/*
- * Copyright (c) 2007-2012 Nicira, Inc.
+ * Copyright (c) 2007-2014 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
#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 <net/ipv6.h>
#include <net/checksum.h>
#include <net/dsfield.h>
+#include <net/sctp/checksum.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,
- struct ovs_key_ipv4_tunnel *tun_key, bool keep_skb);
+ const struct nlattr *attr, int len, bool keep_skb);
static int make_writable(struct sk_buff *skb, int write_len)
{
if (unlikely(err))
return err;
- if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
+ if (skb->ip_summed == CHECKSUM_COMPLETE)
skb->csum = csum_sub(skb->csum, csum_partial(skb->data
- + ETH_HLEN, VLAN_HLEN, 0));
+ + (2 * ETH_ALEN), VLAN_HLEN, 0));
vhdr = (struct vlan_hdr *)(skb->data + ETH_HLEN);
*current_tci = vhdr->h_vlan_TCI;
if (unlikely(err))
return err;
- __vlan_hwaccel_put_tag(skb, ntohs(tci));
+ __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ntohs(tci));
return 0;
}
/* push down current VLAN tag */
current_tag = vlan_tx_tag_get(skb);
- if (!__vlan_put_tag(skb, current_tag))
+ if (!__vlan_put_tag(skb, skb->vlan_proto, current_tag))
return -ENOMEM;
- if (get_ip_summed(skb) == OVS_CSUM_COMPLETE)
+ if (skb->ip_summed == CHECKSUM_COMPLETE)
skb->csum = csum_add(skb->csum, csum_partial(skb->data
- + ETH_HLEN, VLAN_HLEN, 0));
+ + (2 * ETH_ALEN), VLAN_HLEN, 0));
}
- __vlan_hwaccel_put_tag(skb, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT);
+ __vlan_hwaccel_put_tag(skb, vlan->vlan_tpid, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT);
return 0;
}
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);
+ 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;
}
if (likely(transport_len >= sizeof(struct udphdr))) {
struct udphdr *uh = udp_hdr(skb);
- if (uh->check ||
- get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
+ if (uh->check || skb->ip_summed == CHECKSUM_PARTIAL) {
inet_proto_csum_replace4(&uh->check, skb,
*addr, new_addr, 1);
if (!uh->check)
}
csum_replace4(&nh->check, *addr, new_addr);
- skb_clear_rxhash(skb);
+ skb_clear_hash(skb);
*addr = new_addr;
}
if (likely(transport_len >= sizeof(struct udphdr))) {
struct udphdr *uh = udp_hdr(skb);
- if (uh->check ||
- get_ip_summed(skb) == OVS_CSUM_PARTIAL) {
+ if (uh->check || skb->ip_summed == CHECKSUM_PARTIAL) {
inet_proto_csum_replace16(&uh->check, skb,
addr, new_addr, 1);
if (!uh->check)
if (recalculate_csum)
update_ipv6_checksum(skb, l4_proto, addr, new_addr);
- skb_clear_rxhash(skb);
+ skb_clear_hash(skb);
memcpy(addr, new_addr, sizeof(__be32[4]));
}
{
inet_proto_csum_replace2(check, skb, *port, new_port, 0);
*port = new_port;
- skb_clear_rxhash(skb);
+ 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 && get_ip_summed(skb) != OVS_CSUM_PARTIAL) {
+ 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_rxhash(skb);
+ skb_clear_hash(skb);
}
}
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;
const struct nlattr *a;
int rem;
+ BUG_ON(!OVS_CB(skb)->pkt_key);
+
upcall.cmd = OVS_PACKET_CMD_ACTION;
- upcall.key = &OVS_CB(skb)->flow->key;
+ upcall.key = OVS_CB(skb)->pkt_key;
upcall.userdata = NULL;
- upcall.pid = 0;
+ upcall.portid = 0;
for (a = nla_data(attr), rem = nla_len(attr); rem > 0;
a = nla_next(a, &rem)) {
break;
case OVS_USERSPACE_ATTR_PID:
- upcall.pid = nla_get_u32(a);
+ upcall.portid = nla_get_u32(a);
break;
}
}
}
static int sample(struct datapath *dp, struct sk_buff *skb,
- const struct nlattr *attr,
- struct ovs_key_ipv4_tunnel *tun_key)
+ const struct nlattr *attr)
{
const struct nlattr *acts_list = NULL;
const struct nlattr *a;
a = nla_next(a, &rem)) {
switch (nla_type(a)) {
case OVS_SAMPLE_ATTR_PROBABILITY:
- if (net_random() >= nla_get_u32(a))
+ if (prandom_u32() >= nla_get_u32(a))
return 0;
break;
}
return do_execute_actions(dp, skb, nla_data(acts_list),
- nla_len(acts_list), tun_key, true);
+ nla_len(acts_list), true);
+}
+
+static void execute_hash(struct sk_buff *skb, const struct nlattr *attr)
+{
+ struct sw_flow_key *key = OVS_CB(skb)->pkt_key;
+ struct ovs_action_hash *hash_act = nla_data(attr);
+ u32 hash = 0;
+
+ /* OVS_HASH_ALG_L4 is the only possible hash algorithm. */
+ hash = skb_get_hash(skb);
+ hash = jhash_1word(hash, hash_act->hash_basis);
+ if (!hash)
+ hash = 0x1;
+
+ key->ovs_flow_hash = hash;
}
static int execute_set_action(struct sk_buff *skb,
- const struct nlattr *nested_attr,
- struct ovs_key_ipv4_tunnel *tun_key)
+ const struct nlattr *nested_attr)
{
int err = 0;
break;
case OVS_KEY_ATTR_SKB_MARK:
- skb_set_mark(skb, nla_get_u32(nested_attr));
- break;
-
- case OVS_KEY_ATTR_TUN_ID:
- /* If we're only using the TUN_ID action, store the value in a
- * temporary instance of struct ovs_key_ipv4_tunnel on the stack.
- * If both IPV4_TUNNEL and TUN_ID are being used together we
- * can't write into the IPV4_TUNNEL action, so make a copy and
- * write into that version.
- */
- if (!OVS_CB(skb)->tun_key)
- memset(tun_key, 0, sizeof(*tun_key));
- else if (OVS_CB(skb)->tun_key != tun_key)
- memcpy(tun_key, OVS_CB(skb)->tun_key, sizeof(*tun_key));
- OVS_CB(skb)->tun_key = tun_key;
-
- OVS_CB(skb)->tun_key->tun_id = nla_get_be64(nested_attr);
+ skb->mark = nla_get_u32(nested_attr);
break;
case OVS_KEY_ATTR_IPV4_TUNNEL:
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;
}
+static int execute_recirc(struct datapath *dp, struct sk_buff *skb,
+ const struct nlattr *a)
+{
+ struct sw_flow_key recirc_key;
+ const struct vport *p = OVS_CB(skb)->input_vport;
+ uint32_t hash = OVS_CB(skb)->pkt_key->ovs_flow_hash;
+ int err;
+
+ err = ovs_flow_extract(skb, p->port_no, &recirc_key);
+ if (err)
+ return err;
+
+ recirc_key.ovs_flow_hash = hash;
+ recirc_key.recirc_id = nla_get_u32(a);
+
+ ovs_dp_process_packet_with_key(skb, &recirc_key, true);
+
+ return 0;
+}
+
/* 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,
- struct ovs_key_ipv4_tunnel *tun_key, bool keep_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
output_userspace(dp, skb, a);
break;
+ case OVS_ACTION_ATTR_HASH:
+ execute_hash(skb, a);
+ break;
+
case OVS_ACTION_ATTR_PUSH_VLAN:
err = push_vlan(skb, nla_data(a));
if (unlikely(err)) /* skb already freed. */
err = pop_vlan(skb);
break;
+ case OVS_ACTION_ATTR_RECIRC: {
+ struct sk_buff *recirc_skb;
+ const bool last_action = (a->nla_len == rem);
+
+ if (!last_action || keep_skb)
+ recirc_skb = skb_clone(skb, GFP_ATOMIC);
+ else
+ recirc_skb = skb;
+
+ err = execute_recirc(dp, recirc_skb, a);
+
+ if (last_action || err)
+ return err;
+
+ break;
+ }
+
case OVS_ACTION_ATTR_SET:
- err = execute_set_action(skb, nla_data(a), tun_key);
+ err = execute_set_action(skb, nla_data(a));
break;
case OVS_ACTION_ATTR_SAMPLE:
- err = sample(dp, skb, a, tun_key);
+ err = sample(dp, skb, a);
+ if (unlikely(err)) /* skb already freed. */
+ return err;
break;
}
}
/* 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
+ * to avoid blowing out the stack in the event that we have a loop.
+ *
+ * Each loop adds some (estimated) cost to the kernel stack.
+ * The loop terminates when the max cost is exceeded.
+ * */
+#define RECIRC_STACK_COST 1
+#define DEFAULT_STACK_COST 4
+/* Allow up to 4 regular services, and up to 3 recirculations */
+#define MAX_STACK_COST (DEFAULT_STACK_COST * 4 + RECIRC_STACK_COST * 3)
struct loop_counter {
- u8 count; /* Count. */
+ u8 stack_cost; /* loop stack cost. */
bool looping; /* Loop detected? */
};
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);
+ pr_warn("%s: flow loop detected, dropping\n",
+ ovs_dp_name(dp));
actions->actions_len = 0;
return -ELOOP;
}
/* Execute a list of actions against 'skb'. */
-int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb)
+int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb, bool recirc)
{
struct sw_flow_actions *acts = rcu_dereference(OVS_CB(skb)->flow->sf_acts);
+ const u8 stack_cost = recirc ? RECIRC_STACK_COST : DEFAULT_STACK_COST;
struct loop_counter *loop;
int error;
- struct ovs_key_ipv4_tunnel tun_key;
/* Check whether we've looped too much. */
loop = &__get_cpu_var(loop_counters);
- if (unlikely(++loop->count > MAX_LOOPS))
+ loop->stack_cost += stack_cost;
+ if (unlikely(loop->stack_cost > MAX_STACK_COST))
loop->looping = true;
if (unlikely(loop->looping)) {
error = loop_suppress(dp, acts);
OVS_CB(skb)->tun_key = NULL;
error = do_execute_actions(dp, skb, acts->actions,
- acts->actions_len, &tun_key, false);
+ 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)
+ /* Decrement loop stack cost. */
+ loop->stack_cost -= stack_cost;
+ if (!loop->stack_cost)
loop->looping = false;
return error;
git://git.onelab.eu / sliver-openvswitch.git / blobdiff