *uh;
	326	int err;
	327
	328	err = make_writable(skb, skb_transport_offset(skb) +
	329	sizeof(struct udphdr));
	330	if (unlikely(err))
	331	return err;
	332
	333	uh = udp_hdr(skb);
	334	if (udp_port_key->udp_src != uh->source)
81e5d41d	335	set_udp_port(skb, &uh->source, udp_port_key->udp_src);
ccb1352e JG	336
ccb1352e JG	337	if (udp_port_key->udp_dst != uh->dest)
81e5d41d	338	set_udp_port(skb, &uh->dest, udp_port_key->udp_dst);
ccb1352e JG	339
	340	return 0;
	341	}
	342
81e5d41d	343	static int set_tcp(struct sk_buff skb, const struct ovs_key_tcp tcp_port_key)
ccb1352e JG	344	{
	345	struct tcphdr *th;
	346	int err;
	347
	348	err = make_writable(skb, skb_transport_offset(skb) +
	349	sizeof(struct tcphdr));
	350	if (unlikely(err))
	351	return err;
	352
	353	th = tcp_hdr(skb);
	354	if (tcp_port_key->tcp_src != th->source)
	355	set_tp_port(skb, &th->source, tcp_port_key->tcp_src, &th->check);
	356
	357	if (tcp_port_key->tcp_dst != th->dest)
	358	set_tp_port(skb, &th->dest, tcp_port_key->tcp_dst, &th->check);
	359
	360	return 0;
	361	}
	362
a175a723 JS	363	static int set_sctp(struct sk_buff *skb,
	364	const struct ovs_key_sctp *sctp_port_key)
	365	{
	366	struct sctphdr *sh;
	367	int err;
	368	unsigned int sctphoff = skb_transport_offset(skb);
	369
	370	err = make_writable(skb, sctphoff + sizeof(struct sctphdr));
	371	if (unlikely(err))
	372	return err;
	373
	374	sh = sctp_hdr(skb);
	375	if (sctp_port_key->sctp_src != sh->source \|\|
	376	sctp_port_key->sctp_dst != sh->dest) {
	377	__le32 old_correct_csum, new_csum, old_csum;
	378
	379	old_csum = sh->checksum;
	380	old_correct_csum = sctp_compute_cksum(skb, sctphoff);
	381
	382	sh->source = sctp_port_key->sctp_src;
	383	sh->dest = sctp_port_key->sctp_dst;
	384
	385	new_csum = sctp_compute_cksum(skb, sctphoff);
	386
	387	/* Carry any checksum errors through. */
	388	sh->checksum = old_csum ^ old_correct_csum ^ new_csum;
	389
7539fadc	390	skb_clear_hash(skb);
a175a723 JS	391	}
	392
	393	return 0;
	394	}
	395
ccb1352e JG	396	static int do_output(struct datapath dp, struct sk_buff skb, int out_port)
	397	{
	398	struct vport *vport;
	399
	400	if (unlikely(!skb))
	401	return -ENOMEM;
	402
15eac2a7	403	vport = ovs_vport_rcu(dp, out_port);
ccb1352e JG	404	if (unlikely(!vport)) {
	405	kfree_skb(skb);
	406	return -ENODEV;
	407	}
	408
	409	ovs_vport_send(vport, skb);
	410	return 0;
	411	}
	412
	413	static int output_userspace(struct datapath dp, struct sk_buff skb,
2ff3e4e4	414	struct sw_flow_key key, const struct nlattr attr)
ccb1352e JG	415	{
	416	struct dp_upcall_info upcall;
	417	const struct nlattr *a;
	418	int rem;
	419
	420	upcall.cmd = OVS_PACKET_CMD_ACTION;
2ff3e4e4	421	upcall.key = key;
ccb1352e	422	upcall.userdata = NULL;
15e47304	423	upcall.portid = 0;
ccb1352e JG	424
	425	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
	426	a = nla_next(a, &rem)) {
	427	switch (nla_type(a)) {
	428	case OVS_USERSPACE_ATTR_USERDATA:
	429	upcall.userdata = a;
	430	break;
	431
	432	case OVS_USERSPACE_ATTR_PID:
15e47304	433	upcall.portid = nla_get_u32(a);
ccb1352e JG	434	break;
	435	}
	436	}
	437
	438	return ovs_dp_upcall(dp, skb, &upcall);
	439	}
	440
651887b0 SH	441	static bool last_action(const struct nlattr *a, int rem)
	442	{
	443	return a->nla_len == rem;
	444	}
	445
ccb1352e	446	static int sample(struct datapath dp, struct sk_buff skb,
2ff3e4e4	447	struct sw_flow_key key, const struct nlattr attr)
ccb1352e JG	448	{
	449	const struct nlattr *acts_list = NULL;
	450	const struct nlattr *a;
651887b0	451	struct sk_buff *sample_skb;
ccb1352e JG	452	int rem;
	453
	454	for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
	455	a = nla_next(a, &rem)) {
	456	switch (nla_type(a)) {
	457	case OVS_SAMPLE_ATTR_PROBABILITY:
63862b5b	458	if (prandom_u32() >= nla_get_u32(a))
ccb1352e JG	459	return 0;
	460	break;
	461
	462	case OVS_SAMPLE_ATTR_ACTIONS:
	463	acts_list = a;
	464	break;
	465	}
	466	}
	467
651887b0 SH	468	rem = nla_len(acts_list);
	469	a = nla_data(acts_list);
	470
	471	/* Actions list is either empty or only contains a single user-space
	472	* action, the latter being a special case as it is the only known
	473	* usage of the sample action.
	474	* In these special cases don't clone the skb as there are no
	475	* side-effects in the nested actions.
	476	* Otherwise, clone in case the nested actions have side effects.
	477	*/
	478	if (likely(rem == 0 \|\| (nla_type(a) == OVS_ACTION_ATTR_USERSPACE &&
	479	last_action(a, rem)))) {
	480	sample_skb = skb;
	481	skb_get(skb);
	482	} else {
	483	sample_skb = skb_clone(skb, GFP_ATOMIC);
d9e0ecb8 AZ	484	if (!sample_skb) /* Skip sample action when out of memory. */
d9e0ecb8 AZ	485	return 0;
651887b0 SH	486	}
	487
	488	/* Note that do_execute_actions() never consumes skb.
	489	* In the case where skb has been cloned above it is the clone that
	490	* is consumed. Otherwise the skb_get(skb) call prevents
	491	* consumption by do_execute_actions(). Thus, it is safe to simply
	492	* return the error code and let the caller (also
	493	* do_execute_actions()) free skb on error.
	494	*/
2ff3e4e4	495	return do_execute_actions(dp, sample_skb, key, a, rem);
ccb1352e JG	496	}
	497
	498	static int execute_set_action(struct sk_buff *skb,
	499	const struct nlattr *nested_attr)
	500	{
	501	int err = 0;
	502
	503	switch (nla_type(nested_attr)) {
	504	case OVS_KEY_ATTR_PRIORITY:
	505	skb->priority = nla_get_u32(nested_attr);
	506	break;
	507
39c7caeb AA	508	case OVS_KEY_ATTR_SKB_MARK:
	509	skb->mark = nla_get_u32(nested_attr);
	510	break;
	511
7d5437c7	512	case OVS_KEY_ATTR_IPV4_TUNNEL:
8c8b1b83	513	OVS_CB(skb)->egress_tun_key = nla_data(nested_attr);
7d5437c7 PS	514	break;
7d5437c7 PS	515
ccb1352e JG	516	case OVS_KEY_ATTR_ETHERNET:
	517	err = set_eth_addr(skb, nla_data(nested_attr));
	518	break;
	519
	520	case OVS_KEY_ATTR_IPV4:
	521	err = set_ipv4(skb, nla_data(nested_attr));
	522	break;
	523
3fdbd1ce AA	524	case OVS_KEY_ATTR_IPV6:
	525	err = set_ipv6(skb, nla_data(nested_attr));
	526	break;
	527
ccb1352e	528	case OVS_KEY_ATTR_TCP:
81e5d41d	529	err = set_tcp(skb, nla_data(nested_attr));
ccb1352e JG	530	break;
	531
	532	case OVS_KEY_ATTR_UDP:
81e5d41d	533	err = set_udp(skb, nla_data(nested_attr));
ccb1352e	534	break;
a175a723 JS	535
	536	case OVS_KEY_ATTR_SCTP:
	537	err = set_sctp(skb, nla_data(nested_attr));
	538	break;
ccb1352e JG	539	}
	540
	541	return err;
	542	}
	543
	544	/* Execute a list of actions against 'skb'. */
	545	static int do_execute_actions(struct datapath dp, struct sk_buff skb,
2ff3e4e4	546	struct sw_flow_key *key,
651887b0	547	const struct nlattr *attr, int len)
ccb1352e JG	548	{
	549	/* Every output action needs a separate clone of 'skb', but the common
	550	* case is just a single output action, so that doing a clone and
	551	* then freeing the original skbuff is wasteful. So the following code
	552	* is slightly obscure just to avoid that. */
	553	int prev_port = -1;
	554	const struct nlattr *a;
	555	int rem;
	556
	557	for (a = attr, rem = len; rem > 0;
	558	a = nla_next(a, &rem)) {
	559	int err = 0;
	560
	561	if (prev_port != -1) {
	562	do_output(dp, skb_clone(skb, GFP_ATOMIC), prev_port);
	563	prev_port = -1;
	564	}
	565
	566	switch (nla_type(a)) {
	567	case OVS_ACTION_ATTR_OUTPUT:
	568	prev_port = nla_get_u32(a);
	569	break;
	570
	571	case OVS_ACTION_ATTR_USERSPACE:
2ff3e4e4	572	output_userspace(dp, skb, key, a);
ccb1352e JG	573	break;
	574
	575	case OVS_ACTION_ATTR_PUSH_VLAN:
	576	err = push_vlan(skb, nla_data(a));
	577	if (unlikely(err)) /* skb already freed. */
	578	return err;
	579	break;
	580
	581	case OVS_ACTION_ATTR_POP_VLAN:
	582	err = pop_vlan(skb);
	583	break;
	584
	585	case OVS_ACTION_ATTR_SET:
	586	err = execute_set_action(skb, nla_data(a));
	587	break;
	588
	589	case OVS_ACTION_ATTR_SAMPLE:
2ff3e4e4	590	err = sample(dp, skb, key, a);
fe984c08 AZ	591	if (unlikely(err)) /* skb already freed. */
fe984c08 AZ	592	return err;
ccb1352e JG	593	break;
	594	}
	595
	596	if (unlikely(err)) {
	597	kfree_skb(skb);
	598	return err;
	599	}
	600	}
	601
651887b0	602	if (prev_port != -1)
ccb1352e	603	do_output(dp, skb, prev_port);
651887b0	604	else
ccb1352e JG	605	consume_skb(skb);
	606
	607	return 0;
	608	}
	609
	610	/* Execute a list of actions against 'skb'. */
2ff3e4e4 PS	611	int ovs_execute_actions(struct datapath dp, struct sk_buff skb,
2ff3e4e4 PS	612	struct sw_flow_key *key)
ccb1352e JG	613	{
	614	struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts);
	615
2ff3e4e4 PS	616	return do_execute_actions(dp, skb, key,
2ff3e4e4 PS	617	acts->actions, acts->actions_len);
ccb1352e	618	}