/*
- * Copyright (c) 2007, 2008, 2009, 2010, 2011 Nicira Networks.
- * Distributed under the terms of the GNU GPL version 2.
+ * Copyright (c) 2007-2014 Nicira, Inc.
*
- * Significant portions of this file may be copied from parts of the Linux
- * kernel, by Linus Torvalds and others.
+ * 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
*/
-/* Functions for managing the dp interface/device. */
-
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/init.h>
#include <linux/version.h>
#include <linux/ethtool.h>
#include <linux/wait.h>
-#include <asm/system.h>
#include <asm/div64.h>
-#include <asm/bug.h>
#include <linux/highmem.h>
#include <linux/netfilter_bridge.h>
#include <linux/netfilter_ipv4.h>
#include <linux/inetdevice.h>
#include <linux/list.h>
+#include <linux/openvswitch.h>
#include <linux/rculist.h>
#include <linux/dmi.h>
-#include <net/inet_ecn.h>
+#include <linux/genetlink.h>
+#include <net/genetlink.h>
#include <net/genetlink.h>
+#include <net/net_namespace.h>
+#include <net/netns/generic.h>
-#include "openvswitch/datapath-protocol.h"
-#include "checksum.h"
#include "datapath.h"
-#include "actions.h"
#include "flow.h"
-#include "table.h"
+#include "flow_table.h"
+#include "flow_netlink.h"
#include "vlan.h"
#include "vport-internal_dev.h"
+#include "vport-netdev.h"
+
+int ovs_net_id __read_mostly;
-int (*dp_ioctl_hook)(struct net_device *dev, struct ifreq *rq, int cmd);
-EXPORT_SYMBOL(dp_ioctl_hook);
+static struct genl_family dp_packet_genl_family;
+static struct genl_family dp_flow_genl_family;
+static struct genl_family dp_datapath_genl_family;
+
+static struct genl_multicast_group ovs_dp_flow_multicast_group = {
+ .name = OVS_FLOW_MCGROUP
+};
+
+static struct genl_multicast_group ovs_dp_datapath_multicast_group = {
+ .name = OVS_DATAPATH_MCGROUP
+};
+
+struct genl_multicast_group ovs_dp_vport_multicast_group = {
+ .name = OVS_VPORT_MCGROUP
+};
+
+/* Check if need to build a reply message.
+ * OVS userspace sets the NLM_F_ECHO flag if it needs the reply. */
+static bool ovs_must_notify(struct genl_info *info,
+ const struct genl_multicast_group *grp)
+{
+ return info->nlhdr->nlmsg_flags & NLM_F_ECHO ||
+ netlink_has_listeners(genl_info_net(info)->genl_sock, GROUP_ID(grp));
+}
+
+static void ovs_notify(struct genl_family *family, struct genl_multicast_group *grp,
+ struct sk_buff *skb, struct genl_info *info)
+{
+ genl_notify(family, skb, genl_info_net(info),
+ info->snd_portid, GROUP_ID(grp), 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 such as drop frags, 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.
*/
-/* Global list of datapaths to enable dumping them all out.
- * Protected by genl_mutex.
- */
-static LIST_HEAD(dps);
+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_control_packets(struct datapath *, struct sk_buff *,
- const struct dp_upcall_info *);
+static int queue_gso_packets(struct datapath *dp, struct sk_buff *,
+ const struct dp_upcall_info *);
+static int queue_userspace_packet(struct datapath *dp, struct sk_buff *,
+ const struct dp_upcall_info *);
-/* Must be called with rcu_read_lock, genl_mutex, or RTNL lock. */
-struct datapath *get_dp(int dp_ifindex)
+/* 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;
struct net_device *dev;
rcu_read_lock();
- dev = dev_get_by_index_rcu(&init_net, dp_ifindex);
+ dev = dev_get_by_index_rcu(net, dp_ifindex);
if (dev) {
- struct vport *vport = internal_dev_get_vport(dev);
+ struct vport *vport = ovs_internal_dev_get_vport(dev);
if (vport)
dp = vport->dp;
}
return dp;
}
-EXPORT_SYMBOL_GPL(get_dp);
-/* Must be called with genl_mutex. */
-static struct tbl *get_table_protected(struct datapath *dp)
+/* Must be called with rcu_read_lock or ovs_mutex. */
+const char *ovs_dp_name(const struct datapath *dp)
{
- return rcu_dereference_protected(dp->table, lockdep_genl_is_held());
+ struct vport *vport = ovs_vport_ovsl_rcu(dp, OVSP_LOCAL);
+ return vport->ops->get_name(vport);
}
-/* Must be called with rcu_read_lock or RTNL lock. */
-static struct vport *get_vport_protected(struct datapath *dp, u16 port_no)
+static int get_dpifindex(struct datapath *dp)
{
- return rcu_dereference_rtnl(dp->ports[port_no]);
-}
+ struct vport *local;
+ int ifindex;
-/* Must be called with rcu_read_lock or RTNL lock. */
-const char *dp_name(const struct datapath *dp)
-{
- return vport_get_name(rcu_dereference_rtnl(dp->ports[ODPP_LOCAL]));
-}
-
-static inline size_t br_nlmsg_size(void)
-{
- return NLMSG_ALIGN(sizeof(struct ifinfomsg))
- + nla_total_size(IFNAMSIZ) /* IFLA_IFNAME */
- + nla_total_size(MAX_ADDR_LEN) /* IFLA_ADDRESS */
- + nla_total_size(4) /* IFLA_MASTER */
- + nla_total_size(4) /* IFLA_MTU */
- + nla_total_size(4) /* IFLA_LINK */
- + nla_total_size(1); /* IFLA_OPERSTATE */
-}
-
-/* Caller must hold RTNL lock. */
-static int dp_fill_ifinfo(struct sk_buff *skb,
- const struct vport *port,
- int event, unsigned int flags)
-{
- struct datapath *dp = port->dp;
- int ifindex = vport_get_ifindex(port);
- int iflink = vport_get_iflink(port);
- struct ifinfomsg *hdr;
- struct nlmsghdr *nlh;
-
- if (ifindex < 0)
- return ifindex;
-
- if (iflink < 0)
- return iflink;
-
- nlh = nlmsg_put(skb, 0, 0, event, sizeof(*hdr), flags);
- if (nlh == NULL)
- return -EMSGSIZE;
-
- hdr = nlmsg_data(nlh);
- hdr->ifi_family = AF_BRIDGE;
- hdr->__ifi_pad = 0;
- hdr->ifi_type = ARPHRD_ETHER;
- hdr->ifi_index = ifindex;
- hdr->ifi_flags = vport_get_flags(port);
- hdr->ifi_change = 0;
-
- NLA_PUT_STRING(skb, IFLA_IFNAME, vport_get_name(port));
- NLA_PUT_U32(skb, IFLA_MASTER,
- vport_get_ifindex(get_vport_protected(dp, ODPP_LOCAL)));
- NLA_PUT_U32(skb, IFLA_MTU, vport_get_mtu(port));
-#ifdef IFLA_OPERSTATE
- NLA_PUT_U8(skb, IFLA_OPERSTATE,
- vport_is_running(port)
- ? vport_get_operstate(port)
- : IF_OPER_DOWN);
-#endif
-
- NLA_PUT(skb, IFLA_ADDRESS, ETH_ALEN, vport_get_addr(port));
+ rcu_read_lock();
- if (ifindex != iflink)
- NLA_PUT_U32(skb, IFLA_LINK,iflink);
+ local = ovs_vport_rcu(dp, OVSP_LOCAL);
+ if (local)
+ ifindex = netdev_vport_priv(local)->dev->ifindex;
+ else
+ ifindex = 0;
- return nlmsg_end(skb, nlh);
+ rcu_read_unlock();
-nla_put_failure:
- nlmsg_cancel(skb, nlh);
- return -EMSGSIZE;
+ return ifindex;
}
-/* Caller must hold RTNL lock. */
-static void dp_ifinfo_notify(int event, struct vport *port)
+static void destroy_dp_rcu(struct rcu_head *rcu)
{
- struct sk_buff *skb;
- int err = -ENOBUFS;
-
- skb = nlmsg_new(br_nlmsg_size(), GFP_KERNEL);
- if (skb == NULL)
- goto errout;
+ struct datapath *dp = container_of(rcu, struct datapath, rcu);
- err = dp_fill_ifinfo(skb, port, event, 0);
- if (err < 0) {
- /* -EMSGSIZE implies BUG in br_nlmsg_size() */
- WARN_ON(err == -EMSGSIZE);
- kfree_skb(skb);
- goto errout;
- }
- rtnl_notify(skb, &init_net, 0, RTNLGRP_LINK, NULL, GFP_KERNEL);
- return;
-errout:
- if (err < 0)
- rtnl_set_sk_err(&init_net, RTNLGRP_LINK, err);
+ ovs_flow_tbl_destroy(&dp->table);
+ free_percpu(dp->stats_percpu);
+ release_net(ovs_dp_get_net(dp));
+ kfree(dp->ports);
+ kfree(dp);
}
-static void release_dp(struct kobject *kobj)
+static struct hlist_head *vport_hash_bucket(const struct datapath *dp,
+ u16 port_no)
{
- struct datapath *dp = container_of(kobj, struct datapath, ifobj);
- kfree(dp);
+ return &dp->ports[port_no & (DP_VPORT_HASH_BUCKETS - 1)];
}
-static struct kobj_type dp_ktype = {
- .release = release_dp
-};
-
-static void destroy_dp_rcu(struct rcu_head *rcu)
+/* Called with ovs_mutex or RCU read lock. */
+struct vport *ovs_lookup_vport(const struct datapath *dp, u16 port_no)
{
- struct datapath *dp = container_of(rcu, struct datapath, rcu);
+ struct vport *vport;
+ struct hlist_head *head;
- tbl_destroy((struct tbl __force *)dp->table, flow_free_tbl);
- free_percpu(dp->stats_percpu);
- kobject_put(&dp->ifobj);
+ head = vport_hash_bucket(dp, port_no);
+ hlist_for_each_entry_rcu(vport, head, dp_hash_node) {
+ if (vport->port_no == port_no)
+ return vport;
+ }
+ 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;
- vport = vport_add(parms);
+ vport = ovs_vport_add(parms);
if (!IS_ERR(vport)) {
struct datapath *dp = parms->dp;
+ struct hlist_head *head = vport_hash_bucket(dp, vport->port_no);
- rcu_assign_pointer(dp->ports[parms->port_no], vport);
- list_add(&vport->node, &dp->port_list);
-
- dp_ifinfo_notify(RTM_NEWLINK, vport);
+ hlist_add_head_rcu(&vport->dp_hash_node, head);
}
-
return vport;
}
-/* Called with RTNL lock. */
-int dp_detach_port(struct vport *p)
+void ovs_dp_detach_port(struct vport *p)
{
- ASSERT_RTNL();
-
- if (p->port_no != ODPP_LOCAL)
- dp_sysfs_del_if(p);
- dp_ifinfo_notify(RTM_DELLINK, p);
+ ASSERT_OVSL();
/* First drop references to device. */
- list_del(&p->node);
- rcu_assign_pointer(p->dp->ports[p->port_no], NULL);
+ hlist_del_rcu(&p->dp_hash_node);
/* Then destroy it. */
- return vport_del(p);
+ ovs_vport_del(p);
}
-/* Must be called with rcu_read_lock. */
-void dp_process_received_packet(struct vport *p, struct sk_buff *skb)
+void ovs_dp_process_packet_with_key(struct sk_buff *skb,
+ struct sw_flow_key *pkt_key)
{
+ const struct vport *p = OVS_CB(skb)->input_vport;
struct datapath *dp = p->dp;
+ struct sw_flow *flow;
struct dp_stats_percpu *stats;
- int stats_counter_off;
- int error;
-
- OVS_CB(skb)->vport = p;
-
- if (!OVS_CB(skb)->flow) {
- struct sw_flow_key key;
- struct tbl_node *flow_node;
- bool is_frag;
-
- /* Extract flow from 'skb' into 'key'. */
- error = flow_extract(skb, p->port_no, &key, &is_frag);
- if (unlikely(error)) {
- kfree_skb(skb);
- return;
- }
-
- if (is_frag && dp->drop_frags) {
- kfree_skb(skb);
- stats_counter_off = offsetof(struct dp_stats_percpu, n_frags);
- goto out;
- }
-
- /* Look up flow. */
- flow_node = tbl_lookup(rcu_dereference(dp->table), &key,
- flow_hash(&key), flow_cmp);
- if (unlikely(!flow_node)) {
- struct dp_upcall_info upcall;
-
- upcall.cmd = ODP_PACKET_CMD_MISS;
- upcall.key = &key;
- upcall.userdata = 0;
- upcall.sample_pool = 0;
- upcall.actions = NULL;
- upcall.actions_len = 0;
- dp_upcall(dp, skb, &upcall);
- stats_counter_off = offsetof(struct dp_stats_percpu, n_missed);
- goto out;
- }
-
- OVS_CB(skb)->flow = flow_cast(flow_node);
+ u64 *stats_counter;
+ u32 n_mask_hit;
+
+ stats = this_cpu_ptr(dp->stats_percpu);
+
+ /* Look up flow. */
+ flow = ovs_flow_tbl_lookup_stats(&dp->table, pkt_key, &n_mask_hit);
+ if (unlikely(!flow)) {
+ struct dp_upcall_info upcall;
+
+ upcall.cmd = OVS_PACKET_CMD_MISS;
+ upcall.key = pkt_key;
+ upcall.userdata = NULL;
+ upcall.portid = ovs_vport_find_upcall_portid(p, skb);
+ ovs_dp_upcall(dp, skb, &upcall);
+ consume_skb(skb);
+ stats_counter = &stats->n_missed;
+ goto out;
}
- stats_counter_off = offsetof(struct dp_stats_percpu, n_hit);
- flow_used(OVS_CB(skb)->flow, skb);
- execute_actions(dp, skb);
+ OVS_CB(skb)->pkt_key = pkt_key;
+ OVS_CB(skb)->flow = flow;
+
+ ovs_flow_stats_update(OVS_CB(skb)->flow, pkt_key->tp.flags, skb);
+ ovs_execute_actions(dp, skb);
+ stats_counter = &stats->n_hit;
out:
/* Update datapath statistics. */
- local_bh_disable();
- stats = per_cpu_ptr(dp->stats_percpu, smp_processor_id());
-
- write_seqcount_begin(&stats->seqlock);
- (*(u64 *)((u8 *)stats + stats_counter_off))++;
- write_seqcount_end(&stats->seqlock);
-
- local_bh_enable();
-}
-
-static void copy_and_csum_skb(struct sk_buff *skb, void *to)
-{
- u16 csum_start, csum_offset;
- __wsum csum;
-
- get_skb_csum_pointers(skb, &csum_start, &csum_offset);
- csum_start -= skb_headroom(skb);
-
- skb_copy_bits(skb, 0, to, csum_start);
-
- csum = skb_copy_and_csum_bits(skb, csum_start, to + csum_start,
- skb->len - csum_start, 0);
- *(__sum16 *)(to + csum_start + csum_offset) = csum_fold(csum);
+ u64_stats_update_begin(&stats->sync);
+ (*stats_counter)++;
+ stats->n_mask_hit += n_mask_hit;
+ u64_stats_update_end(&stats->sync);
}
-static struct genl_family dp_packet_genl_family = {
- .id = GENL_ID_GENERATE,
- .hdrsize = sizeof(struct odp_header),
- .name = ODP_PACKET_FAMILY,
- .version = 1,
- .maxattr = ODP_PACKET_ATTR_MAX
-};
-
-/* Generic Netlink multicast groups for upcalls.
- *
- * We really want three unique multicast groups per datapath, but we can't even
- * get one, because genl_register_mc_group() takes genl_lock, which is also
- * held during Generic Netlink message processing, so trying to acquire
- * multicast groups during ODP_DP_NEW processing deadlocks. Instead, we
- * preallocate a few groups and use them round-robin for datapaths. Collision
- * isn't fatal--multicast listeners should check that the family is the one
- * that they want and discard others--but it wastes time and memory to receive
- * unwanted messages.
- */
-#define PACKET_N_MC_GROUPS 16
-static struct genl_multicast_group packet_mc_groups[PACKET_N_MC_GROUPS];
-
-static u32 packet_mc_group(struct datapath *dp, u8 cmd)
-{
- u32 idx;
- BUILD_BUG_ON_NOT_POWER_OF_2(PACKET_N_MC_GROUPS);
-
- idx = jhash_2words(dp->dp_ifindex, cmd, 0) & (PACKET_N_MC_GROUPS - 1);
- return packet_mc_groups[idx].id;
-}
-
-static int packet_register_mc_groups(void)
+/* Must be called with rcu_read_lock. */
+void ovs_dp_process_received_packet(struct vport *p, struct sk_buff *skb)
{
- int i;
+ int error;
+ struct sw_flow_key key;
- for (i = 0; i < PACKET_N_MC_GROUPS; i++) {
- struct genl_multicast_group *group = &packet_mc_groups[i];
- int error;
+ OVS_CB(skb)->input_vport = p;
- sprintf(group->name, "packet%d", i);
- error = genl_register_mc_group(&dp_packet_genl_family, group);
- if (error)
- return error;
+ /* Extract flow from 'skb' into 'key'. */
+ error = ovs_flow_extract(skb, p->port_no, &key);
+ if (unlikely(error)) {
+ kfree_skb(skb);
+ return;
}
- return 0;
+
+ ovs_dp_process_packet_with_key(skb, &key);
}
-int dp_upcall(struct datapath *dp, struct sk_buff *skb, const struct dp_upcall_info *upcall_info)
+int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb,
+ const struct dp_upcall_info *upcall_info)
{
struct dp_stats_percpu *stats;
int err;
- WARN_ON_ONCE(skb_shared(skb));
-
- forward_ip_summed(skb);
-
- err = vswitch_skb_checksum_setup(skb);
- if (err)
- goto err_kfree_skb;
-
- /* Break apart GSO packets into their component pieces. Otherwise
- * userspace may try to stuff a 64kB packet into a 1500-byte MTU. */
- if (skb_is_gso(skb)) {
- struct sk_buff *nskb = skb_gso_segment(skb, NETIF_F_SG | NETIF_F_HW_CSUM);
-
- kfree_skb(skb);
- skb = nskb;
- if (IS_ERR(skb)) {
- err = PTR_ERR(skb);
- goto err;
- }
+ if (upcall_info->portid == 0) {
+ err = -ENOTCONN;
+ goto err;
}
- err = queue_control_packets(dp, skb, upcall_info);
+ if (!skb_is_gso(skb))
+ err = queue_userspace_packet(dp, skb, upcall_info);
+ else
+ err = queue_gso_packets(dp, skb, upcall_info);
if (err)
goto err;
return 0;
-err_kfree_skb:
- kfree_skb(skb);
err:
- local_bh_disable();
- stats = per_cpu_ptr(dp->stats_percpu, smp_processor_id());
+ stats = this_cpu_ptr(dp->stats_percpu);
- write_seqcount_begin(&stats->seqlock);
+ u64_stats_update_begin(&stats->sync);
stats->n_lost++;
- write_seqcount_end(&stats->seqlock);
-
- local_bh_enable();
+ u64_stats_update_end(&stats->sync);
return err;
}
-/* Send each packet in the 'skb' list to userspace for 'dp' as directed by
- * 'upcall_info'. There will be only one packet unless we broke up a GSO
- * packet.
- */
-static int queue_control_packets(struct datapath *dp, struct sk_buff *skb,
- const struct dp_upcall_info *upcall_info)
+static int queue_gso_packets(struct datapath *dp, struct sk_buff *skb,
+ const struct dp_upcall_info *upcall_info)
{
- u32 group = packet_mc_group(dp, upcall_info->cmd);
- struct sk_buff *nskb;
+ unsigned short gso_type = skb_shinfo(skb)->gso_type;
+ struct dp_upcall_info later_info;
+ struct sw_flow_key later_key;
+ struct sk_buff *segs, *nskb;
int err;
- do {
- struct odp_header *upcall;
- struct sk_buff *user_skb; /* to be queued to userspace */
- struct nlattr *nla;
- unsigned int len;
-
- nskb = skb->next;
- skb->next = NULL;
-
- err = vlan_deaccel_tag(skb);
- if (unlikely(err))
- goto err_kfree_skbs;
-
- if (nla_attr_size(skb->len) > USHRT_MAX)
- goto err_kfree_skbs;
-
- len = sizeof(struct odp_header);
- len += nla_total_size(skb->len);
- len += nla_total_size(FLOW_BUFSIZE);
- if (upcall_info->userdata)
- len += nla_total_size(8);
- if (upcall_info->sample_pool)
- len += nla_total_size(4);
- if (upcall_info->actions_len)
- len += nla_total_size(upcall_info->actions_len);
-
- user_skb = genlmsg_new(len, GFP_ATOMIC);
- if (!user_skb) {
- netlink_set_err(INIT_NET_GENL_SOCK, 0, group, -ENOBUFS);
- goto err_kfree_skbs;
- }
+ segs = __skb_gso_segment(skb, NETIF_F_SG, false);
+ if (IS_ERR(segs))
+ return PTR_ERR(segs);
- upcall = genlmsg_put(user_skb, 0, 0, &dp_packet_genl_family, 0, upcall_info->cmd);
- upcall->dp_ifindex = dp->dp_ifindex;
-
- nla = nla_nest_start(user_skb, ODP_PACKET_ATTR_KEY);
- flow_to_nlattrs(upcall_info->key, user_skb);
- nla_nest_end(user_skb, nla);
-
- if (upcall_info->userdata)
- nla_put_u64(user_skb, ODP_PACKET_ATTR_USERDATA, upcall_info->userdata);
- if (upcall_info->sample_pool)
- nla_put_u32(user_skb, ODP_PACKET_ATTR_SAMPLE_POOL, upcall_info->sample_pool);
- if (upcall_info->actions_len) {
- const struct nlattr *actions = upcall_info->actions;
- u32 actions_len = upcall_info->actions_len;
+ /* Queue all of the segments. */
+ skb = segs;
+ do {
+ err = queue_userspace_packet(dp, skb, upcall_info);
+ if (err)
+ break;
- nla = nla_nest_start(user_skb, ODP_PACKET_ATTR_ACTIONS);
- memcpy(__skb_put(user_skb, actions_len), actions, actions_len);
- nla_nest_end(user_skb, nla);
+ if (skb == segs && gso_type & SKB_GSO_UDP) {
+ /* The initial flow key extracted by ovs_flow_extract()
+ * in this case is for a first fragment, so we need to
+ * properly mark later fragments.
+ */
+ later_key = *upcall_info->key;
+ later_key.ip.frag = OVS_FRAG_TYPE_LATER;
+
+ later_info = *upcall_info;
+ later_info.key = &later_key;
+ upcall_info = &later_info;
}
+ } while ((skb = skb->next));
- nla = __nla_reserve(user_skb, ODP_PACKET_ATTR_PACKET, skb->len);
- if (skb->ip_summed == CHECKSUM_PARTIAL)
- copy_and_csum_skb(skb, nla_data(nla));
+ /* Free all of the segments. */
+ skb = segs;
+ do {
+ nskb = skb->next;
+ if (err)
+ kfree_skb(skb);
else
- skb_copy_bits(skb, 0, nla_data(nla), skb->len);
+ consume_skb(skb);
+ } while ((skb = nskb));
+ return err;
+}
- err = genlmsg_multicast(user_skb, 0, group, GFP_ATOMIC);
- if (err)
- goto err_kfree_skbs;
+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 */
+}
- kfree_skb(skb);
- skb = nskb;
- } while (skb);
- return 0;
+static size_t upcall_msg_size(const struct nlattr *userdata,
+ unsigned int hdrlen)
+{
+ size_t size = NLMSG_ALIGN(sizeof(struct ovs_header))
+ + nla_total_size(hdrlen) /* OVS_PACKET_ATTR_PACKET */
+ + nla_total_size(key_attr_size()); /* OVS_PACKET_ATTR_KEY */
-err_kfree_skbs:
- kfree_skb(skb);
- while ((skb = nskb) != NULL) {
- nskb = skb->next;
- kfree_skb(skb);
- }
- return err;
+ /* OVS_PACKET_ATTR_USERDATA */
+ if (userdata)
+ size += NLA_ALIGN(userdata->nla_len);
+
+ return size;
}
-/* Called with genl_mutex. */
-static int flush_flows(int dp_ifindex)
+static int queue_userspace_packet(struct datapath *dp, struct sk_buff *skb,
+ const struct dp_upcall_info *upcall_info)
{
- struct tbl *old_table;
- struct tbl *new_table;
- struct datapath *dp;
-
- dp = get_dp(dp_ifindex);
- if (!dp)
+ struct ovs_header *upcall;
+ struct sk_buff *nskb = NULL;
+ struct sk_buff *user_skb; /* to be queued to userspace */
+ struct nlattr *nla;
+ struct genl_info info = {
+#if LINUX_VERSION_CODE >= KERNEL_VERSION(3,14,0)
+ .dst_sk = ovs_dp_get_net(dp)->genl_sock,
+#endif
+ .snd_portid = upcall_info->portid,
+ };
+ size_t len;
+ unsigned int hlen;
+ int err, dp_ifindex;
+
+ dp_ifindex = get_dpifindex(dp);
+ if (!dp_ifindex)
return -ENODEV;
- old_table = get_table_protected(dp);
- new_table = tbl_create(TBL_MIN_BUCKETS);
- if (!new_table)
- return -ENOMEM;
+ if (vlan_tx_tag_present(skb)) {
+ nskb = skb_clone(skb, GFP_ATOMIC);
+ if (!nskb)
+ return -ENOMEM;
- rcu_assign_pointer(dp->table, new_table);
+ nskb = __vlan_put_tag(nskb, nskb->vlan_proto, vlan_tx_tag_get(nskb));
+ if (!nskb)
+ return -ENOMEM;
- tbl_deferred_destroy(old_table, flow_free_tbl);
+ vlan_set_tci(nskb, 0);
- return 0;
-}
-
-static int validate_actions(const struct nlattr *attr)
-{
- const struct nlattr *a;
- int rem;
-
- nla_for_each_nested(a, attr, rem) {
- static const u32 action_lens[ODP_ACTION_ATTR_MAX + 1] = {
- [ODP_ACTION_ATTR_OUTPUT] = 4,
- [ODP_ACTION_ATTR_CONTROLLER] = 8,
- [ODP_ACTION_ATTR_SET_DL_TCI] = 2,
- [ODP_ACTION_ATTR_STRIP_VLAN] = 0,
- [ODP_ACTION_ATTR_SET_DL_SRC] = ETH_ALEN,
- [ODP_ACTION_ATTR_SET_DL_DST] = ETH_ALEN,
- [ODP_ACTION_ATTR_SET_NW_SRC] = 4,
- [ODP_ACTION_ATTR_SET_NW_DST] = 4,
- [ODP_ACTION_ATTR_SET_NW_TOS] = 1,
- [ODP_ACTION_ATTR_SET_TP_SRC] = 2,
- [ODP_ACTION_ATTR_SET_TP_DST] = 2,
- [ODP_ACTION_ATTR_SET_TUNNEL] = 8,
- [ODP_ACTION_ATTR_SET_PRIORITY] = 4,
- [ODP_ACTION_ATTR_POP_PRIORITY] = 0,
- [ODP_ACTION_ATTR_DROP_SPOOFED_ARP] = 0,
- };
- int type = nla_type(a);
-
- if (type > ODP_ACTION_ATTR_MAX || nla_len(a) != action_lens[type])
- return -EINVAL;
-
- switch (type) {
- case ODP_ACTION_ATTR_UNSPEC:
- return -EINVAL;
-
- case ODP_ACTION_ATTR_CONTROLLER:
- case ODP_ACTION_ATTR_STRIP_VLAN:
- case ODP_ACTION_ATTR_SET_DL_SRC:
- case ODP_ACTION_ATTR_SET_DL_DST:
- case ODP_ACTION_ATTR_SET_NW_SRC:
- case ODP_ACTION_ATTR_SET_NW_DST:
- case ODP_ACTION_ATTR_SET_TP_SRC:
- case ODP_ACTION_ATTR_SET_TP_DST:
- case ODP_ACTION_ATTR_SET_TUNNEL:
- case ODP_ACTION_ATTR_SET_PRIORITY:
- case ODP_ACTION_ATTR_POP_PRIORITY:
- case ODP_ACTION_ATTR_DROP_SPOOFED_ARP:
- /* No validation needed. */
- break;
+ skb = nskb;
+ }
- case ODP_ACTION_ATTR_OUTPUT:
- if (nla_get_u32(a) >= DP_MAX_PORTS)
- return -EINVAL;
- break;
+ if (nla_attr_size(skb->len) > USHRT_MAX) {
+ err = -EFBIG;
+ goto out;
+ }
- case ODP_ACTION_ATTR_SET_DL_TCI:
- if (nla_get_be16(a) & htons(VLAN_CFI_MASK))
- return -EINVAL;
- break;
+ /* Complete checksum if needed */
+ if (skb->ip_summed == CHECKSUM_PARTIAL &&
+ (err = skb_checksum_help(skb)))
+ goto out;
- case ODP_ACTION_ATTR_SET_NW_TOS:
- if (nla_get_u8(a) & INET_ECN_MASK)
- return -EINVAL;
- break;
+ /* Older versions of OVS user space enforce alignment of the last
+ * Netlink attribute to NLA_ALIGNTO which would require extensive
+ * padding logic. Only perform zerocopy if padding is not required.
+ */
+ if (dp->user_features & OVS_DP_F_UNALIGNED)
+ hlen = skb_zerocopy_headlen(skb);
+ else
+ hlen = skb->len;
- default:
- return -EOPNOTSUPP;
- }
+ len = upcall_msg_size(upcall_info->userdata, hlen);
+ user_skb = genlmsg_new_unicast(len, &info, GFP_ATOMIC);
+ if (!user_skb) {
+ err = -ENOMEM;
+ goto out;
}
- if (rem > 0)
- return -EINVAL;
+ upcall = genlmsg_put(user_skb, 0, 0, &dp_packet_genl_family,
+ 0, upcall_info->cmd);
+ upcall->dp_ifindex = dp_ifindex;
- return 0;
-}
-static void clear_stats(struct sw_flow *flow)
-{
- flow->used = 0;
- flow->tcp_flags = 0;
- flow->packet_count = 0;
- flow->byte_count = 0;
-}
+ nla = nla_nest_start(user_skb, OVS_PACKET_ATTR_KEY);
+ ovs_nla_put_flow(upcall_info->key, upcall_info->key, user_skb);
+ nla_nest_end(user_skb, nla);
-/* Called with genl_mutex. */
-static int expand_table(struct datapath *dp)
-{
- struct tbl *old_table = get_table_protected(dp);
- struct tbl *new_table;
+ if (upcall_info->userdata)
+ __nla_put(user_skb, OVS_PACKET_ATTR_USERDATA,
+ nla_len(upcall_info->userdata),
+ nla_data(upcall_info->userdata));
- new_table = tbl_expand(old_table);
- if (IS_ERR(new_table))
- return PTR_ERR(new_table);
+ /* Only reserve room for attribute header, packet data is added
+ * in skb_zerocopy() */
+ if (!(nla = nla_reserve(user_skb, OVS_PACKET_ATTR_PACKET, 0))) {
+ err = -ENOBUFS;
+ goto out;
+ }
+ nla->nla_len = nla_attr_size(skb->len);
+
+ err = skb_zerocopy(user_skb, skb, skb->len, hlen);
+ if (err)
+ goto out;
+
+ /* Pad OVS_PACKET_ATTR_PACKET if linear copy was performed */
+ if (!(dp->user_features & OVS_DP_F_UNALIGNED)) {
+ size_t plen = NLA_ALIGN(user_skb->len) - user_skb->len;
+
+ if (plen > 0)
+ memset(skb_put(user_skb, plen), 0, plen);
+ }
- rcu_assign_pointer(dp->table, new_table);
- tbl_deferred_destroy(old_table, NULL);
+ ((struct nlmsghdr *) user_skb->data)->nlmsg_len = user_skb->len;
- return 0;
+ err = genlmsg_unicast(ovs_dp_get_net(dp), user_skb, upcall_info->portid);
+out:
+ if (err)
+ skb_tx_error(skb);
+ kfree_skb(nskb);
+ return err;
}
-static int odp_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
+static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
{
- struct odp_header *odp_header = info->userhdr;
+ struct ovs_header *ovs_header = info->userhdr;
struct nlattr **a = info->attrs;
struct sw_flow_actions *acts;
struct sk_buff *packet;
struct sw_flow *flow;
struct datapath *dp;
struct ethhdr *eth;
- bool is_frag;
+ struct vport *input_vport;
int len;
int err;
err = -EINVAL;
- if (!a[ODP_PACKET_ATTR_PACKET] || !a[ODP_PACKET_ATTR_ACTIONS] ||
- nla_len(a[ODP_PACKET_ATTR_PACKET]) < ETH_HLEN)
- goto err;
-
- err = validate_actions(a[ODP_PACKET_ATTR_ACTIONS]);
- if (err)
+ if (!a[OVS_PACKET_ATTR_PACKET] || !a[OVS_PACKET_ATTR_KEY] ||
+ !a[OVS_PACKET_ATTR_ACTIONS])
goto err;
- len = nla_len(a[ODP_PACKET_ATTR_PACKET]);
+ len = nla_len(a[OVS_PACKET_ATTR_PACKET]);
packet = __dev_alloc_skb(NET_IP_ALIGN + len, GFP_KERNEL);
err = -ENOMEM;
if (!packet)
goto err;
skb_reserve(packet, NET_IP_ALIGN);
- memcpy(__skb_put(packet, len), nla_data(a[ODP_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);
/* Build an sw_flow for sending this packet. */
- flow = flow_alloc();
+ flow = ovs_flow_alloc();
err = PTR_ERR(flow);
if (IS_ERR(flow))
goto err_kfree_skb;
- err = flow_extract(packet, -1, &flow->key, &is_frag);
+ err = ovs_flow_extract(packet, -1, &flow->key);
if (err)
- goto err_flow_put;
- flow->tbl_node.hash = flow_hash(&flow->key);
+ goto err_flow_free;
- acts = flow_actions_alloc(a[ODP_PACKET_ATTR_ACTIONS]);
+ err = ovs_nla_get_flow_metadata(flow, a[OVS_PACKET_ATTR_KEY]);
+ if (err)
+ goto err_flow_free;
+ acts = ovs_nla_alloc_flow_actions(nla_len(a[OVS_PACKET_ATTR_ACTIONS]));
err = PTR_ERR(acts);
if (IS_ERR(acts))
- goto err_flow_put;
+ goto err_flow_free;
+
+ err = ovs_nla_copy_actions(a[OVS_PACKET_ATTR_ACTIONS],
+ &flow->key, 0, &acts);
rcu_assign_pointer(flow->sf_acts, acts);
+ if (err)
+ goto err_flow_free;
OVS_CB(packet)->flow = flow;
+ OVS_CB(packet)->pkt_key = &flow->key;
+ packet->priority = flow->key.phy.priority;
+ packet->mark = flow->key.phy.skb_mark;
rcu_read_lock();
- dp = get_dp(odp_header->dp_ifindex);
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
err = -ENODEV;
if (!dp)
goto err_unlock;
- err = execute_actions(dp, packet);
+
+ input_vport = ovs_vport_rcu(dp, flow->key.phy.in_port);
+ if (!input_vport)
+ input_vport = ovs_vport_rcu(dp, OVSP_LOCAL);
+
+ if (!input_vport)
+ goto err_unlock;
+
+ OVS_CB(packet)->input_vport = input_vport;
+
+ local_bh_disable();
+ err = ovs_execute_actions(dp, packet);
+ local_bh_enable();
rcu_read_unlock();
- flow_put(flow);
+ ovs_flow_free(flow, false);
return err;
err_unlock:
rcu_read_unlock();
-err_flow_put:
- flow_put(flow);
+err_flow_free:
+ ovs_flow_free(flow, false);
err_kfree_skb:
kfree_skb(packet);
err:
return err;
}
-static const struct nla_policy packet_policy[ODP_PACKET_ATTR_MAX + 1] = {
- [ODP_PACKET_ATTR_PACKET] = { .type = NLA_UNSPEC },
- [ODP_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED },
+static const struct nla_policy packet_policy[OVS_PACKET_ATTR_MAX + 1] = {
+ [OVS_PACKET_ATTR_PACKET] = { .len = ETH_HLEN },
+ [OVS_PACKET_ATTR_KEY] = { .type = NLA_NESTED },
+ [OVS_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED },
};
static struct genl_ops dp_packet_genl_ops[] = {
- { .cmd = ODP_PACKET_CMD_EXECUTE,
+ { .cmd = OVS_PACKET_CMD_EXECUTE,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = packet_policy,
- .doit = odp_packet_cmd_execute
+ .doit = ovs_packet_cmd_execute
}
};
-static void get_dp_stats(struct datapath *dp, struct odp_stats *stats)
+static struct genl_family dp_packet_genl_family = {
+ .id = GENL_ID_GENERATE,
+ .hdrsize = sizeof(struct ovs_header),
+ .name = OVS_PACKET_FAMILY,
+ .version = OVS_PACKET_VERSION,
+ .maxattr = OVS_PACKET_ATTR_MAX,
+ .netnsok = true,
+ .parallel_ops = true,
+ .ops = dp_packet_genl_ops,
+ .n_ops = ARRAY_SIZE(dp_packet_genl_ops),
+};
+
+static void get_dp_stats(struct datapath *dp, struct ovs_dp_stats *stats,
+ struct ovs_dp_megaflow_stats *mega_stats)
{
int i;
- stats->n_frags = stats->n_hit = stats->n_missed = stats->n_lost = 0;
+ memset(mega_stats, 0, sizeof(*mega_stats));
+
+ stats->n_flows = ovs_flow_tbl_count(&dp->table);
+ mega_stats->n_masks = ovs_flow_tbl_num_masks(&dp->table);
+
+ stats->n_hit = stats->n_missed = stats->n_lost = 0;
+
for_each_possible_cpu(i) {
const struct dp_stats_percpu *percpu_stats;
struct dp_stats_percpu local_stats;
- unsigned seqcount;
+ unsigned int start;
percpu_stats = per_cpu_ptr(dp->stats_percpu, i);
do {
- seqcount = read_seqcount_begin(&percpu_stats->seqlock);
+ start = u64_stats_fetch_begin_bh(&percpu_stats->sync);
local_stats = *percpu_stats;
- } while (read_seqcount_retry(&percpu_stats->seqlock, seqcount));
+ } while (u64_stats_fetch_retry_bh(&percpu_stats->sync, start));
- stats->n_frags += local_stats.n_frags;
stats->n_hit += local_stats.n_hit;
stats->n_missed += local_stats.n_missed;
stats->n_lost += local_stats.n_lost;
+ mega_stats->n_mask_hit += local_stats.n_mask_hit;
}
}
-/* MTU of the dp pseudo-device: ETH_DATA_LEN or the minimum of the ports.
- * Called with RTNL lock.
- */
-int dp_min_mtu(const struct datapath *dp)
+static size_t ovs_flow_cmd_msg_size(const struct sw_flow_actions *acts)
{
- struct vport *p;
- int mtu = 0;
-
- ASSERT_RTNL();
-
- list_for_each_entry (p, &dp->port_list, node) {
- int dev_mtu;
-
- /* Skip any internal ports, since that's what we're trying to
- * set. */
- if (is_internal_vport(p))
- continue;
-
- dev_mtu = vport_get_mtu(p);
- if (!dev_mtu)
- continue;
- if (!mtu || dev_mtu < mtu)
- mtu = dev_mtu;
- }
-
- return mtu ? mtu : ETH_DATA_LEN;
+ 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 */
}
-/* Sets the MTU of all datapath devices to the minimum of the ports
- * Called with RTNL lock.
- */
-void set_internal_devs_mtu(const struct datapath *dp)
-{
- struct vport *p;
- int mtu;
-
- ASSERT_RTNL();
-
- mtu = dp_min_mtu(dp);
-
- list_for_each_entry (p, &dp->port_list, node) {
- if (is_internal_vport(p))
- vport_set_mtu(p, mtu);
- }
-}
-
-static const struct nla_policy flow_policy[ODP_FLOW_ATTR_MAX + 1] = {
- [ODP_FLOW_ATTR_KEY] = { .type = NLA_NESTED },
- [ODP_FLOW_ATTR_ACTIONS] = { .type = NLA_NESTED },
- [ODP_FLOW_ATTR_CLEAR] = { .type = NLA_FLAG },
-};
-
-static struct genl_family dp_flow_genl_family = {
- .id = GENL_ID_GENERATE,
- .hdrsize = sizeof(struct odp_header),
- .name = ODP_FLOW_FAMILY,
- .version = 1,
- .maxattr = ODP_FLOW_ATTR_MAX
-};
-
-static struct genl_multicast_group dp_flow_multicast_group = {
- .name = ODP_FLOW_MCGROUP
-};
-
-/* Called with genl_lock. */
-static int odp_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
- struct sk_buff *skb, u32 pid, u32 seq, u32 flags, u8 cmd)
+/* Called with ovs_mutex or RCU read lock. */
+static int ovs_flow_cmd_fill_info(const struct sw_flow *flow, int dp_ifindex,
+ 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 odp_flow_stats stats;
- struct odp_header *odp_header;
- struct nlattr *nla;
+ struct nlattr *start;
+ struct ovs_flow_stats stats;
+ __be16 tcp_flags;
unsigned long used;
- u8 tcp_flags;
+ struct ovs_header *ovs_header;
+ struct nlattr *nla;
int err;
- sf_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
-
- odp_header = genlmsg_put(skb, pid, seq, &dp_flow_genl_family, flags, cmd);
- if (!odp_header)
+ ovs_header = genlmsg_put(skb, portid, seq, &dp_flow_genl_family, flags, cmd);
+ if (!ovs_header)
return -EMSGSIZE;
- odp_header->dp_ifindex = dp->dp_ifindex;
+ ovs_header->dp_ifindex = dp_ifindex;
- nla = nla_nest_start(skb, ODP_FLOW_ATTR_KEY);
+ /* Fill flow key. */
+ nla = nla_nest_start(skb, OVS_FLOW_ATTR_KEY);
if (!nla)
goto nla_put_failure;
- err = flow_to_nlattrs(&flow->key, skb);
+
+ err = ovs_nla_put_flow(&flow->unmasked_key, &flow->unmasked_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;
- stats.n_bytes = flow->byte_count;
- tcp_flags = flow->tcp_flags;
- spin_unlock_bh(&flow->lock);
+ nla = nla_nest_start(skb, OVS_FLOW_ATTR_MASK);
+ if (!nla)
+ goto nla_put_failure;
- if (used)
- NLA_PUT_U64(skb, ODP_FLOW_ATTR_USED, flow_used_time(used));
+ err = ovs_nla_put_flow(&flow->key, &flow->mask->key, skb);
+ if (err)
+ goto error;
- if (stats.n_packets)
- NLA_PUT(skb, ODP_FLOW_ATTR_STATS, sizeof(struct odp_flow_stats), &stats);
+ nla_nest_end(skb, nla);
- if (tcp_flags)
- NLA_PUT_U8(skb, ODP_FLOW_ATTR_TCP_FLAGS, tcp_flags);
+ ovs_flow_stats_get(flow, &stats, &used, &tcp_flags);
+
+ if (used &&
+ nla_put_u64(skb, OVS_FLOW_ATTR_USED, ovs_flow_used_time(used)))
+ goto nla_put_failure;
- /* If ODP_FLOW_ATTR_ACTIONS doesn't fit, skip dumping the actions if
+ if (stats.n_packets &&
+ nla_put(skb, OVS_FLOW_ATTR_STATS, sizeof(struct ovs_flow_stats), &stats))
+ goto nla_put_failure;
+
+ if ((u8)ntohs(tcp_flags) &&
+ nla_put_u8(skb, OVS_FLOW_ATTR_TCP_FLAGS, (u8)ntohs(tcp_flags)))
+ goto nla_put_failure;
+
+ /* If OVS_FLOW_ATTR_ACTIONS doesn't fit, skip dumping the actions if
* this is the first flow to be dumped into 'skb'. This is unusual for
* Netlink but individual action lists can be longer than
* NLMSG_GOODSIZE and thus entirely undumpable if we didn't do this.
* This can only fail for dump operations because the skb is always
* properly sized for single flows.
*/
- err = nla_put(skb, ODP_FLOW_ATTR_ACTIONS, sf_acts->actions_len,
- sf_acts->actions);
- if (err < 0 && skb_orig_len)
- goto error;
+ start = nla_nest_start(skb, OVS_FLOW_ATTR_ACTIONS);
+ if (start) {
+ const struct sw_flow_actions *sf_acts;
+
+ sf_acts = rcu_dereference_ovsl(flow->sf_acts);
+ err = ovs_nla_put_actions(sf_acts->actions,
+ sf_acts->actions_len, skb);
+
+ if (!err)
+ nla_nest_end(skb, start);
+ else {
+ if (skb_orig_len)
+ goto error;
- return genlmsg_end(skb, odp_header);
+ nla_nest_cancel(skb, start);
+ }
+ } else if (skb_orig_len)
+ goto nla_put_failure;
+
+ return genlmsg_end(skb, ovs_header);
nla_put_failure:
err = -EMSGSIZE;
error:
- genlmsg_cancel(skb, odp_header);
+ genlmsg_cancel(skb, ovs_header);
return err;
}
-static struct sk_buff *odp_flow_cmd_alloc_info(struct sw_flow *flow)
+/* May not be called with RCU read lock. */
+static struct sk_buff *ovs_flow_cmd_alloc_info(const struct sw_flow_actions *acts,
+ struct genl_info *info,
+ bool always)
{
- const struct sw_flow_actions *sf_acts;
- int len;
+ struct sk_buff *skb;
- sf_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
+ if (!always && !ovs_must_notify(info, &ovs_dp_flow_multicast_group))
+ return NULL;
- len = nla_total_size(FLOW_BUFSIZE); /* ODP_FLOW_ATTR_KEY */
- len += nla_total_size(sf_acts->actions_len); /* ODP_FLOW_ATTR_ACTIONS */
- len += nla_total_size(sizeof(struct odp_flow_stats)); /* ODP_FLOW_ATTR_STATS */
- len += nla_total_size(1); /* ODP_FLOW_ATTR_TCP_FLAGS */
- len += nla_total_size(8); /* ODP_FLOW_ATTR_USED */
- return genlmsg_new(NLMSG_ALIGN(sizeof(struct odp_header)) + len, GFP_KERNEL);
+ skb = genlmsg_new_unicast(ovs_flow_cmd_msg_size(acts), info, GFP_KERNEL);
+
+ if (!skb)
+ return ERR_PTR(-ENOMEM);
+
+ return skb;
}
-static struct sk_buff *odp_flow_cmd_build_info(struct sw_flow *flow, struct datapath *dp,
- u32 pid, u32 seq, u8 cmd)
+/* Called with ovs_mutex. */
+static struct sk_buff *ovs_flow_cmd_build_info(const struct sw_flow *flow,
+ int dp_ifindex,
+ struct genl_info *info, u8 cmd,
+ bool always)
{
struct sk_buff *skb;
int retval;
- skb = odp_flow_cmd_alloc_info(flow);
- if (!skb)
- return ERR_PTR(-ENOMEM);
+ skb = ovs_flow_cmd_alloc_info(ovsl_dereference(flow->sf_acts), info,
+ always);
+ if (!skb || IS_ERR(skb))
+ return skb;
- retval = odp_flow_cmd_fill_info(flow, dp, skb, pid, seq, 0, cmd);
+ retval = ovs_flow_cmd_fill_info(flow, dp_ifindex, skb,
+ info->snd_portid, info->snd_seq, 0,
+ cmd);
BUG_ON(retval < 0);
return skb;
}
-static int odp_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info)
+static int ovs_flow_cmd_new(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
- struct odp_header *odp_header = info->userhdr;
- struct tbl_node *flow_node;
- struct sw_flow_key key;
- struct sw_flow *flow;
+ struct ovs_header *ovs_header = info->userhdr;
+ struct sw_flow *flow, *new_flow;
+ struct sw_flow_mask mask;
struct sk_buff *reply;
struct datapath *dp;
- struct tbl *table;
- u32 hash;
+ struct sw_flow_actions *acts;
+ struct sw_flow_match match;
int error;
- /* Extract key. */
+ /* Must have key and actions. */
error = -EINVAL;
- if (!a[ODP_FLOW_ATTR_KEY])
+ if (!a[OVS_FLOW_ATTR_KEY])
goto error;
- error = flow_from_nlattrs(&key, a[ODP_FLOW_ATTR_KEY]);
- if (error)
+ if (!a[OVS_FLOW_ATTR_ACTIONS])
goto error;
- /* Validate actions. */
- if (a[ODP_FLOW_ATTR_ACTIONS]) {
- error = validate_actions(a[ODP_FLOW_ATTR_ACTIONS]);
- if (error)
- goto error;
- } else if (info->genlhdr->cmd == ODP_FLOW_CMD_NEW) {
- error = -EINVAL;
+ /* Most of the time we need to allocate a new flow, do it before
+ * locking. */
+ new_flow = ovs_flow_alloc();
+ if (IS_ERR(new_flow)) {
+ error = PTR_ERR(new_flow);
goto error;
}
- dp = get_dp(odp_header->dp_ifindex);
- error = -ENODEV;
- if (!dp)
- goto error;
+ /* Extract key. */
+ ovs_match_init(&match, &new_flow->unmasked_key, &mask);
+ error = ovs_nla_get_match(&match,
+ a[OVS_FLOW_ATTR_KEY], a[OVS_FLOW_ATTR_MASK]);
+ if (error)
+ goto err_kfree_flow;
- hash = flow_hash(&key);
- table = get_table_protected(dp);
- flow_node = tbl_lookup(table, &key, hash, flow_cmp);
- if (!flow_node) {
- struct sw_flow_actions *acts;
+ ovs_flow_mask_key(&new_flow->key, &new_flow->unmasked_key, &mask);
- /* Bail out if we're not allowed to create a new flow. */
- error = -ENOENT;
- if (info->genlhdr->cmd == ODP_FLOW_CMD_SET)
- goto error;
+ /* Validate actions. */
+ acts = ovs_nla_alloc_flow_actions(nla_len(a[OVS_FLOW_ATTR_ACTIONS]));
+ error = PTR_ERR(acts);
+ if (IS_ERR(acts))
+ goto err_kfree_flow;
- /* Expand table, if necessary, to make room. */
- if (tbl_count(table) >= tbl_n_buckets(table)) {
- error = expand_table(dp);
- if (error)
- goto error;
- table = get_table_protected(dp);
- }
+ error = ovs_nla_copy_actions(a[OVS_FLOW_ATTR_ACTIONS], &new_flow->key,
+ 0, &acts);
+ if (error) {
+ OVS_NLERR("Flow actions may not be safe on all matching packets.\n");
+ goto err_kfree_acts;
+ }
- /* Allocate flow. */
- flow = flow_alloc();
- if (IS_ERR(flow)) {
- error = PTR_ERR(flow);
- goto error;
- }
- flow->key = key;
- clear_stats(flow);
+ reply = ovs_flow_cmd_alloc_info(acts, info, false);
+ if (IS_ERR(reply)) {
+ error = PTR_ERR(reply);
+ goto err_kfree_acts;
+ }
- /* Obtain actions. */
- acts = flow_actions_alloc(a[ODP_FLOW_ATTR_ACTIONS]);
- error = PTR_ERR(acts);
- if (IS_ERR(acts))
- goto error_free_flow;
- rcu_assign_pointer(flow->sf_acts, acts);
+ ovs_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (unlikely(!dp)) {
+ error = -ENODEV;
+ goto err_unlock_ovs;
+ }
+ /* Check if this is a duplicate flow */
+ flow = ovs_flow_tbl_lookup(&dp->table, &new_flow->unmasked_key);
+ if (likely(!flow)) {
+ rcu_assign_pointer(new_flow->sf_acts, acts);
/* Put flow in bucket. */
- error = tbl_insert(table, &flow->tbl_node, hash);
- if (error)
- goto error_free_flow;
+ error = ovs_flow_tbl_insert(&dp->table, new_flow, &mask);
+ if (unlikely(error)) {
+ acts = NULL;
+ goto err_unlock_ovs;
+ }
- reply = odp_flow_cmd_build_info(flow, dp, info->snd_pid,
- info->snd_seq, ODP_FLOW_CMD_NEW);
+ if (unlikely(reply)) {
+ error = ovs_flow_cmd_fill_info(new_flow,
+ ovs_header->dp_ifindex,
+ reply, info->snd_portid,
+ info->snd_seq, 0,
+ OVS_FLOW_CMD_NEW);
+ BUG_ON(error < 0);
+ }
+ ovs_unlock();
} else {
- /* We found a matching flow. */
struct sw_flow_actions *old_acts;
/* Bail out if we're not allowed to modify an existing flow.
* request. We also accept NLM_F_EXCL in case that bug ever
* gets fixed.
*/
- error = -EEXIST;
- if (info->genlhdr->cmd == ODP_FLOW_CMD_NEW &&
- info->nlhdr->nlmsg_flags & (NLM_F_CREATE | NLM_F_EXCL))
+ if (unlikely(info->nlhdr->nlmsg_flags & (NLM_F_CREATE
+ | NLM_F_EXCL))) {
+ error = -EEXIST;
+ goto err_unlock_ovs;
+ }
+ /* The unmasked key has to be the same for flow updates. */
+ if (unlikely(!ovs_flow_cmp_unmasked_key(flow, &match))) {
+ error = -EEXIST;
+ goto err_unlock_ovs;
+ }
+ /* Update actions. */
+ old_acts = ovsl_dereference(flow->sf_acts);
+ rcu_assign_pointer(flow->sf_acts, acts);
+
+ if (unlikely(reply)) {
+ error = ovs_flow_cmd_fill_info(flow,
+ ovs_header->dp_ifindex,
+ reply, info->snd_portid,
+ info->snd_seq, 0,
+ OVS_FLOW_CMD_NEW);
+ BUG_ON(error < 0);
+ }
+ ovs_unlock();
+
+ ovs_nla_free_flow_actions(old_acts);
+ ovs_flow_free(new_flow, false);
+ }
+
+ if (reply)
+ ovs_notify(&dp_flow_genl_family, &ovs_dp_flow_multicast_group, reply, info);
+ return 0;
+
+err_unlock_ovs:
+ ovs_unlock();
+ kfree_skb(reply);
+err_kfree_acts:
+ kfree(acts);
+err_kfree_flow:
+ ovs_flow_free(new_flow, false);
+error:
+ return error;
+}
+
+static int ovs_flow_cmd_set(struct sk_buff *skb, struct genl_info *info)
+{
+ struct nlattr **a = info->attrs;
+ struct ovs_header *ovs_header = info->userhdr;
+ struct sw_flow_key key, masked_key;
+ struct sw_flow *flow;
+ struct sw_flow_mask mask;
+ struct sk_buff *reply = NULL;
+ struct datapath *dp;
+ struct sw_flow_actions *old_acts = NULL, *acts = NULL;
+ struct sw_flow_match match;
+ int error;
+
+ /* Extract key. */
+ error = -EINVAL;
+ if (!a[OVS_FLOW_ATTR_KEY])
+ goto error;
+
+ ovs_match_init(&match, &key, &mask);
+ error = ovs_nla_get_match(&match,
+ a[OVS_FLOW_ATTR_KEY], a[OVS_FLOW_ATTR_MASK]);
+ if (error)
+ goto error;
+
+ /* Validate actions. */
+ if (a[OVS_FLOW_ATTR_ACTIONS]) {
+ acts = ovs_nla_alloc_flow_actions(nla_len(a[OVS_FLOW_ATTR_ACTIONS]));
+ error = PTR_ERR(acts);
+ if (IS_ERR(acts))
goto error;
- /* Update actions. */
- flow = flow_cast(flow_node);
- old_acts = rcu_dereference_protected(flow->sf_acts,
- lockdep_genl_is_held());
- if (a[ODP_FLOW_ATTR_ACTIONS] &&
- (old_acts->actions_len != nla_len(a[ODP_FLOW_ATTR_ACTIONS]) ||
- memcmp(old_acts->actions, nla_data(a[ODP_FLOW_ATTR_ACTIONS]),
- old_acts->actions_len))) {
- struct sw_flow_actions *new_acts;
-
- new_acts = flow_actions_alloc(a[ODP_FLOW_ATTR_ACTIONS]);
- error = PTR_ERR(new_acts);
- if (IS_ERR(new_acts))
- goto error;
+ ovs_flow_mask_key(&masked_key, &key, &mask);
+ error = ovs_nla_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_acts;
+ }
+ }
- rcu_assign_pointer(flow->sf_acts, new_acts);
- flow_deferred_free_acts(old_acts);
+ /* Can allocate before locking if have acts. */
+ if (acts) {
+ reply = ovs_flow_cmd_alloc_info(acts, info, false);
+ if (IS_ERR(reply)) {
+ error = PTR_ERR(reply);
+ goto err_kfree_acts;
}
+ }
- reply = odp_flow_cmd_build_info(flow, dp, info->snd_pid,
- info->snd_seq, ODP_FLOW_CMD_NEW);
+ ovs_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (unlikely(!dp)) {
+ error = -ENODEV;
+ goto err_unlock_ovs;
+ }
+ /* Check that the flow exists. */
+ flow = ovs_flow_tbl_lookup(&dp->table, &key);
+ if (unlikely(!flow)) {
+ error = -ENOENT;
+ goto err_unlock_ovs;
+ }
+ /* The unmasked key has to be the same for flow updates. */
+ if (unlikely(!ovs_flow_cmp_unmasked_key(flow, &match))) {
+ error = -EEXIST;
+ goto err_unlock_ovs;
+ }
+ /* Update actions, if present. */
+ if (likely(acts)) {
+ old_acts = ovsl_dereference(flow->sf_acts);
+ rcu_assign_pointer(flow->sf_acts, acts);
- /* Clear stats. */
- if (a[ODP_FLOW_ATTR_CLEAR]) {
- spin_lock_bh(&flow->lock);
- clear_stats(flow);
- spin_unlock_bh(&flow->lock);
+ if (unlikely(reply)) {
+ error = ovs_flow_cmd_fill_info(flow,
+ ovs_header->dp_ifindex,
+ reply, info->snd_portid,
+ info->snd_seq, 0,
+ OVS_FLOW_CMD_NEW);
+ BUG_ON(error < 0);
+ }
+ } else {
+ /* Could not alloc without acts before locking. */
+ reply = ovs_flow_cmd_build_info(flow, ovs_header->dp_ifindex,
+ info, OVS_FLOW_CMD_NEW, false);
+ if (unlikely(IS_ERR(reply))) {
+ error = PTR_ERR(reply);
+ goto err_unlock_ovs;
}
}
- if (!IS_ERR(reply))
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_flow_multicast_group.id, info->nlhdr, GFP_KERNEL);
- else
- netlink_set_err(INIT_NET_GENL_SOCK, 0,
- dp_flow_multicast_group.id, PTR_ERR(reply));
+ /* Clear stats. */
+ if (a[OVS_FLOW_ATTR_CLEAR])
+ ovs_flow_stats_clear(flow);
+ ovs_unlock();
+
+ if (reply)
+ ovs_notify(&dp_flow_genl_family, &ovs_dp_flow_multicast_group, reply, info);
+ if (old_acts)
+ ovs_nla_free_flow_actions(old_acts);
return 0;
-error_free_flow:
- flow_put(flow);
+err_unlock_ovs:
+ ovs_unlock();
+ kfree_skb(reply);
+err_kfree_acts:
+ kfree(acts);
error:
return error;
}
-static int odp_flow_cmd_get(struct sk_buff *skb, struct genl_info *info)
+static int ovs_flow_cmd_get(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
- struct odp_header *odp_header = info->userhdr;
+ struct ovs_header *ovs_header = info->userhdr;
struct sw_flow_key key;
- struct tbl_node *flow_node;
struct sk_buff *reply;
struct sw_flow *flow;
struct datapath *dp;
- struct tbl *table;
+ struct sw_flow_match match;
int err;
- if (!a[ODP_FLOW_ATTR_KEY])
+ if (!a[OVS_FLOW_ATTR_KEY]) {
+ OVS_NLERR("Flow get message rejected, Key attribute missing.\n");
return -EINVAL;
- err = flow_from_nlattrs(&key, a[ODP_FLOW_ATTR_KEY]);
+ }
+
+ ovs_match_init(&match, &key, NULL);
+ err = ovs_nla_get_match(&match, a[OVS_FLOW_ATTR_KEY], NULL);
if (err)
return err;
- dp = get_dp(odp_header->dp_ifindex);
- if (!dp)
- return -ENODEV;
+ ovs_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (!dp) {
+ err = -ENODEV;
+ goto unlock;
+ }
- table = get_table_protected(dp);
- flow_node = tbl_lookup(table, &key, flow_hash(&key), flow_cmp);
- if (!flow_node)
- return -ENOENT;
+ flow = ovs_flow_tbl_lookup(&dp->table, &key);
+ if (!flow || !ovs_flow_cmp_unmasked_key(flow, &match)) {
+ err = -ENOENT;
+ goto unlock;
+ }
- flow = flow_cast(flow_node);
- reply = odp_flow_cmd_build_info(flow, dp, info->snd_pid, info->snd_seq, ODP_FLOW_CMD_NEW);
- if (IS_ERR(reply))
- return PTR_ERR(reply);
+ reply = ovs_flow_cmd_build_info(flow, ovs_header->dp_ifindex, info,
+ OVS_FLOW_CMD_NEW, true);
+ if (IS_ERR(reply)) {
+ err = PTR_ERR(reply);
+ goto unlock;
+ }
+ ovs_unlock();
return genlmsg_reply(reply, info);
+unlock:
+ ovs_unlock();
+ return err;
}
-static int odp_flow_cmd_del(struct sk_buff *skb, struct genl_info *info)
+static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
- struct odp_header *odp_header = info->userhdr;
+ struct ovs_header *ovs_header = info->userhdr;
struct sw_flow_key key;
- struct tbl_node *flow_node;
struct sk_buff *reply;
struct sw_flow *flow;
struct datapath *dp;
- struct tbl *table;
+ struct sw_flow_match match;
int err;
- if (!a[ODP_FLOW_ATTR_KEY])
- return flush_flows(odp_header->dp_ifindex);
- err = flow_from_nlattrs(&key, a[ODP_FLOW_ATTR_KEY]);
- if (err)
- return err;
-
- dp = get_dp(odp_header->dp_ifindex);
- if (!dp)
- return -ENODEV;
-
- table = get_table_protected(dp);
- flow_node = tbl_lookup(table, &key, flow_hash(&key), flow_cmp);
- if (!flow_node)
- return -ENOENT;
- flow = flow_cast(flow_node);
-
- reply = odp_flow_cmd_alloc_info(flow);
- if (!reply)
- return -ENOMEM;
-
- err = tbl_remove(table, flow_node);
- if (err) {
- kfree_skb(reply);
- return err;
+ if (likely(a[OVS_FLOW_ATTR_KEY])) {
+ ovs_match_init(&match, &key, NULL);
+ err = ovs_nla_get_match(&match, a[OVS_FLOW_ATTR_KEY], NULL);
+ if (unlikely(err))
+ return err;
}
- err = odp_flow_cmd_fill_info(flow, dp, reply, info->snd_pid,
- info->snd_seq, 0, ODP_FLOW_CMD_DEL);
- BUG_ON(err < 0);
+ ovs_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (unlikely(!dp)) {
+ err = -ENODEV;
+ goto unlock;
+ }
+ if (unlikely(!a[OVS_FLOW_ATTR_KEY])) {
+ err = ovs_flow_tbl_flush(&dp->table);
+ goto unlock;
+ }
+ flow = ovs_flow_tbl_lookup(&dp->table, &key);
+ if (unlikely(!flow || !ovs_flow_cmp_unmasked_key(flow, &match))) {
+ err = -ENOENT;
+ goto unlock;
+ }
- flow_deferred_free(flow);
+ ovs_flow_tbl_remove(&dp->table, flow);
+ ovs_unlock();
+
+ reply = ovs_flow_cmd_alloc_info((const struct sw_flow_actions __force *)flow->sf_acts,
+ info, false);
+
+ if (likely(reply)) {
+ if (likely(!IS_ERR(reply))) {
+ rcu_read_lock(); /* Keep RCU checker happy. */
+ err = ovs_flow_cmd_fill_info(flow,
+ ovs_header->dp_ifindex,
+ reply, info->snd_portid,
+ info->snd_seq, 0,
+ OVS_FLOW_CMD_DEL);
+ rcu_read_unlock();
+ BUG_ON(err < 0);
+ ovs_notify(&dp_flow_genl_family, &ovs_dp_flow_multicast_group, reply, info);
+ } else {
+ genl_set_err(&dp_flow_genl_family, sock_net(skb->sk), 0,
+ GROUP_ID(&ovs_dp_flow_multicast_group), PTR_ERR(reply));
+
+ }
+ }
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_flow_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_flow_free(flow, true);
return 0;
+unlock:
+ ovs_unlock();
+ return err;
}
-static int odp_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
+static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
- struct odp_header *odp_header = genlmsg_data(nlmsg_data(cb->nlh));
+ struct ovs_header *ovs_header = genlmsg_data(nlmsg_data(cb->nlh));
+ struct table_instance *ti;
struct datapath *dp;
- dp = get_dp(odp_header->dp_ifindex);
- if (!dp)
+ rcu_read_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (!dp) {
+ rcu_read_unlock();
return -ENODEV;
+ }
+ ti = rcu_dereference(dp->table.ti);
for (;;) {
- struct tbl_node *flow_node;
struct sw_flow *flow;
u32 bucket, obj;
bucket = cb->args[0];
obj = cb->args[1];
- flow_node = tbl_next(get_table_protected(dp), &bucket, &obj);
- if (!flow_node)
+ flow = ovs_flow_tbl_dump_next(ti, &bucket, &obj);
+ if (!flow)
break;
- flow = flow_cast(flow_node);
- if (odp_flow_cmd_fill_info(flow, dp, skb, NETLINK_CB(cb->skb).pid,
+ if (ovs_flow_cmd_fill_info(flow, ovs_header->dp_ifindex, skb,
+ NETLINK_CB(cb->skb).portid,
cb->nlh->nlmsg_seq, NLM_F_MULTI,
- ODP_FLOW_CMD_NEW) < 0)
+ OVS_FLOW_CMD_NEW) < 0)
break;
cb->args[0] = bucket;
cb->args[1] = obj;
}
+ rcu_read_unlock();
return skb->len;
}
+static const struct nla_policy flow_policy[OVS_FLOW_ATTR_MAX + 1] = {
+ [OVS_FLOW_ATTR_KEY] = { .type = NLA_NESTED },
+ [OVS_FLOW_ATTR_ACTIONS] = { .type = NLA_NESTED },
+ [OVS_FLOW_ATTR_CLEAR] = { .type = NLA_FLAG },
+};
+
static struct genl_ops dp_flow_genl_ops[] = {
- { .cmd = ODP_FLOW_CMD_NEW,
+ { .cmd = OVS_FLOW_CMD_NEW,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = flow_policy,
- .doit = odp_flow_cmd_new_or_set
+ .doit = ovs_flow_cmd_new
},
- { .cmd = ODP_FLOW_CMD_DEL,
+ { .cmd = OVS_FLOW_CMD_DEL,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = flow_policy,
- .doit = odp_flow_cmd_del
+ .doit = ovs_flow_cmd_del
},
- { .cmd = ODP_FLOW_CMD_GET,
+ { .cmd = OVS_FLOW_CMD_GET,
.flags = 0, /* OK for unprivileged users. */
.policy = flow_policy,
- .doit = odp_flow_cmd_get,
- .dumpit = odp_flow_cmd_dump
+ .doit = ovs_flow_cmd_get,
+ .dumpit = ovs_flow_cmd_dump
},
- { .cmd = ODP_FLOW_CMD_SET,
+ { .cmd = OVS_FLOW_CMD_SET,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = flow_policy,
- .doit = odp_flow_cmd_new_or_set,
+ .doit = ovs_flow_cmd_set,
},
};
-static const struct nla_policy datapath_policy[ODP_DP_ATTR_MAX + 1] = {
-#ifdef HAVE_NLA_NUL_STRING
- [ODP_DP_ATTR_NAME] = { .type = NLA_NUL_STRING, .len = IFNAMSIZ - 1 },
-#endif
- [ODP_DP_ATTR_IPV4_FRAGS] = { .type = NLA_U32 },
- [ODP_DP_ATTR_SAMPLING] = { .type = NLA_U32 },
-};
-
-static struct genl_family dp_datapath_genl_family = {
+static struct genl_family dp_flow_genl_family = {
.id = GENL_ID_GENERATE,
- .hdrsize = sizeof(struct odp_header),
- .name = ODP_DATAPATH_FAMILY,
- .version = 1,
- .maxattr = ODP_DP_ATTR_MAX
+ .hdrsize = sizeof(struct ovs_header),
+ .name = OVS_FLOW_FAMILY,
+ .version = OVS_FLOW_VERSION,
+ .maxattr = OVS_FLOW_ATTR_MAX,
+ .netnsok = true,
+ .parallel_ops = true,
+ .ops = dp_flow_genl_ops,
+ .n_ops = ARRAY_SIZE(dp_flow_genl_ops),
+ .mcgrps = &ovs_dp_flow_multicast_group,
+ .n_mcgrps = 1,
};
-static struct genl_multicast_group dp_datapath_multicast_group = {
- .name = ODP_DATAPATH_MCGROUP
-};
+static size_t ovs_dp_cmd_msg_size(void)
+{
+ size_t msgsize = NLMSG_ALIGN(sizeof(struct ovs_header));
-static int odp_dp_cmd_fill_info(struct datapath *dp, struct sk_buff *skb,
- u32 pid, u32 seq, u32 flags, u8 cmd)
+ msgsize += nla_total_size(IFNAMSIZ);
+ msgsize += nla_total_size(sizeof(struct ovs_dp_stats));
+ msgsize += nla_total_size(sizeof(struct ovs_dp_megaflow_stats));
+ msgsize += nla_total_size(sizeof(u32)); /* OVS_DP_ATTR_USER_FEATURES */
+
+ return msgsize;
+}
+
+/* Called with ovs_mutex or RCU read lock. */
+static int ovs_dp_cmd_fill_info(struct datapath *dp, struct sk_buff *skb,
+ u32 portid, u32 seq, u32 flags, u8 cmd)
{
- struct odp_header *odp_header;
- struct nlattr *nla;
+ struct ovs_header *ovs_header;
+ struct ovs_dp_stats dp_stats;
+ struct ovs_dp_megaflow_stats dp_megaflow_stats;
int err;
- odp_header = genlmsg_put(skb, pid, seq, &dp_datapath_genl_family,
+ ovs_header = genlmsg_put(skb, portid, seq, &dp_datapath_genl_family,
flags, cmd);
- if (!odp_header)
+ if (!ovs_header)
goto error;
- odp_header->dp_ifindex = dp->dp_ifindex;
+ ovs_header->dp_ifindex = get_dpifindex(dp);
- rcu_read_lock();
- err = nla_put_string(skb, ODP_DP_ATTR_NAME, dp_name(dp));
- rcu_read_unlock();
+ err = nla_put_string(skb, OVS_DP_ATTR_NAME, ovs_dp_name(dp));
if (err)
goto nla_put_failure;
- nla = nla_reserve(skb, ODP_DP_ATTR_STATS, sizeof(struct odp_stats));
- if (!nla)
+ get_dp_stats(dp, &dp_stats, &dp_megaflow_stats);
+ if (nla_put(skb, OVS_DP_ATTR_STATS, sizeof(struct ovs_dp_stats),
+ &dp_stats))
goto nla_put_failure;
- get_dp_stats(dp, nla_data(nla));
-
- NLA_PUT_U32(skb, ODP_DP_ATTR_IPV4_FRAGS,
- dp->drop_frags ? ODP_DP_FRAG_DROP : ODP_DP_FRAG_ZERO);
- if (dp->sflow_probability)
- NLA_PUT_U32(skb, ODP_DP_ATTR_SAMPLING, dp->sflow_probability);
+ if (nla_put(skb, OVS_DP_ATTR_MEGAFLOW_STATS,
+ sizeof(struct ovs_dp_megaflow_stats),
+ &dp_megaflow_stats))
+ goto nla_put_failure;
- nla = nla_nest_start(skb, ODP_DP_ATTR_MCGROUPS);
- if (!nla)
+ if (nla_put_u32(skb, OVS_DP_ATTR_USER_FEATURES, dp->user_features))
goto nla_put_failure;
- NLA_PUT_U32(skb, ODP_PACKET_CMD_MISS, packet_mc_group(dp, ODP_PACKET_CMD_MISS));
- NLA_PUT_U32(skb, ODP_PACKET_CMD_ACTION, packet_mc_group(dp, ODP_PACKET_CMD_ACTION));
- NLA_PUT_U32(skb, ODP_PACKET_CMD_SAMPLE, packet_mc_group(dp, ODP_PACKET_CMD_SAMPLE));
- nla_nest_end(skb, nla);
- return genlmsg_end(skb, odp_header);
+ return genlmsg_end(skb, ovs_header);
nla_put_failure:
- genlmsg_cancel(skb, odp_header);
+ genlmsg_cancel(skb, ovs_header);
error:
return -EMSGSIZE;
}
-static struct sk_buff *odp_dp_cmd_build_info(struct datapath *dp, u32 pid,
- u32 seq, u8 cmd)
+static struct sk_buff *ovs_dp_cmd_alloc_info(struct genl_info *info)
{
- struct sk_buff *skb;
- int retval;
-
- skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
- if (!skb)
- return ERR_PTR(-ENOMEM);
-
- retval = odp_dp_cmd_fill_info(dp, skb, pid, seq, 0, cmd);
- if (retval < 0) {
- kfree_skb(skb);
- return ERR_PTR(retval);
- }
- return skb;
+ return genlmsg_new_unicast(ovs_dp_cmd_msg_size(), info, GFP_KERNEL);
}
-static int odp_dp_cmd_validate(struct nlattr *a[ODP_DP_ATTR_MAX + 1])
+/* Called with rcu_read_lock or ovs_mutex. */
+static struct datapath *lookup_datapath(struct net *net,
+ struct ovs_header *ovs_header,
+ struct nlattr *a[OVS_DP_ATTR_MAX + 1])
{
- if (a[ODP_DP_ATTR_IPV4_FRAGS]) {
- u32 frags = nla_get_u32(a[ODP_DP_ATTR_IPV4_FRAGS]);
+ struct datapath *dp;
- if (frags != ODP_DP_FRAG_ZERO && frags != ODP_DP_FRAG_DROP)
- return -EINVAL;
- }
+ if (!a[OVS_DP_ATTR_NAME])
+ dp = get_dp(net, ovs_header->dp_ifindex);
+ else {
+ struct vport *vport;
- return CHECK_NUL_STRING(a[ODP_DP_ATTR_NAME], IFNAMSIZ - 1);
+ vport = ovs_vport_locate(net, nla_data(a[OVS_DP_ATTR_NAME]));
+ dp = vport && vport->port_no == OVSP_LOCAL ? vport->dp : NULL;
+ }
+ return dp ? dp : ERR_PTR(-ENODEV);
}
-/* Called with genl_mutex and optionally with RTNL lock also. */
-static struct datapath *lookup_datapath(struct odp_header *odp_header, struct nlattr *a[ODP_DP_ATTR_MAX + 1])
+static void ovs_dp_reset_user_features(struct sk_buff *skb, struct genl_info *info)
{
struct datapath *dp;
- if (!a[ODP_DP_ATTR_NAME])
- dp = get_dp(odp_header->dp_ifindex);
- else {
- struct vport *vport;
+ dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
+ if (IS_ERR(dp))
+ return;
- rcu_read_lock();
- vport = vport_locate(nla_data(a[ODP_DP_ATTR_NAME]));
- dp = vport && vport->port_no == ODPP_LOCAL ? vport->dp : NULL;
- rcu_read_unlock();
- }
- return dp ? dp : ERR_PTR(-ENODEV);
+ WARN(dp->user_features, "Dropping previously announced user features\n");
+ dp->user_features = 0;
}
-/* Called with genl_mutex. */
-static void change_datapath(struct datapath *dp, struct nlattr *a[ODP_DP_ATTR_MAX + 1])
+static void ovs_dp_change(struct datapath *dp, struct nlattr **a)
{
- if (a[ODP_DP_ATTR_IPV4_FRAGS])
- dp->drop_frags = nla_get_u32(a[ODP_DP_ATTR_IPV4_FRAGS]) == ODP_DP_FRAG_DROP;
- if (a[ODP_DP_ATTR_SAMPLING])
- dp->sflow_probability = nla_get_u32(a[ODP_DP_ATTR_SAMPLING]);
+ if (a[OVS_DP_ATTR_USER_FEATURES])
+ dp->user_features = nla_get_u32(a[OVS_DP_ATTR_USER_FEATURES]);
}
-static int odp_dp_cmd_new(struct sk_buff *skb, struct genl_info *info)
+static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
struct vport_parms parms;
struct sk_buff *reply;
struct datapath *dp;
struct vport *vport;
- int err;
+ struct ovs_net *ovs_net;
+ int err, i;
err = -EINVAL;
- if (!a[ODP_DP_ATTR_NAME])
- goto err;
-
- err = odp_dp_cmd_validate(a);
- if (err)
+ if (!a[OVS_DP_ATTR_NAME] || !a[OVS_DP_ATTR_UPCALL_PID])
goto err;
- rtnl_lock();
- err = -ENODEV;
- if (!try_module_get(THIS_MODULE))
- goto err_unlock_rtnl;
+ reply = ovs_dp_cmd_alloc_info(info);
+ if (!reply)
+ return -ENOMEM;
err = -ENOMEM;
dp = kzalloc(sizeof(*dp), GFP_KERNEL);
if (dp == NULL)
- goto err_put_module;
- INIT_LIST_HEAD(&dp->port_list);
+ goto err_free_reply;
- /* Initialize kobject for bridge. This will be added as
- * /sys/class/net/<devname>/brif later, if sysfs is enabled. */
- dp->ifobj.kset = NULL;
- kobject_init(&dp->ifobj, &dp_ktype);
+ ovs_dp_set_net(dp, hold_net(sock_net(skb->sk)));
/* Allocate table. */
- err = -ENOMEM;
- rcu_assign_pointer(dp->table, tbl_create(TBL_MIN_BUCKETS));
- if (!dp->table)
+ err = ovs_flow_tbl_init(&dp->table);
+ if (err)
goto err_free_dp;
+ dp->stats_percpu = alloc_percpu(struct dp_stats_percpu);
+ if (!dp->stats_percpu) {
+ err = -ENOMEM;
+ goto err_destroy_table;
+ }
+
+ for_each_possible_cpu(i) {
+ struct dp_stats_percpu *dpath_stats;
+ dpath_stats = per_cpu_ptr(dp->stats_percpu, i);
+ u64_stats_init(&dpath_stats->sync);
+ }
+
+ dp->ports = kmalloc(DP_VPORT_HASH_BUCKETS * sizeof(struct hlist_head),
+ GFP_KERNEL);
+ if (!dp->ports) {
+ err = -ENOMEM;
+ goto err_destroy_percpu;
+ }
+
+ for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++)
+ INIT_HLIST_HEAD(&dp->ports[i]);
+
/* Set up our datapath device. */
- parms.name = nla_data(a[ODP_DP_ATTR_NAME]);
- parms.type = ODP_VPORT_TYPE_INTERNAL;
+ parms.name = nla_data(a[OVS_DP_ATTR_NAME]);
+ parms.type = OVS_VPORT_TYPE_INTERNAL;
parms.options = NULL;
parms.dp = dp;
- parms.port_no = ODPP_LOCAL;
+ parms.port_no = OVSP_LOCAL;
+ parms.upcall_portids = a[OVS_DP_ATTR_UPCALL_PID];
+
+ ovs_dp_change(dp, a);
+
+ /* So far only local changes have been made, now need the lock. */
+ ovs_lock();
+
vport = new_vport(&parms);
if (IS_ERR(vport)) {
err = PTR_ERR(vport);
if (err == -EBUSY)
err = -EEXIST;
- goto err_destroy_table;
- }
- dp->dp_ifindex = vport_get_ifindex(vport);
+ if (err == -EEXIST) {
+ /* An outdated user space instance that does not understand
+ * the concept of user_features has attempted to create a new
+ * datapath and is likely to reuse it. Drop all user features.
+ */
+ if (info->genlhdr->version < OVS_DP_VER_FEATURES)
+ ovs_dp_reset_user_features(skb, info);
+ }
- dp->drop_frags = 0;
- dp->stats_percpu = alloc_percpu(struct dp_stats_percpu);
- if (!dp->stats_percpu) {
- err = -ENOMEM;
- goto err_destroy_local_port;
+ goto err_destroy_ports_array;
}
- change_datapath(dp, a);
-
- reply = odp_dp_cmd_build_info(dp, info->snd_pid, info->snd_seq, ODP_DP_CMD_NEW);
- err = PTR_ERR(reply);
- if (IS_ERR(reply))
- goto err_destroy_local_port;
+ err = ovs_dp_cmd_fill_info(dp, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_DP_CMD_NEW);
+ BUG_ON(err < 0);
- list_add_tail(&dp->list_node, &dps);
- dp_sysfs_add_dp(dp);
+ ovs_net = net_generic(ovs_dp_get_net(dp), ovs_net_id);
+ list_add_tail_rcu(&dp->list_node, &ovs_net->dps);
- rtnl_unlock();
+ ovs_unlock();
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_datapath_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_notify(&dp_datapath_genl_family, &ovs_dp_datapath_multicast_group, reply, info);
return 0;
-err_destroy_local_port:
- dp_detach_port(get_vport_protected(dp, ODPP_LOCAL));
+err_destroy_ports_array:
+ ovs_unlock();
+ kfree(dp->ports);
+err_destroy_percpu:
+ free_percpu(dp->stats_percpu);
err_destroy_table:
- tbl_destroy(get_table_protected(dp), NULL);
+ ovs_flow_tbl_destroy(&dp->table);
err_free_dp:
+ release_net(ovs_dp_get_net(dp));
kfree(dp);
-err_put_module:
- module_put(THIS_MODULE);
-err_unlock_rtnl:
- rtnl_unlock();
+err_free_reply:
+ kfree_skb(reply);
err:
return err;
}
-static int odp_dp_cmd_del(struct sk_buff *skb, struct genl_info *info)
+/* Called with ovs_mutex. */
+static void __dp_destroy(struct datapath *dp)
+{
+ int i;
+
+ for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) {
+ struct vport *vport;
+ struct hlist_node *n;
+
+ hlist_for_each_entry_safe(vport, n, &dp->ports[i], dp_hash_node)
+ if (vport->port_no != OVSP_LOCAL)
+ ovs_dp_detach_port(vport);
+ }
+
+ list_del_rcu(&dp->list_node);
+
+ /* OVSP_LOCAL is datapath internal port. We need to make sure that
+ * all ports in datapath are destroyed first before freeing datapath.
+ */
+ ovs_dp_detach_port(ovs_vport_ovsl(dp, OVSP_LOCAL));
+
+ /* RCU destroy the flow table */
+ call_rcu(&dp->rcu, destroy_dp_rcu);
+}
+
+static int ovs_dp_cmd_del(struct sk_buff *skb, struct genl_info *info)
{
- struct vport *vport, *next_vport;
struct sk_buff *reply;
struct datapath *dp;
int err;
- err = odp_dp_cmd_validate(info->attrs);
- if (err)
- goto exit;
+ reply = ovs_dp_cmd_alloc_info(info);
+ if (!reply)
+ return -ENOMEM;
- rtnl_lock();
- dp = lookup_datapath(info->userhdr, info->attrs);
+ ovs_lock();
+ dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
err = PTR_ERR(dp);
if (IS_ERR(dp))
- goto exit_unlock;
-
- reply = odp_dp_cmd_build_info(dp, info->snd_pid, info->snd_seq, ODP_DP_CMD_DEL);
- err = PTR_ERR(reply);
- if (IS_ERR(reply))
- goto exit_unlock;
-
- list_for_each_entry_safe (vport, next_vport, &dp->port_list, node)
- if (vport->port_no != ODPP_LOCAL)
- dp_detach_port(vport);
+ goto err_unlock_free;
- dp_sysfs_del_dp(dp);
- list_del(&dp->list_node);
- dp_detach_port(get_vport_protected(dp, ODPP_LOCAL));
-
- /* 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();
-
- call_rcu(&dp->rcu, destroy_dp_rcu);
- module_put(THIS_MODULE);
+ err = ovs_dp_cmd_fill_info(dp, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_DP_CMD_DEL);
+ BUG_ON(err < 0);
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_datapath_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ __dp_destroy(dp);
+ ovs_unlock();
+ ovs_notify(&dp_datapath_genl_family, &ovs_dp_datapath_multicast_group, reply, info);
return 0;
-exit_unlock:
- rtnl_unlock();
-exit:
+err_unlock_free:
+ ovs_unlock();
+ kfree_skb(reply);
return err;
}
-static int odp_dp_cmd_set(struct sk_buff *skb, struct genl_info *info)
+static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info)
{
struct sk_buff *reply;
struct datapath *dp;
int err;
- err = odp_dp_cmd_validate(info->attrs);
- if (err)
- return err;
+ reply = ovs_dp_cmd_alloc_info(info);
+ if (!reply)
+ return -ENOMEM;
- dp = lookup_datapath(info->userhdr, info->attrs);
+ 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 err_unlock_free;
- change_datapath(dp, info->attrs);
+ ovs_dp_change(dp, info->attrs);
- reply = odp_dp_cmd_build_info(dp, info->snd_pid, info->snd_seq, ODP_DP_CMD_NEW);
- if (IS_ERR(reply)) {
- err = PTR_ERR(reply);
- netlink_set_err(INIT_NET_GENL_SOCK, 0,
- dp_datapath_multicast_group.id, err);
- return 0;
- }
+ err = ovs_dp_cmd_fill_info(dp, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_DP_CMD_NEW);
+ BUG_ON(err < 0);
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_datapath_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_unlock();
+ ovs_notify(&dp_datapath_genl_family, &ovs_dp_datapath_multicast_group, reply, info);
return 0;
+
+err_unlock_free:
+ ovs_unlock();
+ kfree_skb(reply);
+ return err;
}
-static int odp_dp_cmd_get(struct sk_buff *skb, struct genl_info *info)
+static int ovs_dp_cmd_get(struct sk_buff *skb, struct genl_info *info)
{
struct sk_buff *reply;
struct datapath *dp;
int err;
- err = odp_dp_cmd_validate(info->attrs);
- if (err)
- return err;
-
- dp = lookup_datapath(info->userhdr, info->attrs);
- if (IS_ERR(dp))
- return PTR_ERR(dp);
+ reply = ovs_dp_cmd_alloc_info(info);
+ if (!reply)
+ return -ENOMEM;
- reply = odp_dp_cmd_build_info(dp, info->snd_pid, info->snd_seq, ODP_DP_CMD_NEW);
- if (IS_ERR(reply))
- return PTR_ERR(reply);
+ rcu_read_lock();
+ dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs);
+ if (IS_ERR(dp)) {
+ err = PTR_ERR(dp);
+ goto err_unlock_free;
+ }
+ err = ovs_dp_cmd_fill_info(dp, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_DP_CMD_NEW);
+ BUG_ON(err < 0);
+ rcu_read_unlock();
return genlmsg_reply(reply, info);
+
+err_unlock_free:
+ rcu_read_unlock();
+ kfree_skb(reply);
+ return err;
}
-static int odp_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
+static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
+ struct ovs_net *ovs_net = net_generic(sock_net(skb->sk), ovs_net_id);
struct datapath *dp;
int skip = cb->args[0];
int i = 0;
- list_for_each_entry (dp, &dps, list_node) {
- if (i < skip)
- continue;
- if (odp_dp_cmd_fill_info(dp, skb, NETLINK_CB(cb->skb).pid,
+ 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,
- ODP_DP_CMD_NEW) < 0)
+ OVS_DP_CMD_NEW) < 0)
break;
i++;
}
+ rcu_read_unlock();
cb->args[0] = i;
return skb->len;
}
+static const struct nla_policy datapath_policy[OVS_DP_ATTR_MAX + 1] = {
+ [OVS_DP_ATTR_NAME] = { .type = NLA_NUL_STRING, .len = IFNAMSIZ - 1 },
+ [OVS_DP_ATTR_UPCALL_PID] = { .type = NLA_U32 },
+ [OVS_DP_ATTR_USER_FEATURES] = { .type = NLA_U32 },
+};
+
static struct genl_ops dp_datapath_genl_ops[] = {
- { .cmd = ODP_DP_CMD_NEW,
+ { .cmd = OVS_DP_CMD_NEW,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = datapath_policy,
- .doit = odp_dp_cmd_new
+ .doit = ovs_dp_cmd_new
},
- { .cmd = ODP_DP_CMD_DEL,
+ { .cmd = OVS_DP_CMD_DEL,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = datapath_policy,
- .doit = odp_dp_cmd_del
+ .doit = ovs_dp_cmd_del
},
- { .cmd = ODP_DP_CMD_GET,
+ { .cmd = OVS_DP_CMD_GET,
.flags = 0, /* OK for unprivileged users. */
.policy = datapath_policy,
- .doit = odp_dp_cmd_get,
- .dumpit = odp_dp_cmd_dump
+ .doit = ovs_dp_cmd_get,
+ .dumpit = ovs_dp_cmd_dump
},
- { .cmd = ODP_DP_CMD_SET,
+ { .cmd = OVS_DP_CMD_SET,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = datapath_policy,
- .doit = odp_dp_cmd_set,
+ .doit = ovs_dp_cmd_set,
},
};
-static const struct nla_policy vport_policy[ODP_VPORT_ATTR_MAX + 1] = {
-#ifdef HAVE_NLA_NUL_STRING
- [ODP_VPORT_ATTR_NAME] = { .type = NLA_NUL_STRING, .len = IFNAMSIZ - 1 },
- [ODP_VPORT_ATTR_PORT_NO] = { .type = NLA_U32 },
- [ODP_VPORT_ATTR_TYPE] = { .type = NLA_U32 },
- [ODP_VPORT_ATTR_STATS] = { .len = sizeof(struct rtnl_link_stats64) },
- [ODP_VPORT_ATTR_ADDRESS] = { .len = ETH_ALEN },
-#else
- [ODP_VPORT_ATTR_STATS] = { .minlen = sizeof(struct rtnl_link_stats64) },
- [ODP_VPORT_ATTR_ADDRESS] = { .minlen = ETH_ALEN },
-#endif
- [ODP_VPORT_ATTR_MTU] = { .type = NLA_U32 },
- [ODP_VPORT_ATTR_OPTIONS] = { .type = NLA_NESTED },
-};
-
-static struct genl_family dp_vport_genl_family = {
+static struct genl_family dp_datapath_genl_family = {
.id = GENL_ID_GENERATE,
- .hdrsize = sizeof(struct odp_header),
- .name = ODP_VPORT_FAMILY,
- .version = 1,
- .maxattr = ODP_VPORT_ATTR_MAX
-};
-
-static struct genl_multicast_group dp_vport_multicast_group = {
- .name = ODP_VPORT_MCGROUP
+ .hdrsize = sizeof(struct ovs_header),
+ .name = OVS_DATAPATH_FAMILY,
+ .version = OVS_DATAPATH_VERSION,
+ .maxattr = OVS_DP_ATTR_MAX,
+ .netnsok = true,
+ .parallel_ops = true,
+ .ops = dp_datapath_genl_ops,
+ .n_ops = ARRAY_SIZE(dp_datapath_genl_ops),
+ .mcgrps = &ovs_dp_datapath_multicast_group,
+ .n_mcgrps = 1,
};
-/* Called with RTNL lock or RCU read lock. */
-static int odp_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb,
- u32 pid, u32 seq, u32 flags, u8 cmd)
+/* 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)
{
- struct odp_header *odp_header;
- struct nlattr *nla;
- int ifindex, iflink;
- int mtu;
+ struct ovs_header *ovs_header;
+ struct ovs_vport_stats vport_stats;
int err;
- odp_header = genlmsg_put(skb, pid, seq, &dp_vport_genl_family,
+ ovs_header = genlmsg_put(skb, portid, seq, &dp_vport_genl_family,
flags, cmd);
- if (!odp_header)
+ if (!ovs_header)
return -EMSGSIZE;
- odp_header->dp_ifindex = vport->dp->dp_ifindex;
+ ovs_header->dp_ifindex = get_dpifindex(vport->dp);
- NLA_PUT_U32(skb, ODP_VPORT_ATTR_PORT_NO, vport->port_no);
- NLA_PUT_U32(skb, ODP_VPORT_ATTR_TYPE, vport_get_type(vport));
- NLA_PUT_STRING(skb, ODP_VPORT_ATTR_NAME, vport_get_name(vport));
-
- nla = nla_reserve(skb, ODP_VPORT_ATTR_STATS, sizeof(struct rtnl_link_stats64));
- if (!nla)
+ if (nla_put_u32(skb, OVS_VPORT_ATTR_PORT_NO, vport->port_no) ||
+ nla_put_u32(skb, OVS_VPORT_ATTR_TYPE, vport->ops->type) ||
+ nla_put_string(skb, OVS_VPORT_ATTR_NAME, vport->ops->get_name(vport)))
goto nla_put_failure;
- if (vport_get_stats(vport, nla_data(nla)))
- __skb_trim(skb, skb->len - nla->nla_len);
- NLA_PUT(skb, ODP_VPORT_ATTR_ADDRESS, ETH_ALEN, vport_get_addr(vport));
+ ovs_vport_get_stats(vport, &vport_stats);
+ if (nla_put(skb, OVS_VPORT_ATTR_STATS, sizeof(struct ovs_vport_stats),
+ &vport_stats))
+ goto nla_put_failure;
- mtu = vport_get_mtu(vport);
- if (mtu)
- NLA_PUT_U32(skb, ODP_VPORT_ATTR_MTU, mtu);
+ if (ovs_vport_get_upcall_portids(vport, skb))
+ goto nla_put_failure;
- err = vport_get_options(vport, skb);
+ err = ovs_vport_get_options(vport, skb);
if (err == -EMSGSIZE)
goto error;
- ifindex = vport_get_ifindex(vport);
- if (ifindex > 0)
- NLA_PUT_U32(skb, ODP_VPORT_ATTR_IFINDEX, ifindex);
-
- iflink = vport_get_iflink(vport);
- if (iflink > 0)
- NLA_PUT_U32(skb, ODP_VPORT_ATTR_IFLINK, iflink);
-
- return genlmsg_end(skb, odp_header);
+ return genlmsg_end(skb, ovs_header);
nla_put_failure:
err = -EMSGSIZE;
error:
- genlmsg_cancel(skb, odp_header);
+ genlmsg_cancel(skb, ovs_header);
return err;
}
-/* Called with RTNL lock or RCU read lock. */
-static struct sk_buff *odp_vport_cmd_build_info(struct vport *vport, u32 pid,
- u32 seq, u8 cmd)
+static struct sk_buff *ovs_vport_cmd_alloc_info(void)
+{
+ return nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL);
+}
+
+/* Called with ovs_mutex, only via ovs_dp_notify_wq(). */
+struct sk_buff *ovs_vport_cmd_build_info(struct vport *vport, u32 portid,
+ u32 seq, u8 cmd)
{
struct sk_buff *skb;
int retval;
if (!skb)
return ERR_PTR(-ENOMEM);
- retval = odp_vport_cmd_fill_info(vport, skb, pid, seq, 0, cmd);
- if (retval < 0) {
- kfree_skb(skb);
- return ERR_PTR(retval);
- }
- return skb;
-}
+ retval = ovs_vport_cmd_fill_info(vport, skb, portid, seq, 0, cmd);
+ BUG_ON(retval < 0);
-static int odp_vport_cmd_validate(struct nlattr *a[ODP_VPORT_ATTR_MAX + 1])
-{
- return CHECK_NUL_STRING(a[ODP_VPORT_ATTR_NAME], IFNAMSIZ - 1);
+ return skb;
}
-/* Called with RTNL lock or RCU read lock. */
-static struct vport *lookup_vport(struct odp_header *odp_header,
- struct nlattr *a[ODP_VPORT_ATTR_MAX + 1])
+/* 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])
{
struct datapath *dp;
struct vport *vport;
- if (a[ODP_VPORT_ATTR_NAME]) {
- vport = vport_locate(nla_data(a[ODP_VPORT_ATTR_NAME]));
+ if (a[OVS_VPORT_ATTR_NAME]) {
+ vport = ovs_vport_locate(net, nla_data(a[OVS_VPORT_ATTR_NAME]));
if (!vport)
return ERR_PTR(-ENODEV);
+ if (ovs_header->dp_ifindex &&
+ ovs_header->dp_ifindex != get_dpifindex(vport->dp))
+ return ERR_PTR(-ENODEV);
return vport;
- } else if (a[ODP_VPORT_ATTR_PORT_NO]) {
- u32 port_no = nla_get_u32(a[ODP_VPORT_ATTR_PORT_NO]);
+ } else if (a[OVS_VPORT_ATTR_PORT_NO]) {
+ u32 port_no = nla_get_u32(a[OVS_VPORT_ATTR_PORT_NO]);
if (port_no >= DP_MAX_PORTS)
return ERR_PTR(-EFBIG);
- dp = get_dp(odp_header->dp_ifindex);
+ dp = get_dp(net, ovs_header->dp_ifindex);
if (!dp)
return ERR_PTR(-ENODEV);
- vport = get_vport_protected(dp, port_no);
+ vport = ovs_vport_ovsl_rcu(dp, port_no);
if (!vport)
- return ERR_PTR(-ENOENT);
+ return ERR_PTR(-ENODEV);
return vport;
} else
return ERR_PTR(-EINVAL);
}
-/* Called with RTNL lock. */
-static int change_vport(struct vport *vport, struct nlattr *a[ODP_VPORT_ATTR_MAX + 1])
-{
- int err = 0;
- if (a[ODP_VPORT_ATTR_STATS])
- err = vport_set_stats(vport, nla_data(a[ODP_VPORT_ATTR_STATS]));
- if (!err && a[ODP_VPORT_ATTR_ADDRESS])
- err = vport_set_addr(vport, nla_data(a[ODP_VPORT_ATTR_ADDRESS]));
- if (!err && a[ODP_VPORT_ATTR_MTU])
- err = vport_set_mtu(vport, nla_get_u32(a[ODP_VPORT_ATTR_MTU]));
- return err;
-}
-
-static int odp_vport_cmd_new(struct sk_buff *skb, struct genl_info *info)
+static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
- struct odp_header *odp_header = info->userhdr;
+ struct ovs_header *ovs_header = info->userhdr;
struct vport_parms parms;
struct sk_buff *reply;
struct vport *vport;
u32 port_no;
int err;
- err = -EINVAL;
- if (!a[ODP_VPORT_ATTR_NAME] || !a[ODP_VPORT_ATTR_TYPE])
- goto exit;
+ if (!a[OVS_VPORT_ATTR_NAME] || !a[OVS_VPORT_ATTR_TYPE] ||
+ !a[OVS_VPORT_ATTR_UPCALL_PID])
+ return -EINVAL;
- err = odp_vport_cmd_validate(a);
- if (err)
- goto exit;
+ port_no = a[OVS_VPORT_ATTR_PORT_NO]
+ ? nla_get_u32(a[OVS_VPORT_ATTR_PORT_NO]) : 0;
+ if (port_no >= DP_MAX_PORTS)
+ return -EFBIG;
+
+ reply = ovs_vport_cmd_alloc_info();
+ if (!reply)
+ return -ENOMEM;
- rtnl_lock();
- dp = get_dp(odp_header->dp_ifindex);
+ ovs_lock();
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
err = -ENODEV;
if (!dp)
- goto exit_unlock;
+ goto exit_unlock_free;
- if (a[ODP_VPORT_ATTR_PORT_NO]) {
- port_no = nla_get_u32(a[ODP_VPORT_ATTR_PORT_NO]);
-
- err = -EFBIG;
- if (port_no >= DP_MAX_PORTS)
- goto exit_unlock;
-
- vport = get_vport_protected(dp, port_no);
+ if (port_no) {
+ vport = ovs_vport_ovsl(dp, port_no);
err = -EBUSY;
if (vport)
- goto exit_unlock;
+ goto exit_unlock_free;
} else {
for (port_no = 1; ; port_no++) {
if (port_no >= DP_MAX_PORTS) {
err = -EFBIG;
- goto exit_unlock;
+ goto exit_unlock_free;
}
- vport = get_vport_protected(dp, port_no);
+ vport = ovs_vport_ovsl(dp, port_no);
if (!vport)
break;
}
}
- parms.name = nla_data(a[ODP_VPORT_ATTR_NAME]);
- parms.type = nla_get_u32(a[ODP_VPORT_ATTR_TYPE]);
- parms.options = a[ODP_VPORT_ATTR_OPTIONS];
+ parms.name = nla_data(a[OVS_VPORT_ATTR_NAME]);
+ parms.type = nla_get_u32(a[OVS_VPORT_ATTR_TYPE]);
+ parms.options = a[OVS_VPORT_ATTR_OPTIONS];
parms.dp = dp;
parms.port_no = port_no;
+ parms.upcall_portids = a[OVS_VPORT_ATTR_UPCALL_PID];
vport = new_vport(&parms);
err = PTR_ERR(vport);
if (IS_ERR(vport))
- goto exit_unlock;
+ goto exit_unlock_free;
- set_internal_devs_mtu(dp);
- dp_sysfs_add_if(vport);
+ err = 0;
+ if (a[OVS_VPORT_ATTR_STATS])
+ ovs_vport_set_stats(vport, nla_data(a[OVS_VPORT_ATTR_STATS]));
- err = change_vport(vport, a);
- if (!err) {
- reply = odp_vport_cmd_build_info(vport, info->snd_pid,
- info->snd_seq, ODP_VPORT_CMD_NEW);
- if (IS_ERR(reply))
- err = PTR_ERR(reply);
- }
- if (err) {
- dp_detach_port(vport);
- goto exit_unlock;
- }
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_VPORT_CMD_NEW);
+ BUG_ON(err < 0);
+ ovs_unlock();
+ ovs_notify(&dp_vport_genl_family, &ovs_dp_vport_multicast_group, reply, info);
+ return 0;
-exit_unlock:
- rtnl_unlock();
-exit:
+exit_unlock_free:
+ ovs_unlock();
+ kfree_skb(reply);
return err;
}
-static int odp_vport_cmd_set(struct sk_buff *skb, struct genl_info *info)
+static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
struct sk_buff *reply;
struct vport *vport;
int err;
- err = odp_vport_cmd_validate(a);
- if (err)
- goto exit;
+ reply = ovs_vport_cmd_alloc_info();
+ if (!reply)
+ return -ENOMEM;
- rtnl_lock();
- vport = lookup_vport(info->userhdr, a);
+ ovs_lock();
+ vport = lookup_vport(sock_net(skb->sk), info->userhdr, a);
err = PTR_ERR(vport);
if (IS_ERR(vport))
- goto exit_unlock;
+ goto exit_unlock_free;
- err = 0;
- if (a[ODP_VPORT_ATTR_OPTIONS])
- err = vport_set_options(vport, a[ODP_VPORT_ATTR_OPTIONS]);
- if (!err)
- err = change_vport(vport, a);
+ if (a[OVS_VPORT_ATTR_TYPE] &&
+ nla_get_u32(a[OVS_VPORT_ATTR_TYPE]) != vport->ops->type) {
+ err = -EINVAL;
+ goto exit_unlock_free;
+ }
- reply = odp_vport_cmd_build_info(vport, info->snd_pid, info->snd_seq,
- ODP_VPORT_CMD_NEW);
- if (IS_ERR(reply)) {
- err = PTR_ERR(reply);
- netlink_set_err(INIT_NET_GENL_SOCK, 0,
- dp_vport_multicast_group.id, err);
- return 0;
+ if (a[OVS_VPORT_ATTR_OPTIONS]) {
+ err = ovs_vport_set_options(vport, a[OVS_VPORT_ATTR_OPTIONS]);
+ if (err)
+ goto exit_unlock_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]) {
+ err = ovs_vport_set_upcall_portids(vport,
+ a[OVS_VPORT_ATTR_UPCALL_PID]);
+ if (err)
+ goto exit_unlock_free;
}
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_VPORT_CMD_NEW);
+ BUG_ON(err < 0);
+ ovs_unlock();
+
+ ovs_notify(&dp_vport_genl_family, &ovs_dp_vport_multicast_group, reply, info);
+ return 0;
-exit_unlock:
- rtnl_unlock();
-exit:
+exit_unlock_free:
+ ovs_unlock();
+ kfree_skb(reply);
return err;
}
-static int odp_vport_cmd_del(struct sk_buff *skb, struct genl_info *info)
+static int ovs_vport_cmd_del(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
struct sk_buff *reply;
struct vport *vport;
int err;
- err = odp_vport_cmd_validate(a);
- if (err)
- goto exit;
+ reply = ovs_vport_cmd_alloc_info();
+ if (!reply)
+ return -ENOMEM;
- rtnl_lock();
- vport = lookup_vport(info->userhdr, a);
+ ovs_lock();
+ vport = lookup_vport(sock_net(skb->sk), info->userhdr, a);
err = PTR_ERR(vport);
if (IS_ERR(vport))
- goto exit_unlock;
+ goto exit_unlock_free;
- if (vport->port_no == ODPP_LOCAL) {
+ if (vport->port_no == OVSP_LOCAL) {
err = -EINVAL;
- goto exit_unlock;
+ goto exit_unlock_free;
}
- reply = odp_vport_cmd_build_info(vport, info->snd_pid, info->snd_seq,
- ODP_VPORT_CMD_DEL);
- err = PTR_ERR(reply);
- if (IS_ERR(reply))
- goto exit_unlock;
-
- err = dp_detach_port(vport);
+ err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_VPORT_CMD_DEL);
+ BUG_ON(err < 0);
+ ovs_dp_detach_port(vport);
+ ovs_unlock();
- genl_notify(reply, genl_info_net(info), info->snd_pid,
- dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
+ ovs_notify(&dp_vport_genl_family, &ovs_dp_vport_multicast_group, reply, info);
+ return 0;
-exit_unlock:
- rtnl_unlock();
-exit:
+exit_unlock_free:
+ ovs_unlock();
+ kfree_skb(reply);
return err;
}
-static int odp_vport_cmd_get(struct sk_buff *skb, struct genl_info *info)
+static int ovs_vport_cmd_get(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr **a = info->attrs;
- struct odp_header *odp_header = info->userhdr;
+ struct ovs_header *ovs_header = info->userhdr;
struct sk_buff *reply;
struct vport *vport;
int err;
- err = odp_vport_cmd_validate(a);
- if (err)
- goto exit;
+ reply = ovs_vport_cmd_alloc_info();
+ if (!reply)
+ return -ENOMEM;
rcu_read_lock();
- vport = lookup_vport(odp_header, a);
+ vport = lookup_vport(sock_net(skb->sk), ovs_header, a);
err = PTR_ERR(vport);
if (IS_ERR(vport))
- goto exit_unlock;
-
- reply = odp_vport_cmd_build_info(vport, info->snd_pid, info->snd_seq,
- ODP_VPORT_CMD_NEW);
- err = PTR_ERR(reply);
- if (IS_ERR(reply))
- goto exit_unlock;
+ goto exit_unlock_free;
+ err = ovs_vport_cmd_fill_info(vport, reply, info->snd_portid,
+ info->snd_seq, 0, OVS_VPORT_CMD_NEW);
+ BUG_ON(err < 0);
+ rcu_read_unlock();
- err = genlmsg_reply(reply, info);
+ return genlmsg_reply(reply, info);
-exit_unlock:
+exit_unlock_free:
rcu_read_unlock();
-exit:
+ kfree_skb(reply);
return err;
}
-static int odp_vport_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
+static int ovs_vport_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
- struct odp_header *odp_header = genlmsg_data(nlmsg_data(cb->nlh));
+ struct ovs_header *ovs_header = genlmsg_data(nlmsg_data(cb->nlh));
struct datapath *dp;
- u32 port_no;
- int retval;
-
- dp = get_dp(odp_header->dp_ifindex);
- if (!dp)
- return -ENODEV;
+ int bucket = cb->args[0], skip = cb->args[1];
+ int i, j = 0;
rcu_read_lock();
- for (port_no = cb->args[0]; port_no < DP_MAX_PORTS; port_no++) {
+ dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex);
+ if (!dp) {
+ rcu_read_unlock();
+ return -ENODEV;
+ }
+ for (i = bucket; i < DP_VPORT_HASH_BUCKETS; i++) {
struct vport *vport;
- vport = get_vport_protected(dp, port_no);
- if (!vport)
- continue;
-
- if (odp_vport_cmd_fill_info(vport, skb, NETLINK_CB(cb->skb).pid,
- cb->nlh->nlmsg_seq, NLM_F_MULTI,
- ODP_VPORT_CMD_NEW) < 0)
- break;
+ j = 0;
+ hlist_for_each_entry_rcu(vport, &dp->ports[i], dp_hash_node) {
+ if (j >= skip &&
+ ovs_vport_cmd_fill_info(vport, skb,
+ NETLINK_CB(cb->skb).portid,
+ cb->nlh->nlmsg_seq,
+ NLM_F_MULTI,
+ OVS_VPORT_CMD_NEW) < 0)
+ goto out;
+
+ j++;
+ }
+ skip = 0;
}
+out:
rcu_read_unlock();
- cb->args[0] = port_no;
- retval = skb->len;
+ cb->args[0] = i;
+ cb->args[1] = j;
- return retval;
+ return skb->len;
}
+static const struct nla_policy vport_policy[OVS_VPORT_ATTR_MAX + 1] = {
+ [OVS_VPORT_ATTR_NAME] = { .type = NLA_NUL_STRING, .len = IFNAMSIZ - 1 },
+ [OVS_VPORT_ATTR_STATS] = { .len = sizeof(struct ovs_vport_stats) },
+ [OVS_VPORT_ATTR_PORT_NO] = { .type = NLA_U32 },
+ [OVS_VPORT_ATTR_TYPE] = { .type = NLA_U32 },
+ [OVS_VPORT_ATTR_UPCALL_PID] = { .type = NLA_U32 },
+ [OVS_VPORT_ATTR_OPTIONS] = { .type = NLA_NESTED },
+};
+
static struct genl_ops dp_vport_genl_ops[] = {
- { .cmd = ODP_VPORT_CMD_NEW,
+ { .cmd = OVS_VPORT_CMD_NEW,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = vport_policy,
- .doit = odp_vport_cmd_new
+ .doit = ovs_vport_cmd_new
},
- { .cmd = ODP_VPORT_CMD_DEL,
+ { .cmd = OVS_VPORT_CMD_DEL,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = vport_policy,
- .doit = odp_vport_cmd_del
+ .doit = ovs_vport_cmd_del
},
- { .cmd = ODP_VPORT_CMD_GET,
+ { .cmd = OVS_VPORT_CMD_GET,
.flags = 0, /* OK for unprivileged users. */
.policy = vport_policy,
- .doit = odp_vport_cmd_get,
- .dumpit = odp_vport_cmd_dump
+ .doit = ovs_vport_cmd_get,
+ .dumpit = ovs_vport_cmd_dump
},
- { .cmd = ODP_VPORT_CMD_SET,
+ { .cmd = OVS_VPORT_CMD_SET,
.flags = GENL_ADMIN_PERM, /* Requires CAP_NET_ADMIN privilege. */
.policy = vport_policy,
- .doit = odp_vport_cmd_set,
+ .doit = ovs_vport_cmd_set,
},
};
-struct genl_family_and_ops {
- struct genl_family *family;
- struct genl_ops *ops;
- int n_ops;
- struct genl_multicast_group *group;
+struct genl_family dp_vport_genl_family = {
+ .id = GENL_ID_GENERATE,
+ .hdrsize = sizeof(struct ovs_header),
+ .name = OVS_VPORT_FAMILY,
+ .version = OVS_VPORT_VERSION,
+ .maxattr = OVS_VPORT_ATTR_MAX,
+ .netnsok = true,
+ .parallel_ops = true,
+ .ops = dp_vport_genl_ops,
+ .n_ops = ARRAY_SIZE(dp_vport_genl_ops),
+ .mcgrps = &ovs_dp_vport_multicast_group,
+ .n_mcgrps = 1,
};
-static const struct genl_family_and_ops dp_genl_families[] = {
- { &dp_datapath_genl_family,
- dp_datapath_genl_ops, ARRAY_SIZE(dp_datapath_genl_ops),
- &dp_datapath_multicast_group },
- { &dp_vport_genl_family,
- dp_vport_genl_ops, ARRAY_SIZE(dp_vport_genl_ops),
- &dp_vport_multicast_group },
- { &dp_flow_genl_family,
- dp_flow_genl_ops, ARRAY_SIZE(dp_flow_genl_ops),
- &dp_flow_multicast_group },
- { &dp_packet_genl_family,
- dp_packet_genl_ops, ARRAY_SIZE(dp_packet_genl_ops),
- NULL },
+static struct genl_family *dp_genl_families[] = {
+ &dp_datapath_genl_family,
+ &dp_vport_genl_family,
+ &dp_flow_genl_family,
+ &dp_packet_genl_family,
};
static void dp_unregister_genl(int n_families)
{
int i;
- for (i = 0; i < n_families; i++) {
- genl_unregister_family(dp_genl_families[i].family);
- }
+ for (i = 0; i < n_families; i++)
+ genl_unregister_family(dp_genl_families[i]);
}
static int dp_register_genl(void)
{
- int n_registered;
int err;
int i;
- n_registered = 0;
for (i = 0; i < ARRAY_SIZE(dp_genl_families); i++) {
- const struct genl_family_and_ops *f = &dp_genl_families[i];
- err = genl_register_family_with_ops(f->family, f->ops,
- f->n_ops);
+ err = genl_register_family(dp_genl_families[i]);
if (err)
goto error;
- n_registered++;
-
- if (f->group) {
- err = genl_register_mc_group(f->family, f->group);
- if (err)
- goto error;
- }
}
- err = packet_register_mc_groups();
- if (err)
- goto error;
return 0;
error:
- dp_unregister_genl(n_registered);
+ dp_unregister_genl(i);
return err;
}
+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);
+
+ 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 = {
+ .init = ovs_init_net,
+ .exit = ovs_exit_net,
+ .id = &ovs_net_id,
+ .size = sizeof(struct ovs_net),
+};
+
+DEFINE_COMPAT_PNET_REG_FUNC(device);
+
static int __init dp_init(void)
{
- struct sk_buff *dummy_skb;
int err;
- BUILD_BUG_ON(sizeof(struct ovs_skb_cb) > sizeof(dummy_skb->cb));
+ BUILD_BUG_ON(sizeof(struct ovs_skb_cb) > FIELD_SIZEOF(struct sk_buff, cb));
- printk("Open vSwitch %s, built "__DATE__" "__TIME__"\n", VERSION BUILDNR);
+ pr_info("Open vSwitch switching datapath %s, built "__DATE__" "__TIME__"\n",
+ VERSION);
- err = flow_init();
+ err = ovs_flow_init();
if (err)
goto error;
- err = vport_init();
+ err = ovs_vport_init();
if (err)
goto error_flow_exit;
- err = register_netdevice_notifier(&dp_device_notifier);
+ err = register_pernet_device(&ovs_net_ops);
if (err)
goto error_vport_exit;
+ err = register_netdevice_notifier(&ovs_dp_device_notifier);
+ if (err)
+ goto error_netns_exit;
+
err = dp_register_genl();
if (err < 0)
goto error_unreg_notifier;
return 0;
error_unreg_notifier:
- unregister_netdevice_notifier(&dp_device_notifier);
+ unregister_netdevice_notifier(&ovs_dp_device_notifier);
+error_netns_exit:
+ unregister_pernet_device(&ovs_net_ops);
error_vport_exit:
- vport_exit();
+ ovs_vport_exit();
error_flow_exit:
- flow_exit();
+ ovs_flow_exit();
error:
return err;
}
static void dp_cleanup(void)
{
- rcu_barrier();
dp_unregister_genl(ARRAY_SIZE(dp_genl_families));
- unregister_netdevice_notifier(&dp_device_notifier);
- vport_exit();
- flow_exit();
+ unregister_netdevice_notifier(&ovs_dp_device_notifier);
+ unregister_pernet_device(&ovs_net_ops);
+ rcu_barrier();
+ ovs_vport_exit();
+ ovs_flow_exit();
}
module_init(dp_init);
MODULE_DESCRIPTION("Open vSwitch switching datapath");
MODULE_LICENSE("GPL");
+MODULE_VERSION(VERSION);
git://git.onelab.eu / sliver-openvswitch.git / blobdiff