/*
- * Copyright (c) 2007-2012 Nicira, Inc.
+ * 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
#include <linux/openvswitch.h>
#include <linux/rculist.h>
#include <linux/dmi.h>
+#include <linux/genetlink.h>
+#include <net/genetlink.h>
#include <net/genetlink.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>
#include "checksum.h"
#include "datapath.h"
#include "flow.h"
-#include "genl_exec.h"
#include "vlan.h"
-#include "tunnel.h"
#include "vport-internal_dev.h"
-
-#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18) || \
- LINUX_VERSION_CODE >= KERNEL_VERSION(3,9,0)
-#error Kernels before 2.6.18 or after 3.8 are not supported by this version of Open vSwitch.
-#endif
+#include "vport-netdev.h"
#define REHASH_FLOW_INTERVAL (10 * 60 * HZ)
static void rehash_flow_table(struct work_struct *work);
int ovs_net_id __read_mostly;
+static void ovs_notify(struct sk_buff *skb, struct genl_info *info,
+ struct genl_multicast_group *grp)
+{
+ genl_notify(skb, genl_info_net(info), info->snd_portid,
+ grp->id, info->nlhdr, GFP_KERNEL);
+}
+
/**
* DOC: Locking:
*
- * Writes to device state (add/remove datapath, port, set operations on vports,
- * etc.) are protected by RTNL.
- *
- * Writes to other state (flow table modifications, set miscellaneous datapath
- * parameters, etc.) are protected by genl_mutex. The RTNL lock nests inside
- * genl_mutex.
+ * All writes e.g. Writes to device state (add/remove datapath, port, set
+ * operations on vports, etc.), Writes to other state (flow table
+ * modifications, set miscellaneous datapath parameters, etc.) are protected
+ * by ovs_lock.
*
* Reads are protected by RCU.
*
* There are a few special cases (mostly stats) that have their own
* synchronization but they nest under all of above and don't interact with
* each other.
+ *
+ * The RTNL lock nests inside ovs_mutex.
*/
+static DEFINE_MUTEX(ovs_mutex);
+
+void ovs_lock(void)
+{
+ mutex_lock(&ovs_mutex);
+}
+
+void ovs_unlock(void)
+{
+ mutex_unlock(&ovs_mutex);
+}
+
+#ifdef CONFIG_LOCKDEP
+int lockdep_ovsl_is_held(void)
+{
+ if (debug_locks)
+ return lockdep_is_held(&ovs_mutex);
+ else
+ return 1;
+}
+#endif
+
static struct vport *new_vport(const struct vport_parms *);
static int queue_gso_packets(struct net *, int dp_ifindex, struct sk_buff *,
const struct dp_upcall_info *);
struct sk_buff *,
const struct dp_upcall_info *);
-/* Must be called with rcu_read_lock, genl_mutex, or RTNL lock. */
+/* Must be called with rcu_read_lock or ovs_mutex. */
static struct datapath *get_dp(struct net *net, int dp_ifindex)
{
struct datapath *dp = NULL;
return dp;
}
-/* Must be called with rcu_read_lock or RTNL lock. */
+/* Must be called with rcu_read_lock or ovs_mutex. */
const char *ovs_dp_name(const struct datapath *dp)
{
- struct vport *vport = ovs_vport_rtnl_rcu(dp, OVSP_LOCAL);
+ struct vport *vport = ovs_vport_ovsl_rcu(dp, OVSP_LOCAL);
return vport->ops->get_name(vport);
}
local = ovs_vport_rcu(dp, OVSP_LOCAL);
if (local)
- ifindex = local->ops->get_ifindex(local);
+ ifindex = netdev_vport_priv(local)->dev->ifindex;
else
ifindex = 0;
{
struct datapath *dp = container_of(rcu, struct datapath, rcu);
- ovs_flow_tbl_destroy((__force struct flow_table *)dp->table);
+ ovs_flow_tbl_destroy((__force struct flow_table *)dp->table, false);
free_percpu(dp->stats_percpu);
release_net(ovs_dp_get_net(dp));
kfree(dp->ports);
return NULL;
}
-/* Called with RTNL lock and genl_lock. */
+/* Called with ovs_mutex. */
static struct vport *new_vport(const struct vport_parms *parms)
{
struct vport *vport;
return vport;
}
-/* Called with RTNL lock. */
void ovs_dp_detach_port(struct vport *p)
{
- ASSERT_RTNL();
+ ASSERT_OVSL();
/* First drop references to device. */
hlist_del_rcu(&p->dp_hash_node);
struct datapath *dp = p->dp;
struct sw_flow *flow;
struct dp_stats_percpu *stats;
+ struct sw_flow_key key;
u64 *stats_counter;
int error;
stats = this_cpu_ptr(dp->stats_percpu);
- if (!OVS_CB(skb)->flow) {
- struct sw_flow_key key;
- int key_len;
-
- /* Extract flow from 'skb' into 'key'. */
- error = ovs_flow_extract(skb, p->port_no, &key, &key_len);
- if (unlikely(error)) {
- kfree_skb(skb);
- return;
- }
-
- /* Look up flow. */
- flow = ovs_flow_tbl_lookup(rcu_dereference(dp->table),
- &key, key_len);
- if (unlikely(!flow)) {
- struct dp_upcall_info upcall;
-
- upcall.cmd = OVS_PACKET_CMD_MISS;
- upcall.key = &key;
- upcall.userdata = NULL;
- upcall.portid = p->upcall_portid;
- ovs_dp_upcall(dp, skb, &upcall);
- consume_skb(skb);
- stats_counter = &stats->n_missed;
- goto out;
- }
+ /* Extract flow from 'skb' into 'key'. */
+ error = ovs_flow_extract(skb, p->port_no, &key);
+ if (unlikely(error)) {
+ kfree_skb(skb);
+ return;
+ }
- OVS_CB(skb)->flow = flow;
+ /* Look up flow. */
+ flow = ovs_flow_lookup(rcu_dereference(dp->table), &key);
+ if (unlikely(!flow)) {
+ struct dp_upcall_info upcall;
+
+ upcall.cmd = OVS_PACKET_CMD_MISS;
+ upcall.key = &key;
+ upcall.userdata = NULL;
+ upcall.portid = p->upcall_portid;
+ ovs_dp_upcall(dp, skb, &upcall);
+ consume_skb(skb);
+ stats_counter = &stats->n_missed;
+ goto out;
}
+ OVS_CB(skb)->flow = flow;
+ OVS_CB(skb)->pkt_key = &key;
+
stats_counter = &stats->n_hit;
ovs_flow_used(OVS_CB(skb)->flow, skb);
ovs_execute_actions(dp, skb);
.version = OVS_PACKET_VERSION,
.maxattr = OVS_PACKET_ATTR_MAX,
SET_NETNSOK
+ SET_PARALLEL_OPS
};
int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb,
return err;
}
+static size_t key_attr_size(void)
+{
+ return nla_total_size(4) /* OVS_KEY_ATTR_PRIORITY */
+ + nla_total_size(0) /* OVS_KEY_ATTR_TUNNEL */
+ + nla_total_size(8) /* OVS_TUNNEL_KEY_ATTR_ID */
+ + nla_total_size(4) /* OVS_TUNNEL_KEY_ATTR_IPV4_SRC */
+ + nla_total_size(4) /* OVS_TUNNEL_KEY_ATTR_IPV4_DST */
+ + nla_total_size(1) /* OVS_TUNNEL_KEY_ATTR_TOS */
+ + nla_total_size(1) /* OVS_TUNNEL_KEY_ATTR_TTL */
+ + nla_total_size(0) /* OVS_TUNNEL_KEY_ATTR_DONT_FRAGMENT */
+ + nla_total_size(0) /* OVS_TUNNEL_KEY_ATTR_CSUM */
+ + nla_total_size(4) /* OVS_KEY_ATTR_IN_PORT */
+ + nla_total_size(4) /* OVS_KEY_ATTR_SKB_MARK */
+ + nla_total_size(12) /* OVS_KEY_ATTR_ETHERNET */
+ + nla_total_size(2) /* OVS_KEY_ATTR_ETHERTYPE */
+ + nla_total_size(4) /* OVS_KEY_ATTR_8021Q */
+ + nla_total_size(0) /* OVS_KEY_ATTR_ENCAP */
+ + nla_total_size(2) /* OVS_KEY_ATTR_ETHERTYPE */
+ + nla_total_size(40) /* OVS_KEY_ATTR_IPV6 */
+ + nla_total_size(2) /* OVS_KEY_ATTR_ICMPV6 */
+ + nla_total_size(28); /* OVS_KEY_ATTR_ND */
+}
+
+static size_t upcall_msg_size(const struct sk_buff *skb,
+ const struct nlattr *userdata)
+{
+ size_t size = NLMSG_ALIGN(sizeof(struct ovs_header))
+ + nla_total_size(skb->len) /* OVS_PACKET_ATTR_PACKET */
+ + nla_total_size(key_attr_size()); /* OVS_PACKET_ATTR_KEY */
+
+ /* OVS_PACKET_ATTR_USERDATA */
+ if (userdata)
+ size += NLA_ALIGN(userdata->nla_len);
+
+ return size;
+}
+
static int queue_userspace_packet(struct net *net, int dp_ifindex,
struct sk_buff *skb,
const struct dp_upcall_info *upcall_info)
struct sk_buff *nskb = NULL;
struct sk_buff *user_skb; /* to be queued to userspace */
struct nlattr *nla;
- unsigned int len;
int err;
if (vlan_tx_tag_present(skb)) {
goto out;
}
- len = sizeof(struct ovs_header);
- len += nla_total_size(skb->len);
- len += nla_total_size(FLOW_BUFSIZE);
- if (upcall_info->userdata)
- len += NLA_ALIGN(upcall_info->userdata->nla_len);
-
- user_skb = genlmsg_new(len, GFP_ATOMIC);
+ user_skb = genlmsg_new(upcall_msg_size(skb, upcall_info->userdata), GFP_ATOMIC);
if (!user_skb) {
err = -ENOMEM;
goto out;
upcall->dp_ifindex = dp_ifindex;
nla = nla_nest_start(user_skb, OVS_PACKET_ATTR_KEY);
- ovs_flow_to_nlattrs(upcall_info->key, user_skb);
+ ovs_flow_to_nlattrs(upcall_info->key, upcall_info->key, user_skb);
nla_nest_end(user_skb, nla);
if (upcall_info->userdata)
return err;
}
-/* Called with genl_mutex. */
+/* Called with ovs_mutex. */
static int flush_flows(struct datapath *dp)
{
struct flow_table *old_table;
struct flow_table *new_table;
- old_table = genl_dereference(dp->table);
+ old_table = ovsl_dereference(dp->table);
new_table = ovs_flow_tbl_alloc(TBL_MIN_BUCKETS);
if (!new_table)
return -ENOMEM;
rcu_assign_pointer(dp->table, new_table);
- ovs_flow_tbl_deferred_destroy(old_table);
+ ovs_flow_tbl_destroy(old_table, true);
return 0;
}
static int validate_and_copy_set_tun(const struct nlattr *attr,
struct sw_flow_actions **sfa)
{
- struct ovs_key_ipv4_tunnel tun_key;
+ struct sw_flow_match match;
+ struct sw_flow_key key;
int err, start;
- err = ipv4_tun_from_nlattr(nla_data(attr), &tun_key);
+ ovs_match_init(&match, &key, NULL);
+ err = ovs_ipv4_tun_from_nlattr(nla_data(attr), &match, false);
if (err)
return err;
if (start < 0)
return start;
- err = add_action(sfa, OVS_KEY_ATTR_IPV4_TUNNEL, &tun_key, sizeof(tun_key));
+ err = add_action(sfa, OVS_KEY_ATTR_IPV4_TUNNEL, &match.key->tun_key,
+ sizeof(match.key->tun_key));
add_nested_action_end(*sfa, start);
return err;
return validate_tp_port(flow_key);
+ case OVS_KEY_ATTR_SCTP:
+ if (flow_key->ip.proto != IPPROTO_SCTP)
+ return -EINVAL;
+
+ return validate_tp_port(flow_key);
+
default:
return -EINVAL;
}
struct ethhdr *eth;
int len;
int err;
- int key_len;
err = -EINVAL;
if (!a[OVS_PACKET_ATTR_PACKET] || !a[OVS_PACKET_ATTR_KEY] ||
- !a[OVS_PACKET_ATTR_ACTIONS] ||
- nla_len(a[OVS_PACKET_ATTR_PACKET]) < ETH_HLEN)
+ !a[OVS_PACKET_ATTR_ACTIONS])
goto err;
len = nla_len(a[OVS_PACKET_ATTR_PACKET]);
goto err;
skb_reserve(packet, NET_IP_ALIGN);
- memcpy(__skb_put(packet, len), nla_data(a[OVS_PACKET_ATTR_PACKET]), len);
+ nla_memcpy(__skb_put(packet, len), a[OVS_PACKET_ATTR_PACKET], len);
skb_reset_mac_header(packet);
eth = eth_hdr(packet);
/* Normally, setting the skb 'protocol' field would be handled by a
* call to eth_type_trans(), but it assumes there's a sending
* device, which we may not have. */
- if (ntohs(eth->h_proto) >= 1536)
+ if (ntohs(eth->h_proto) >= ETH_P_802_3_MIN)
packet->protocol = eth->h_proto;
else
packet->protocol = htons(ETH_P_802_2);
if (IS_ERR(flow))
goto err_kfree_skb;
- err = ovs_flow_extract(packet, -1, &flow->key, &key_len);
+ err = ovs_flow_extract(packet, -1, &flow->key);
if (err)
goto err_flow_free;
- err = ovs_flow_metadata_from_nlattrs(flow, key_len, a[OVS_PACKET_ATTR_KEY]);
+ err = ovs_flow_metadata_from_nlattrs(flow, a[OVS_PACKET_ATTR_KEY]);
if (err)
goto err_flow_free;
acts = ovs_flow_actions_alloc(nla_len(a[OVS_PACKET_ATTR_ACTIONS]));
goto err_flow_free;
OVS_CB(packet)->flow = flow;
+ OVS_CB(packet)->pkt_key = &flow->key;
packet->priority = flow->key.phy.priority;
skb_set_mark(packet, flow->key.phy.skb_mark);
local_bh_enable();
rcu_read_unlock();
- ovs_flow_free(flow);
+ ovs_flow_free(flow, false);
return err;
err_unlock:
rcu_read_unlock();
err_flow_free:
- ovs_flow_free(flow);
+ ovs_flow_free(flow, false);
err_kfree_skb:
kfree_skb(packet);
err:
}
static const struct nla_policy packet_policy[OVS_PACKET_ATTR_MAX + 1] = {
- [OVS_PACKET_ATTR_PACKET] = { .type = NLA_UNSPEC },
+#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,18)
+ [OVS_PACKET_ATTR_PACKET] = { .len = ETH_HLEN },
+#else
+ [OVS_PACKET_ATTR_PACKET] = { .minlen = ETH_HLEN },
+#endif
[OVS_PACKET_ATTR_KEY] = { .type = NLA_NESTED },
[OVS_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED },
};
static void get_dp_stats(struct datapath *dp, struct ovs_dp_stats *stats)
{
+ struct flow_table *table;
int i;
- struct flow_table *table = genl_dereference(dp->table);
+ table = rcu_dereference_check(dp->table, lockdep_ovsl_is_held());
stats->n_flows = ovs_flow_tbl_count(table);
stats->n_hit = stats->n_missed = stats->n_lost = 0;
.version = OVS_FLOW_VERSION,
.maxattr = OVS_FLOW_ATTR_MAX,
SET_NETNSOK
+ SET_PARALLEL_OPS
};
static struct genl_multicast_group ovs_dp_flow_multicast_group = {
if (!start)
return -EMSGSIZE;
- err = ipv4_tun_to_nlattr(skb, nla_data(ovs_key));
+ err = ovs_ipv4_tun_to_nlattr(skb, nla_data(ovs_key),
+ nla_data(ovs_key));
if (err)
return err;
nla_nest_end(skb, start);
return 0;
}
-/* Called with genl_lock. */
+static size_t ovs_flow_cmd_msg_size(const struct sw_flow_actions *acts)
+{
+ return NLMSG_ALIGN(sizeof(struct ovs_header))
+ + nla_total_size(key_attr_size()) /* OVS_FLOW_ATTR_KEY */
+ + nla_total_size(key_attr_size()) /* OVS_FLOW_ATTR_MASK */
+ + nla_total_size(sizeof(struct ovs_flow_stats)) /* OVS_FLOW_ATTR_STATS */
+ + nla_total_size(1) /* OVS_FLOW_ATTR_TCP_FLAGS */
+ + nla_total_size(8) /* OVS_FLOW_ATTR_USED */
+ + nla_total_size(acts->actions_len); /* OVS_FLOW_ATTR_ACTIONS */
+}
+
+/* Called with ovs_mutex. */
static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
struct sk_buff *skb, u32 portid,
u32 seq, u32 flags, u8 cmd)
{
const int skb_orig_len = skb->len;
- const struct sw_flow_actions *sf_acts;
struct nlattr *start;
struct ovs_flow_stats stats;
struct ovs_header *ovs_header;
u8 tcp_flags;
int err;
- sf_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
-
ovs_header = genlmsg_put(skb, portid, seq, &dp_flow_genl_family, flags, cmd);
if (!ovs_header)
return -EMSGSIZE;
ovs_header->dp_ifindex = get_dpifindex(dp);
+ /* Fill flow key. */
nla = nla_nest_start(skb, OVS_FLOW_ATTR_KEY);
if (!nla)
goto nla_put_failure;
- err = ovs_flow_to_nlattrs(&flow->key, skb);
+
+ err = ovs_flow_to_nlattrs(&flow->unmasked_key,
+ &flow->unmasked_key, skb);
if (err)
goto error;
nla_nest_end(skb, nla);
+ nla = nla_nest_start(skb, OVS_FLOW_ATTR_MASK);
+ if (!nla)
+ goto nla_put_failure;
+
+ err = ovs_flow_to_nlattrs(&flow->key, &flow->mask->key, skb);
+ if (err)
+ goto error;
+
+ nla_nest_end(skb, nla);
+
spin_lock_bh(&flow->lock);
used = flow->used;
stats.n_packets = flow->packet_count;
*/
start = nla_nest_start(skb, OVS_FLOW_ATTR_ACTIONS);
if (start) {
+ const struct sw_flow_actions *sf_acts;
+
+ sf_acts = rcu_dereference_check(flow->sf_acts,
+ lockdep_ovsl_is_held());
+
err = actions_to_attr(sf_acts->actions, sf_acts->actions_len, skb);
if (!err)
nla_nest_end(skb, start);
static struct sk_buff *ovs_flow_cmd_alloc_info(struct sw_flow *flow)
{
const struct sw_flow_actions *sf_acts;
- int len;
-
- sf_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
-
- /* OVS_FLOW_ATTR_KEY */
- len = nla_total_size(FLOW_BUFSIZE);
- /* OVS_FLOW_ATTR_ACTIONS */
- len += nla_total_size(sf_acts->actions_len);
- /* OVS_FLOW_ATTR_STATS */
- len += nla_total_size(sizeof(struct ovs_flow_stats));
- /* OVS_FLOW_ATTR_TCP_FLAGS */
- len += nla_total_size(1);
- /* OVS_FLOW_ATTR_USED */
- len += nla_total_size(8);
- len += NLMSG_ALIGN(sizeof(struct ovs_header));
+ sf_acts = ovsl_dereference(flow->sf_acts);
- return genlmsg_new(len, GFP_KERNEL);
+ return genlmsg_new(ovs_flow_cmd_msg_size(sf_acts), GFP_KERNEL);
}
static struct sk_buff *ovs_flow_cmd_build_info(struct sw_flow *flow,
{
struct nlattr **a = info->attrs;
struct ovs_header *ovs_header = info->userhdr;
- struct sw_flow_key key;
- struct sw_flow *flow;
+ struct sw_flow_key key, masked_key;
+ struct sw_flow *flow = NULL;
+ struct sw_flow_mask mask;
struct sk_buff *reply;
struct datapath *dp;
struct flow_table *table;
struct sw_flow_actions *acts = NULL;
+ struct sw_flow_match match;
int error;
- int key_len;
/* Extract key. */
error = -EINVAL;
if (!a[OVS_FLOW_ATTR_KEY])
goto error;
- error = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
+
+ ovs_match_init(&match, &key, &mask);
+ error = ovs_match_from_nlattrs(&match,
+ a[OVS_FLOW_ATTR_KEY], a[OVS_FLOW_ATTR_MASK]);
if (error)
goto error;
if (IS_ERR(acts))
goto error;
- error = validate_and_copy_actions(a[OVS_FLOW_ATTR_ACTIONS], &key, 0, &acts);
- if (error)
+ ovs_flow_key_mask(&masked_key, &key, &mask);
+ error = validate_and_copy_actions(a[OVS_FLOW_ATTR_ACTIONS],
+ &masked_key, 0, &acts);
+ if (error) {
+ OVS_NLERR("Flow actions may not be safe on all matching packets.\n");
goto err_kfree;
+ }
} else if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW) {
error = -EINVAL;
goto error;
}
+ ovs_lock();
dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
error = -ENODEV;
if (!dp)
- goto err_kfree;
+ goto err_unlock_ovs;
+
+ table = ovsl_dereference(dp->table);
- table = genl_dereference(dp->table);
- flow = ovs_flow_tbl_lookup(table, &key, key_len);
+ /* Check if this is a duplicate flow */
+ flow = ovs_flow_lookup(table, &key);
if (!flow) {
+ struct sw_flow_mask *mask_p;
/* Bail out if we're not allowed to create a new flow. */
error = -ENOENT;
if (info->genlhdr->cmd == OVS_FLOW_CMD_SET)
- goto err_kfree;
+ goto err_unlock_ovs;
/* Expand table, if necessary, to make room. */
if (ovs_flow_tbl_need_to_expand(table)) {
new_table = ovs_flow_tbl_expand(table);
if (!IS_ERR(new_table)) {
rcu_assign_pointer(dp->table, new_table);
- ovs_flow_tbl_deferred_destroy(table);
- table = genl_dereference(dp->table);
+ ovs_flow_tbl_destroy(table, true);
+ table = ovsl_dereference(dp->table);
}
}
flow = ovs_flow_alloc();
if (IS_ERR(flow)) {
error = PTR_ERR(flow);
- goto err_kfree;
+ goto err_unlock_ovs;
}
clear_stats(flow);
+ flow->key = masked_key;
+ flow->unmasked_key = key;
+
+ /* Make sure mask is unique in the system */
+ mask_p = ovs_sw_flow_mask_find(table, &mask);
+ if (!mask_p) {
+ /* Allocate a new mask if none exsits. */
+ mask_p = ovs_sw_flow_mask_alloc();
+ if (!mask_p)
+ goto err_flow_free;
+ mask_p->key = mask.key;
+ mask_p->range = mask.range;
+ ovs_sw_flow_mask_insert(table, mask_p);
+ }
+
+ ovs_sw_flow_mask_add_ref(mask_p);
+ flow->mask = mask_p;
rcu_assign_pointer(flow->sf_acts, acts);
/* Put flow in bucket. */
- ovs_flow_tbl_insert(table, flow, &key, key_len);
+ ovs_flow_insert(table, flow);
reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
- info->snd_seq,
- OVS_FLOW_CMD_NEW);
+ info->snd_seq, OVS_FLOW_CMD_NEW);
} else {
/* We found a matching flow. */
struct sw_flow_actions *old_acts;
error = -EEXIST;
if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW &&
info->nlhdr->nlmsg_flags & (NLM_F_CREATE | NLM_F_EXCL))
- goto err_kfree;
+ goto err_unlock_ovs;
+
+ /* The unmasked key has to be the same for flow updates. */
+ error = -EINVAL;
+ if (!ovs_flow_cmp_unmasked_key(flow, &key, match.range.end)) {
+ OVS_NLERR("Flow modification message rejected, unmasked key does not match.\n");
+ goto err_unlock_ovs;
+ }
/* Update actions. */
- old_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
+ old_acts = ovsl_dereference(flow->sf_acts);
rcu_assign_pointer(flow->sf_acts, acts);
ovs_flow_deferred_free_acts(old_acts);
spin_unlock_bh(&flow->lock);
}
}
+ ovs_unlock();
if (!IS_ERR(reply))
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_flow_multicast_group.id, info->nlhdr,
- GFP_KERNEL);
+ ovs_notify(reply, info, &ovs_dp_flow_multicast_group);
else
netlink_set_err(GENL_SOCK(sock_net(skb->sk)), 0,
ovs_dp_flow_multicast_group.id, PTR_ERR(reply));
return 0;
+err_flow_free:
+ ovs_flow_free(flow, false);
+err_unlock_ovs:
+ ovs_unlock();
err_kfree:
kfree(acts);
error:
struct sw_flow *flow;
struct datapath *dp;
struct flow_table *table;
+ struct sw_flow_match match;
int err;
- int key_len;
- if (!a[OVS_FLOW_ATTR_KEY])
+ if (!a[OVS_FLOW_ATTR_KEY]) {
+ OVS_NLERR("Flow get message rejected, Key attribute missing.\n");
return -EINVAL;
- err = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
+ }
+
+ ovs_match_init(&match, &key, NULL);
+ err = ovs_match_from_nlattrs(&match, a[OVS_FLOW_ATTR_KEY], NULL);
if (err)
return err;
+ ovs_lock();
dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
- if (!dp)
- return -ENODEV;
+ if (!dp) {
+ err = -ENODEV;
+ goto unlock;
+ }
- table = genl_dereference(dp->table);
- flow = ovs_flow_tbl_lookup(table, &key, key_len);
- if (!flow)
- return -ENOENT;
+ table = ovsl_dereference(dp->table);
+ flow = ovs_flow_lookup_unmasked_key(table, &match);
+ if (!flow) {
+ err = -ENOENT;
+ goto unlock;
+ }
reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid,
info->snd_seq, OVS_FLOW_CMD_NEW);
- if (IS_ERR(reply))
- return PTR_ERR(reply);
+ if (IS_ERR(reply)) {
+ err = PTR_ERR(reply);
+ goto unlock;
+ }
+ ovs_unlock();
return genlmsg_reply(reply, info);
+unlock:
+ ovs_unlock();
+ return err;
}
static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info)
struct sw_flow *flow;
struct datapath *dp;
struct flow_table *table;
+ struct sw_flow_match match;
int err;
- int key_len;
+ ovs_lock();
dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
- if (!dp)
- return -ENODEV;
+ if (!dp) {
+ err = -ENODEV;
+ goto unlock;
+ }
- if (!a[OVS_FLOW_ATTR_KEY])
- return flush_flows(dp);
+ if (!a[OVS_FLOW_ATTR_KEY]) {
+ err = flush_flows(dp);
+ goto unlock;
+ }
- err = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]);
+ ovs_match_init(&match, &key, NULL);
+ err = ovs_match_from_nlattrs(&match, a[OVS_FLOW_ATTR_KEY], NULL);
if (err)
- return err;
+ goto unlock;
- table = genl_dereference(dp->table);
- flow = ovs_flow_tbl_lookup(table, &key, key_len);
- if (!flow)
- return -ENOENT;
+ table = ovsl_dereference(dp->table);
+ flow = ovs_flow_lookup_unmasked_key(table, &match);
+ if (!flow) {
+ err = -ENOENT;
+ goto unlock;
+ }
reply = ovs_flow_cmd_alloc_info(flow);
- if (!reply)
- return -ENOMEM;
+ if (!reply) {
+ err = -ENOMEM;
+ goto unlock;
+ }
- ovs_flow_tbl_remove(table, flow);
+ ovs_flow_remove(table, flow);
err = ovs_flow_cmd_fill_info(flow, dp, reply, info->snd_portid,
info->snd_seq, 0, OVS_FLOW_CMD_DEL);
BUG_ON(err < 0);
- ovs_flow_deferred_free(flow);
+ ovs_flow_free(flow, true);
+ ovs_unlock();
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_flow_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_notify(reply, info, &ovs_dp_flow_multicast_group);
return 0;
+unlock:
+ ovs_unlock();
+ return err;
}
static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
struct datapath *dp;
struct flow_table *table;
+ rcu_read_lock();
dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
- if (!dp)
+ if (!dp) {
+ rcu_read_unlock();
return -ENODEV;
+ }
- table = genl_dereference(dp->table);
-
+ table = rcu_dereference(dp->table);
for (;;) {
struct sw_flow *flow;
u32 bucket, obj;
bucket = cb->args[0];
obj = cb->args[1];
- flow = ovs_flow_tbl_next(table, &bucket, &obj);
+ flow = ovs_flow_dump_next(table, &bucket, &obj);
if (!flow)
break;
cb->args[0] = bucket;
cb->args[1] = obj;
}
+ rcu_read_unlock();
return skb->len;
}
.version = OVS_DATAPATH_VERSION,
.maxattr = OVS_DP_ATTR_MAX,
SET_NETNSOK
+ SET_PARALLEL_OPS
};
static struct genl_multicast_group ovs_dp_datapath_multicast_group = {
.name = OVS_DATAPATH_MCGROUP
};
+static size_t ovs_dp_cmd_msg_size(void)
+{
+ size_t msgsize = NLMSG_ALIGN(sizeof(struct ovs_header));
+
+ msgsize += nla_total_size(IFNAMSIZ);
+ msgsize += nla_total_size(sizeof(struct ovs_dp_stats));
+
+ return msgsize;
+}
+
static int ovs_dp_cmd_fill_info(struct datapath *dp, struct sk_buff *skb,
u32 portid, u32 seq, u32 flags, u8 cmd)
{
struct sk_buff *skb;
int retval;
- skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
+ skb = genlmsg_new(ovs_dp_cmd_msg_size(), GFP_KERNEL);
if (!skb)
return ERR_PTR(-ENOMEM);
return CHECK_NUL_STRING(a[OVS_DP_ATTR_NAME], IFNAMSIZ - 1);
}
-/* Called with genl_mutex and optionally with RTNL lock also. */
+/* Called with ovs_mutex. */
static struct datapath *lookup_datapath(struct net *net,
struct ovs_header *ovs_header,
struct nlattr *a[OVS_DP_ATTR_MAX + 1])
if (err)
goto err;
- rtnl_lock();
+ ovs_lock();
err = -ENOMEM;
dp = kzalloc(sizeof(*dp), GFP_KERNEL);
if (dp == NULL)
- goto err_unlock_rtnl;
+ goto err_unlock_ovs;
ovs_dp_set_net(dp, hold_net(sock_net(skb->sk)));
goto err_destroy_local_port;
ovs_net = net_generic(ovs_dp_get_net(dp), ovs_net_id);
- list_add_tail(&dp->list_node, &ovs_net->dps);
+ list_add_tail_rcu(&dp->list_node, &ovs_net->dps);
- rtnl_unlock();
+ ovs_unlock();
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_datapath_multicast_group.id, info->nlhdr,
- GFP_KERNEL);
+ ovs_notify(reply, info, &ovs_dp_datapath_multicast_group);
return 0;
err_destroy_local_port:
- ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_LOCAL));
+ ovs_dp_detach_port(ovs_vport_ovsl(dp, OVSP_LOCAL));
err_destroy_ports_array:
kfree(dp->ports);
err_destroy_percpu:
free_percpu(dp->stats_percpu);
err_destroy_table:
- ovs_flow_tbl_destroy(genl_dereference(dp->table));
+ ovs_flow_tbl_destroy(ovsl_dereference(dp->table), false);
err_free_dp:
release_net(ovs_dp_get_net(dp));
kfree(dp);
-err_unlock_rtnl:
- rtnl_unlock();
+err_unlock_ovs:
+ ovs_unlock();
err:
return err;
}
-/* Called with genl_mutex. */
+/* Called with ovs_mutex. */
static void __dp_destroy(struct datapath *dp)
{
int i;
- rtnl_lock();
-
for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) {
struct vport *vport;
struct hlist_node *n;
ovs_dp_detach_port(vport);
}
- list_del(&dp->list_node);
- ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_LOCAL));
+ list_del_rcu(&dp->list_node);
- /* rtnl_unlock() will wait until all the references to devices that
- * are pending unregistration have been dropped. We do it here to
- * ensure that any internal devices (which contain DP pointers) are
- * fully destroyed before freeing the datapath.
- */
- rtnl_unlock();
+ /* OVSP_LOCAL is datapath internal port. We need to make sure that
+ * all port in datapath are destroyed first before freeing datapath.
+ */
+ ovs_dp_detach_port(ovs_vport_ovsl(dp, OVSP_LOCAL));
call_rcu(&dp->rcu, destroy_dp_rcu);
}
if (err)
return err;
+ ovs_lock();
dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
err = PTR_ERR(dp);
if (IS_ERR(dp))
- return err;
+ goto unlock;
reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
info->snd_seq, OVS_DP_CMD_DEL);
err = PTR_ERR(reply);
if (IS_ERR(reply))
- return err;
+ goto unlock;
__dp_destroy(dp);
+ ovs_unlock();
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_datapath_multicast_group.id, info->nlhdr,
- GFP_KERNEL);
+ ovs_notify(reply, info, &ovs_dp_datapath_multicast_group);
return 0;
+unlock:
+ ovs_unlock();
+ return err;
}
static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info)
if (err)
return err;
+ ovs_lock();
dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
+ err = PTR_ERR(dp);
if (IS_ERR(dp))
- return PTR_ERR(dp);
+ goto unlock;
reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
info->snd_seq, OVS_DP_CMD_NEW);
err = PTR_ERR(reply);
netlink_set_err(GENL_SOCK(sock_net(skb->sk)), 0,
ovs_dp_datapath_multicast_group.id, err);
- return 0;
+ err = 0;
+ goto unlock;
}
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_datapath_multicast_group.id, info->nlhdr,
- GFP_KERNEL);
+ ovs_unlock();
+ ovs_notify(reply, info, &ovs_dp_datapath_multicast_group);
return 0;
+unlock:
+ ovs_unlock();
+ return err;
}
static int ovs_dp_cmd_get(struct sk_buff *skb, struct genl_info *info)
if (err)
return err;
+ ovs_lock();
dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
- if (IS_ERR(dp))
- return PTR_ERR(dp);
+ if (IS_ERR(dp)) {
+ err = PTR_ERR(dp);
+ goto unlock;
+ }
reply = ovs_dp_cmd_build_info(dp, info->snd_portid,
info->snd_seq, OVS_DP_CMD_NEW);
- if (IS_ERR(reply))
- return PTR_ERR(reply);
+ if (IS_ERR(reply)) {
+ err = PTR_ERR(reply);
+ goto unlock;
+ }
+ ovs_unlock();
return genlmsg_reply(reply, info);
+
+unlock:
+ ovs_unlock();
+ return err;
}
static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
int skip = cb->args[0];
int i = 0;
- list_for_each_entry(dp, &ovs_net->dps, list_node) {
+ rcu_read_lock();
+ list_for_each_entry_rcu(dp, &ovs_net->dps, list_node) {
if (i >= skip &&
ovs_dp_cmd_fill_info(dp, skb, NETLINK_CB(cb->skb).portid,
cb->nlh->nlmsg_seq, NLM_F_MULTI,
break;
i++;
}
+ rcu_read_unlock();
cb->args[0] = i;
.version = OVS_VPORT_VERSION,
.maxattr = OVS_VPORT_ATTR_MAX,
SET_NETNSOK
+ SET_PARALLEL_OPS
};
struct genl_multicast_group ovs_dp_vport_multicast_group = {
.name = OVS_VPORT_MCGROUP
};
-/* Called with RTNL lock or RCU read lock. */
+/* Called with ovs_mutex or RCU read lock. */
static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb,
u32 portid, u32 seq, u32 flags, u8 cmd)
{
return err;
}
-/* Called with RTNL lock or RCU read lock. */
+/* Called with ovs_mutex or RCU read lock. */
struct sk_buff *ovs_vport_cmd_build_info(struct vport *vport, u32 portid,
u32 seq, u8 cmd)
{
return ERR_PTR(-ENOMEM);
retval = ovs_vport_cmd_fill_info(vport, skb, portid, seq, 0, cmd);
- if (retval < 0) {
- kfree_skb(skb);
- return ERR_PTR(retval);
- }
+ BUG_ON(retval < 0);
+
return skb;
}
return CHECK_NUL_STRING(a[OVS_VPORT_ATTR_NAME], IFNAMSIZ - 1);
}
-/* Called with RTNL lock or RCU read lock. */
+/* Called with ovs_mutex or RCU read lock. */
static struct vport *lookup_vport(struct net *net,
struct ovs_header *ovs_header,
struct nlattr *a[OVS_VPORT_ATTR_MAX + 1])
if (!dp)
return ERR_PTR(-ENODEV);
- vport = ovs_vport_rtnl_rcu(dp, port_no);
+ vport = ovs_vport_ovsl_rcu(dp, port_no);
if (!vport)
return ERR_PTR(-ENODEV);
return vport;
if (err)
goto exit;
- rtnl_lock();
+ ovs_lock();
dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
err = -ENODEV;
if (!dp)
if (port_no >= DP_MAX_PORTS)
goto exit_unlock;
- vport = ovs_vport_rtnl(dp, port_no);
+ vport = ovs_vport_ovsl(dp, port_no);
err = -EBUSY;
if (vport)
goto exit_unlock;
err = -EFBIG;
goto exit_unlock;
}
- vport = ovs_vport_rtnl(dp, port_no);
+ vport = ovs_vport_ovsl(dp, port_no);
if (!vport)
break;
}
ovs_dp_detach_port(vport);
goto exit_unlock;
}
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+
+ ovs_notify(reply, info, &ovs_dp_vport_multicast_group);
exit_unlock:
- rtnl_unlock();
+ ovs_unlock();
exit:
return err;
}
if (err)
goto exit;
- rtnl_lock();
+ ovs_lock();
vport = lookup_vport(sock_net(skb->sk), info->userhdr, a);
err = PTR_ERR(vport);
if (IS_ERR(vport))
goto exit_unlock;
- err = 0;
if (a[OVS_VPORT_ATTR_TYPE] &&
- nla_get_u32(a[OVS_VPORT_ATTR_TYPE]) != vport->ops->type)
+ nla_get_u32(a[OVS_VPORT_ATTR_TYPE]) != vport->ops->type) {
err = -EINVAL;
+ goto exit_unlock;
+ }
- if (!err && a[OVS_VPORT_ATTR_OPTIONS])
- err = ovs_vport_set_options(vport, a[OVS_VPORT_ATTR_OPTIONS]);
- if (err)
+ reply = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
+ if (!reply) {
+ err = -ENOMEM;
goto exit_unlock;
+ }
+
+ if (a[OVS_VPORT_ATTR_OPTIONS]) {
+ err = ovs_vport_set_options(vport, a[OVS_VPORT_ATTR_OPTIONS]);
+ if (err)
+ goto exit_free;
+ }
if (a[OVS_VPORT_ATTR_STATS])
ovs_vport_set_stats(vport, nla_data(a[OVS_VPORT_ATTR_STATS]));
if (a[OVS_VPORT_ATTR_UPCALL_PID])
vport->upcall_portid = nla_get_u32(a[OVS_VPORT_ATTR_UPCALL_PID]);
- reply = ovs_vport_cmd_build_info(vport, info->snd_portid,
- info->snd_seq, OVS_VPORT_CMD_NEW);
- if (IS_ERR(reply)) {
- netlink_set_err(GENL_SOCK(sock_net(skb->sk)), 0,
- ovs_dp_vport_multicast_group.id, PTR_ERR(reply));
- goto exit_unlock;
- }
+ err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_VPORT_CMD_NEW);
+ BUG_ON(err < 0);
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_unlock();
+ ovs_notify(reply, info, &ovs_dp_vport_multicast_group);
+ return 0;
+exit_free:
+ kfree_skb(reply);
exit_unlock:
- rtnl_unlock();
+ ovs_unlock();
exit:
return err;
}
if (err)
goto exit;
- rtnl_lock();
+ ovs_lock();
vport = lookup_vport(sock_net(skb->sk), info->userhdr, a);
err = PTR_ERR(vport);
if (IS_ERR(vport))
err = 0;
ovs_dp_detach_port(vport);
- genl_notify(reply, genl_info_net(info), info->snd_portid,
- ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_notify(reply, info, &ovs_dp_vport_multicast_group);
exit_unlock:
- rtnl_unlock();
+ ovs_unlock();
exit:
return err;
}
return err;
}
-static int __rehash_flow_table(void *dummy)
+static void rehash_flow_table(struct work_struct *work)
{
struct datapath *dp;
struct net *net;
+ ovs_lock();
rtnl_lock();
for_each_net(net) {
struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
list_for_each_entry(dp, &ovs_net->dps, list_node) {
- struct flow_table *old_table = genl_dereference(dp->table);
+ struct flow_table *old_table = ovsl_dereference(dp->table);
struct flow_table *new_table;
new_table = ovs_flow_tbl_rehash(old_table);
if (!IS_ERR(new_table)) {
rcu_assign_pointer(dp->table, new_table);
- ovs_flow_tbl_deferred_destroy(old_table);
+ ovs_flow_tbl_destroy(old_table, true);
}
}
}
rtnl_unlock();
- return 0;
-}
-
-static void rehash_flow_table(struct work_struct *work)
-{
- genl_exec(__rehash_flow_table, NULL);
+ ovs_unlock();
schedule_delayed_work(&rehash_flow_wq, REHASH_FLOW_INTERVAL);
}
-static int dp_destroy_all(void *data)
-{
- struct datapath *dp, *dp_next;
- struct ovs_net *ovs_net = data;
-
- list_for_each_entry_safe(dp, dp_next, &ovs_net->dps, list_node)
- __dp_destroy(dp);
-
- return 0;
-}
-
static int __net_init ovs_init_net(struct net *net)
{
struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
INIT_LIST_HEAD(&ovs_net->dps);
+ INIT_WORK(&ovs_net->dp_notify_work, ovs_dp_notify_wq);
return 0;
}
static void __net_exit ovs_exit_net(struct net *net)
{
+ struct datapath *dp, *dp_next;
struct ovs_net *ovs_net = net_generic(net, ovs_net_id);
- genl_exec(dp_destroy_all, ovs_net);
+ ovs_lock();
+ list_for_each_entry_safe(dp, dp_next, &ovs_net->dps, list_node)
+ __dp_destroy(dp);
+ ovs_unlock();
+
+ cancel_work_sync(&ovs_net->dp_notify_work);
}
static struct pernet_operations ovs_net_ops = {
.size = sizeof(struct ovs_net),
};
+DEFINE_COMPAT_PNET_REG_FUNC(device);
+
static int __init dp_init(void)
{
int err;
pr_info("Open vSwitch switching datapath %s, built "__DATE__" "__TIME__"\n",
VERSION);
- err = genl_exec_init();
- if (err)
- goto error;
-
err = ovs_workqueues_init();
if (err)
- goto error_genl_exec;
+ goto error;
err = ovs_flow_init();
if (err)
ovs_flow_exit();
error_wq:
ovs_workqueues_exit();
-error_genl_exec:
- genl_exec_exit();
error:
return err;
}
ovs_vport_exit();
ovs_flow_exit();
ovs_workqueues_exit();
- genl_exec_exit();
}
module_init(dp_init);
git://git.onelab.eu / sliver-openvswitch.git / blobdiff