diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2013-05-01 14:08:52 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2013-05-01 14:08:52 -0700 |
commit | 73287a43cc79ca06629a88d1a199cd283f42456a (patch) | |
tree | acf4456e260115bea77ee31a29f10ce17f0db45c /net/openvswitch | |
parent | 251df49db3327c64bf917bfdba94491fde2b4ee0 (diff) | |
parent | 20074f357da4a637430aec2879c9d864c5d2c23c (diff) | |
download | linux-73287a43cc79ca06629a88d1a199cd283f42456a.tar.bz2 |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next
Pull networking updates from David Miller:
"Highlights (1721 non-merge commits, this has to be a record of some
sort):
1) Add 'random' mode to team driver, from Jiri Pirko and Eric
Dumazet.
2) Make it so that any driver that supports configuration of multiple
MAC addresses can provide the forwarding database add and del
calls by providing a default implementation and hooking that up if
the driver doesn't have an explicit set of handlers. From Vlad
Yasevich.
3) Support GSO segmentation over tunnels and other encapsulating
devices such as VXLAN, from Pravin B Shelar.
4) Support L2 GRE tunnels in the flow dissector, from Michael Dalton.
5) Implement Tail Loss Probe (TLP) detection in TCP, from Nandita
Dukkipati.
6) In the PHY layer, allow supporting wake-on-lan in situations where
the PHY registers have to be written for it to be configured.
Use it to support wake-on-lan in mv643xx_eth.
From Michael Stapelberg.
7) Significantly improve firewire IPV6 support, from YOSHIFUJI
Hideaki.
8) Allow multiple packets to be sent in a single transmission using
network coding in batman-adv, from Martin Hundebøll.
9) Add support for T5 cxgb4 chips, from Santosh Rastapur.
10) Generalize the VXLAN forwarding tables so that there is more
flexibility in configurating various aspects of the endpoints.
From David Stevens.
11) Support RSS and TSO in hardware over GRE tunnels in bxn2x driver,
from Dmitry Kravkov.
12) Zero copy support in nfnelink_queue, from Eric Dumazet and Pablo
Neira Ayuso.
13) Start adding networking selftests.
14) In situations of overload on the same AF_PACKET fanout socket, or
per-cpu packet receive queue, minimize drop by distributing the
load to other cpus/fanouts. From Willem de Bruijn and Eric
Dumazet.
15) Add support for new payload offset BPF instruction, from Daniel
Borkmann.
16) Convert several drivers over to mdoule_platform_driver(), from
Sachin Kamat.
17) Provide a minimal BPF JIT image disassembler userspace tool, from
Daniel Borkmann.
18) Rewrite F-RTO implementation in TCP to match the final
specification of it in RFC4138 and RFC5682. From Yuchung Cheng.
19) Provide netlink socket diag of netlink sockets ("Yo dawg, I hear
you like netlink, so I implemented netlink dumping of netlink
sockets.") From Andrey Vagin.
20) Remove ugly passing of rtnetlink attributes into rtnl_doit
functions, from Thomas Graf.
21) Allow userspace to be able to see if a configuration change occurs
in the middle of an address or device list dump, from Nicolas
Dichtel.
22) Support RFC3168 ECN protection for ipv6 fragments, from Hannes
Frederic Sowa.
23) Increase accuracy of packet length used by packet scheduler, from
Jason Wang.
24) Beginning set of changes to make ipv4/ipv6 fragment handling more
scalable and less susceptible to overload and locking contention,
from Jesper Dangaard Brouer.
25) Get rid of using non-type-safe NLMSG_* macros and use nlmsg_*()
instead. From Hong Zhiguo.
26) Optimize route usage in IPVS by avoiding reference counting where
possible, from Julian Anastasov.
27) Convert IPVS schedulers to RCU, also from Julian Anastasov.
28) Support cpu fanouts in xt_NFQUEUE netfilter target, from Holger
Eitzenberger.
29) Network namespace support for nf_log, ebt_log, xt_LOG, ipt_ULOG,
nfnetlink_log, and nfnetlink_queue. From Gao feng.
30) Implement RFC3168 ECN protection, from Hannes Frederic Sowa.
31) Support several new r8169 chips, from Hayes Wang.
32) Support tokenized interface identifiers in ipv6, from Daniel
Borkmann.
33) Use usbnet_link_change() helper in USB net driver, from Ming Lei.
34) Add 802.1ad vlan offload support, from Patrick McHardy.
35) Support mmap() based netlink communication, also from Patrick
McHardy.
36) Support HW timestamping in mlx4 driver, from Amir Vadai.
37) Rationalize AF_PACKET packet timestamping when transmitting, from
Willem de Bruijn and Daniel Borkmann.
38) Bring parity to what's provided by /proc/net/packet socket dumping
and the info provided by netlink socket dumping of AF_PACKET
sockets. From Nicolas Dichtel.
39) Fix peeking beyond zero sized SKBs in AF_UNIX, from Benjamin
Poirier"
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1722 commits)
filter: fix va_list build error
af_unix: fix a fatal race with bit fields
bnx2x: Prevent memory leak when cnic is absent
bnx2x: correct reading of speed capabilities
net: sctp: attribute printl with __printf for gcc fmt checks
netlink: kconfig: move mmap i/o into netlink kconfig
netpoll: convert mutex into a semaphore
netlink: Fix skb ref counting.
net_sched: act_ipt forward compat with xtables
mlx4_en: fix a build error on 32bit arches
Revert "bnx2x: allow nvram test to run when device is down"
bridge: avoid OOPS if root port not found
drivers: net: cpsw: fix kernel warn on cpsw irq enable
sh_eth: use random MAC address if no valid one supplied
3c509.c: call SET_NETDEV_DEV for all device types (ISA/ISAPnP/EISA)
tg3: fix to append hardware time stamping flags
unix/stream: fix peeking with an offset larger than data in queue
unix/dgram: fix peeking with an offset larger than data in queue
unix/dgram: peek beyond 0-sized skbs
openvswitch: Remove unneeded ovs_netdev_get_ifindex()
...
Diffstat (limited to 'net/openvswitch')
-rw-r--r-- | net/openvswitch/actions.c | 6 | ||||
-rw-r--r-- | net/openvswitch/datapath.c | 421 | ||||
-rw-r--r-- | net/openvswitch/datapath.h | 72 | ||||
-rw-r--r-- | net/openvswitch/dp_notify.c | 82 | ||||
-rw-r--r-- | net/openvswitch/flow.c | 8 | ||||
-rw-r--r-- | net/openvswitch/flow.h | 21 | ||||
-rw-r--r-- | net/openvswitch/vport-internal_dev.c | 22 | ||||
-rw-r--r-- | net/openvswitch/vport-netdev.c | 15 | ||||
-rw-r--r-- | net/openvswitch/vport-netdev.h | 1 | ||||
-rw-r--r-- | net/openvswitch/vport.c | 58 | ||||
-rw-r--r-- | net/openvswitch/vport.h | 21 |
11 files changed, 420 insertions, 307 deletions
diff --git a/net/openvswitch/actions.c b/net/openvswitch/actions.c index d4d5363c7ba7..894b6cbdd929 100644 --- a/net/openvswitch/actions.c +++ b/net/openvswitch/actions.c @@ -98,7 +98,7 @@ static int pop_vlan(struct sk_buff *skb) if (unlikely(err)) return err; - __vlan_hwaccel_put_tag(skb, ntohs(tci)); + __vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ntohs(tci)); return 0; } @@ -110,7 +110,7 @@ static int push_vlan(struct sk_buff *skb, const struct ovs_action_push_vlan *vla /* push down current VLAN tag */ current_tag = vlan_tx_tag_get(skb); - if (!__vlan_put_tag(skb, current_tag)) + if (!__vlan_put_tag(skb, skb->vlan_proto, current_tag)) return -ENOMEM; if (skb->ip_summed == CHECKSUM_COMPLETE) @@ -118,7 +118,7 @@ static int push_vlan(struct sk_buff *skb, const struct ovs_action_push_vlan *vla + (2 * ETH_ALEN), VLAN_HLEN, 0)); } - __vlan_hwaccel_put_tag(skb, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT); + __vlan_hwaccel_put_tag(skb, vlan->vlan_tpid, ntohs(vlan->vlan_tci) & ~VLAN_TAG_PRESENT); return 0; } diff --git a/net/openvswitch/datapath.c b/net/openvswitch/datapath.c index 6980c3e6f066..d12d6b8b5e8b 100644 --- a/net/openvswitch/datapath.c +++ b/net/openvswitch/datapath.c @@ -44,6 +44,7 @@ #include <linux/netfilter_ipv4.h> #include <linux/inetdevice.h> #include <linux/list.h> +#include <linux/lockdep.h> #include <linux/openvswitch.h> #include <linux/rculist.h> #include <linux/dmi.h> @@ -55,39 +56,61 @@ #include "datapath.h" #include "flow.h" #include "vport-internal_dev.h" +#include "vport-netdev.h" -/** - * struct ovs_net - Per net-namespace data for ovs. - * @dps: List of datapaths to enable dumping them all out. - * Protected by genl_mutex. - */ -struct ovs_net { - struct list_head dps; -}; - -static int ovs_net_id __read_mostly; #define REHASH_FLOW_INTERVAL (10 * 60 * HZ) static void rehash_flow_table(struct work_struct *work); static DECLARE_DELAYED_WORK(rehash_flow_wq, rehash_flow_table); +int ovs_net_id __read_mostly; + +static void ovs_notify(struct sk_buff *skb, struct genl_info *info, + struct genl_multicast_group *grp) +{ + genl_notify(skb, genl_info_net(info), info->snd_portid, + grp->id, info->nlhdr, GFP_KERNEL); +} + /** * DOC: Locking: * - * Writes to device state (add/remove datapath, port, set operations on vports, - * etc.) are protected by RTNL. - * - * Writes to other state (flow table modifications, set miscellaneous datapath - * parameters, etc.) are protected by genl_mutex. The RTNL lock nests inside - * genl_mutex. + * All writes e.g. Writes to device state (add/remove datapath, port, set + * operations on vports, etc.), Writes to other state (flow table + * modifications, set miscellaneous datapath parameters, etc.) are protected + * by ovs_lock. * * Reads are protected by RCU. * * There are a few special cases (mostly stats) that have their own * synchronization but they nest under all of above and don't interact with * each other. + * + * The RTNL lock nests inside ovs_mutex. */ +static DEFINE_MUTEX(ovs_mutex); + +void ovs_lock(void) +{ + mutex_lock(&ovs_mutex); +} + +void ovs_unlock(void) +{ + mutex_unlock(&ovs_mutex); +} + +#ifdef CONFIG_LOCKDEP +int lockdep_ovsl_is_held(void) +{ + if (debug_locks) + return lockdep_is_held(&ovs_mutex); + else + return 1; +} +#endif + static struct vport *new_vport(const struct vport_parms *); static int queue_gso_packets(struct net *, int dp_ifindex, struct sk_buff *, const struct dp_upcall_info *); @@ -95,7 +118,7 @@ static int queue_userspace_packet(struct net *, int dp_ifindex, struct sk_buff *, const struct dp_upcall_info *); -/* Must be called with rcu_read_lock, genl_mutex, or RTNL lock. */ +/* Must be called with rcu_read_lock or ovs_mutex. */ static struct datapath *get_dp(struct net *net, int dp_ifindex) { struct datapath *dp = NULL; @@ -113,10 +136,10 @@ static struct datapath *get_dp(struct net *net, int dp_ifindex) return dp; } -/* Must be called with rcu_read_lock or RTNL lock. */ +/* Must be called with rcu_read_lock or ovs_mutex. */ const char *ovs_dp_name(const struct datapath *dp) { - struct vport *vport = ovs_vport_rtnl_rcu(dp, OVSP_LOCAL); + struct vport *vport = ovs_vport_ovsl_rcu(dp, OVSP_LOCAL); return vport->ops->get_name(vport); } @@ -129,7 +152,7 @@ static int get_dpifindex(struct datapath *dp) local = ovs_vport_rcu(dp, OVSP_LOCAL); if (local) - ifindex = local->ops->get_ifindex(local); + ifindex = netdev_vport_priv(local)->dev->ifindex; else ifindex = 0; @@ -168,7 +191,7 @@ struct vport *ovs_lookup_vport(const struct datapath *dp, u16 port_no) 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; @@ -180,14 +203,12 @@ static struct vport *new_vport(const struct vport_parms *parms) hlist_add_head_rcu(&vport->dp_hash_node, head); } - return vport; } -/* Called with RTNL lock. */ void ovs_dp_detach_port(struct vport *p) { - ASSERT_RTNL(); + ASSERT_OVSL(); /* First drop references to device. */ hlist_del_rcu(&p->dp_hash_node); @@ -250,7 +271,8 @@ static struct genl_family dp_packet_genl_family = { .name = OVS_PACKET_FAMILY, .version = OVS_PACKET_VERSION, .maxattr = OVS_PACKET_ATTR_MAX, - .netnsok = true + .netnsok = true, + .parallel_ops = true, }; int ovs_dp_upcall(struct datapath *dp, struct sk_buff *skb, @@ -337,6 +359,35 @@ static int queue_gso_packets(struct net *net, int dp_ifindex, return err; } +static size_t key_attr_size(void) +{ + return nla_total_size(4) /* OVS_KEY_ATTR_PRIORITY */ + + nla_total_size(4) /* OVS_KEY_ATTR_IN_PORT */ + + nla_total_size(4) /* OVS_KEY_ATTR_SKB_MARK */ + + nla_total_size(12) /* OVS_KEY_ATTR_ETHERNET */ + + nla_total_size(2) /* OVS_KEY_ATTR_ETHERTYPE */ + + nla_total_size(4) /* OVS_KEY_ATTR_8021Q */ + + nla_total_size(0) /* OVS_KEY_ATTR_ENCAP */ + + nla_total_size(2) /* OVS_KEY_ATTR_ETHERTYPE */ + + nla_total_size(40) /* OVS_KEY_ATTR_IPV6 */ + + nla_total_size(2) /* OVS_KEY_ATTR_ICMPV6 */ + + nla_total_size(28); /* OVS_KEY_ATTR_ND */ +} + +static size_t upcall_msg_size(const struct sk_buff *skb, + const struct nlattr *userdata) +{ + size_t size = NLMSG_ALIGN(sizeof(struct ovs_header)) + + nla_total_size(skb->len) /* OVS_PACKET_ATTR_PACKET */ + + nla_total_size(key_attr_size()); /* OVS_PACKET_ATTR_KEY */ + + /* OVS_PACKET_ATTR_USERDATA */ + if (userdata) + size += NLA_ALIGN(userdata->nla_len); + + return size; +} + static int queue_userspace_packet(struct net *net, int dp_ifindex, struct sk_buff *skb, const struct dp_upcall_info *upcall_info) @@ -345,7 +396,6 @@ static int queue_userspace_packet(struct net *net, int dp_ifindex, struct sk_buff *nskb = NULL; struct sk_buff *user_skb; /* to be queued to userspace */ struct nlattr *nla; - unsigned int len; int err; if (vlan_tx_tag_present(skb)) { @@ -353,7 +403,7 @@ static int queue_userspace_packet(struct net *net, int dp_ifindex, if (!nskb) return -ENOMEM; - nskb = __vlan_put_tag(nskb, vlan_tx_tag_get(nskb)); + nskb = __vlan_put_tag(nskb, nskb->vlan_proto, vlan_tx_tag_get(nskb)); if (!nskb) return -ENOMEM; @@ -366,13 +416,7 @@ static int queue_userspace_packet(struct net *net, int dp_ifindex, goto out; } - len = sizeof(struct ovs_header); - len += nla_total_size(skb->len); - len += nla_total_size(FLOW_BUFSIZE); - if (upcall_info->cmd == OVS_PACKET_CMD_ACTION) - len += nla_total_size(8); - - user_skb = genlmsg_new(len, GFP_ATOMIC); + user_skb = genlmsg_new(upcall_msg_size(skb, upcall_info->userdata), GFP_ATOMIC); if (!user_skb) { err = -ENOMEM; goto out; @@ -387,8 +431,9 @@ static int queue_userspace_packet(struct net *net, int dp_ifindex, nla_nest_end(user_skb, nla); if (upcall_info->userdata) - nla_put_u64(user_skb, OVS_PACKET_ATTR_USERDATA, - nla_get_u64(upcall_info->userdata)); + __nla_put(user_skb, OVS_PACKET_ATTR_USERDATA, + nla_len(upcall_info->userdata), + nla_data(upcall_info->userdata)); nla = __nla_reserve(user_skb, OVS_PACKET_ATTR_PACKET, skb->len); @@ -402,13 +447,13 @@ out: return err; } -/* Called with genl_mutex. */ +/* Called with ovs_mutex. */ static int flush_flows(struct datapath *dp) { struct flow_table *old_table; struct flow_table *new_table; - old_table = genl_dereference(dp->table); + old_table = ovsl_dereference(dp->table); new_table = ovs_flow_tbl_alloc(TBL_MIN_BUCKETS); if (!new_table) return -ENOMEM; @@ -544,7 +589,7 @@ static int validate_userspace(const struct nlattr *attr) { static const struct nla_policy userspace_policy[OVS_USERSPACE_ATTR_MAX + 1] = { [OVS_USERSPACE_ATTR_PID] = {.type = NLA_U32 }, - [OVS_USERSPACE_ATTR_USERDATA] = {.type = NLA_U64 }, + [OVS_USERSPACE_ATTR_USERDATA] = {.type = NLA_UNSPEC }, }; struct nlattr *a[OVS_USERSPACE_ATTR_MAX + 1]; int error; @@ -661,8 +706,7 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info) err = -EINVAL; if (!a[OVS_PACKET_ATTR_PACKET] || !a[OVS_PACKET_ATTR_KEY] || - !a[OVS_PACKET_ATTR_ACTIONS] || - nla_len(a[OVS_PACKET_ATTR_PACKET]) < ETH_HLEN) + !a[OVS_PACKET_ATTR_ACTIONS]) goto err; len = nla_len(a[OVS_PACKET_ATTR_PACKET]); @@ -672,7 +716,7 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info) goto err; skb_reserve(packet, NET_IP_ALIGN); - memcpy(__skb_put(packet, len), nla_data(a[OVS_PACKET_ATTR_PACKET]), len); + nla_memcpy(__skb_put(packet, len), a[OVS_PACKET_ATTR_PACKET], len); skb_reset_mac_header(packet); eth = eth_hdr(packet); @@ -680,7 +724,7 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info) /* 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); @@ -743,7 +787,7 @@ err: } static const struct nla_policy packet_policy[OVS_PACKET_ATTR_MAX + 1] = { - [OVS_PACKET_ATTR_PACKET] = { .type = NLA_UNSPEC }, + [OVS_PACKET_ATTR_PACKET] = { .len = ETH_HLEN }, [OVS_PACKET_ATTR_KEY] = { .type = NLA_NESTED }, [OVS_PACKET_ATTR_ACTIONS] = { .type = NLA_NESTED }, }; @@ -759,7 +803,7 @@ static struct genl_ops dp_packet_genl_ops[] = { static void get_dp_stats(struct datapath *dp, struct ovs_dp_stats *stats) { int i; - struct flow_table *table = genl_dereference(dp->table); + struct flow_table *table = ovsl_dereference(dp->table); stats->n_flows = ovs_flow_tbl_count(table); @@ -794,14 +838,25 @@ static struct genl_family dp_flow_genl_family = { .name = OVS_FLOW_FAMILY, .version = OVS_FLOW_VERSION, .maxattr = OVS_FLOW_ATTR_MAX, - .netnsok = true + .netnsok = true, + .parallel_ops = true, }; static struct genl_multicast_group ovs_dp_flow_multicast_group = { .name = OVS_FLOW_MCGROUP }; -/* Called with genl_lock. */ +static size_t ovs_flow_cmd_msg_size(const struct sw_flow_actions *acts) +{ + return NLMSG_ALIGN(sizeof(struct ovs_header)) + + nla_total_size(key_attr_size()) /* OVS_FLOW_ATTR_KEY */ + + nla_total_size(sizeof(struct ovs_flow_stats)) /* OVS_FLOW_ATTR_STATS */ + + nla_total_size(1) /* OVS_FLOW_ATTR_TCP_FLAGS */ + + nla_total_size(8) /* OVS_FLOW_ATTR_USED */ + + nla_total_size(acts->actions_len); /* OVS_FLOW_ATTR_ACTIONS */ +} + +/* Called with ovs_mutex. */ static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp, struct sk_buff *skb, u32 portid, u32 seq, u32 flags, u8 cmd) @@ -815,8 +870,7 @@ static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp, u8 tcp_flags; int err; - sf_acts = rcu_dereference_protected(flow->sf_acts, - lockdep_genl_is_held()); + sf_acts = ovsl_dereference(flow->sf_acts); ovs_header = genlmsg_put(skb, portid, seq, &dp_flow_genl_family, flags, cmd); if (!ovs_header) @@ -879,25 +933,10 @@ error: static struct sk_buff *ovs_flow_cmd_alloc_info(struct sw_flow *flow) { const struct sw_flow_actions *sf_acts; - int len; - sf_acts = rcu_dereference_protected(flow->sf_acts, - lockdep_genl_is_held()); + sf_acts = ovsl_dereference(flow->sf_acts); - /* OVS_FLOW_ATTR_KEY */ - len = nla_total_size(FLOW_BUFSIZE); - /* OVS_FLOW_ATTR_ACTIONS */ - len += nla_total_size(sf_acts->actions_len); - /* OVS_FLOW_ATTR_STATS */ - len += nla_total_size(sizeof(struct ovs_flow_stats)); - /* OVS_FLOW_ATTR_TCP_FLAGS */ - len += nla_total_size(1); - /* OVS_FLOW_ATTR_USED */ - len += nla_total_size(8); - - len += NLMSG_ALIGN(sizeof(struct ovs_header)); - - return genlmsg_new(len, GFP_KERNEL); + return genlmsg_new(ovs_flow_cmd_msg_size(sf_acts), GFP_KERNEL); } static struct sk_buff *ovs_flow_cmd_build_info(struct sw_flow *flow, @@ -946,12 +985,13 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) goto error; } + ovs_lock(); dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex); error = -ENODEV; if (!dp) - goto error; + goto err_unlock_ovs; - table = genl_dereference(dp->table); + table = ovsl_dereference(dp->table); flow = ovs_flow_tbl_lookup(table, &key, key_len); if (!flow) { struct sw_flow_actions *acts; @@ -959,7 +999,7 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) /* Bail out if we're not allowed to create a new flow. */ error = -ENOENT; if (info->genlhdr->cmd == OVS_FLOW_CMD_SET) - goto error; + goto err_unlock_ovs; /* Expand table, if necessary, to make room. */ if (ovs_flow_tbl_need_to_expand(table)) { @@ -969,7 +1009,7 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) if (!IS_ERR(new_table)) { rcu_assign_pointer(dp->table, new_table); ovs_flow_tbl_deferred_destroy(table); - table = genl_dereference(dp->table); + table = ovsl_dereference(dp->table); } } @@ -977,7 +1017,7 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) flow = ovs_flow_alloc(); if (IS_ERR(flow)) { error = PTR_ERR(flow); - goto error; + goto err_unlock_ovs; } flow->key = key; clear_stats(flow); @@ -1010,11 +1050,10 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) error = -EEXIST; if (info->genlhdr->cmd == OVS_FLOW_CMD_NEW && info->nlhdr->nlmsg_flags & (NLM_F_CREATE | NLM_F_EXCL)) - goto error; + goto err_unlock_ovs; /* Update actions. */ - old_acts = rcu_dereference_protected(flow->sf_acts, - lockdep_genl_is_held()); + old_acts = ovsl_dereference(flow->sf_acts); acts_attrs = a[OVS_FLOW_ATTR_ACTIONS]; if (acts_attrs && (old_acts->actions_len != nla_len(acts_attrs) || @@ -1025,7 +1064,7 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) new_acts = ovs_flow_actions_alloc(acts_attrs); error = PTR_ERR(new_acts); if (IS_ERR(new_acts)) - goto error; + goto err_unlock_ovs; rcu_assign_pointer(flow->sf_acts, new_acts); ovs_flow_deferred_free_acts(old_acts); @@ -1041,11 +1080,10 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) spin_unlock_bh(&flow->lock); } } + ovs_unlock(); if (!IS_ERR(reply)) - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_flow_multicast_group.id, info->nlhdr, - GFP_KERNEL); + ovs_notify(reply, info, &ovs_dp_flow_multicast_group); else netlink_set_err(sock_net(skb->sk)->genl_sock, 0, ovs_dp_flow_multicast_group.id, PTR_ERR(reply)); @@ -1053,6 +1091,8 @@ static int ovs_flow_cmd_new_or_set(struct sk_buff *skb, struct genl_info *info) error_free_flow: ovs_flow_free(flow); +err_unlock_ovs: + ovs_unlock(); error: return error; } @@ -1075,21 +1115,32 @@ static int ovs_flow_cmd_get(struct sk_buff *skb, struct genl_info *info) if (err) return err; + ovs_lock(); dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex); - if (!dp) - return -ENODEV; + if (!dp) { + err = -ENODEV; + goto unlock; + } - table = genl_dereference(dp->table); + table = ovsl_dereference(dp->table); flow = ovs_flow_tbl_lookup(table, &key, key_len); - if (!flow) - return -ENOENT; + if (!flow) { + err = -ENOENT; + goto unlock; + } reply = ovs_flow_cmd_build_info(flow, dp, info->snd_portid, info->snd_seq, OVS_FLOW_CMD_NEW); - if (IS_ERR(reply)) - return PTR_ERR(reply); + if (IS_ERR(reply)) { + err = PTR_ERR(reply); + goto unlock; + } + ovs_unlock(); return genlmsg_reply(reply, info); +unlock: + ovs_unlock(); + return err; } static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info) @@ -1104,25 +1155,33 @@ static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info) int err; int key_len; + ovs_lock(); dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex); - if (!dp) - return -ENODEV; - - if (!a[OVS_FLOW_ATTR_KEY]) - return flush_flows(dp); + if (!dp) { + err = -ENODEV; + goto unlock; + } + if (!a[OVS_FLOW_ATTR_KEY]) { + err = flush_flows(dp); + goto unlock; + } err = ovs_flow_from_nlattrs(&key, &key_len, a[OVS_FLOW_ATTR_KEY]); if (err) - return err; + goto unlock; - table = genl_dereference(dp->table); + table = ovsl_dereference(dp->table); flow = ovs_flow_tbl_lookup(table, &key, key_len); - if (!flow) - return -ENOENT; + if (!flow) { + err = -ENOENT; + goto unlock; + } reply = ovs_flow_cmd_alloc_info(flow); - if (!reply) - return -ENOMEM; + if (!reply) { + err = -ENOMEM; + goto unlock; + } ovs_flow_tbl_remove(table, flow); @@ -1131,10 +1190,13 @@ static int ovs_flow_cmd_del(struct sk_buff *skb, struct genl_info *info) BUG_ON(err < 0); ovs_flow_deferred_free(flow); + ovs_unlock(); - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_flow_multicast_group.id, info->nlhdr, GFP_KERNEL); + ovs_notify(reply, info, &ovs_dp_flow_multicast_group); return 0; +unlock: + ovs_unlock(); + return err; } static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) @@ -1143,11 +1205,14 @@ static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) struct datapath *dp; struct flow_table *table; + ovs_lock(); dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex); - if (!dp) + if (!dp) { + ovs_unlock(); return -ENODEV; + } - table = genl_dereference(dp->table); + table = ovsl_dereference(dp->table); for (;;) { struct sw_flow *flow; @@ -1168,6 +1233,7 @@ static int ovs_flow_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) cb->args[0] = bucket; cb->args[1] = obj; } + ovs_unlock(); return skb->len; } @@ -1206,13 +1272,24 @@ static struct genl_family dp_datapath_genl_family = { .name = OVS_DATAPATH_FAMILY, .version = OVS_DATAPATH_VERSION, .maxattr = OVS_DP_ATTR_MAX, - .netnsok = true + .netnsok = true, + .parallel_ops = true, }; static struct genl_multicast_group ovs_dp_datapath_multicast_group = { .name = OVS_DATAPATH_MCGROUP }; +static size_t ovs_dp_cmd_msg_size(void) +{ + size_t msgsize = NLMSG_ALIGN(sizeof(struct ovs_header)); + + msgsize += nla_total_size(IFNAMSIZ); + msgsize += nla_total_size(sizeof(struct ovs_dp_stats)); + + return msgsize; +} + static int ovs_dp_cmd_fill_info(struct datapath *dp, struct sk_buff *skb, u32 portid, u32 seq, u32 flags, u8 cmd) { @@ -1251,7 +1328,7 @@ static struct sk_buff *ovs_dp_cmd_build_info(struct datapath *dp, u32 portid, struct sk_buff *skb; int retval; - skb = nlmsg_new(NLMSG_DEFAULT_SIZE, GFP_KERNEL); + skb = genlmsg_new(ovs_dp_cmd_msg_size(), GFP_KERNEL); if (!skb) return ERR_PTR(-ENOMEM); @@ -1263,7 +1340,7 @@ static struct sk_buff *ovs_dp_cmd_build_info(struct datapath *dp, u32 portid, return skb; } -/* Called with genl_mutex and optionally with RTNL lock also. */ +/* Called with ovs_mutex. */ static struct datapath *lookup_datapath(struct net *net, struct ovs_header *ovs_header, struct nlattr *a[OVS_DP_ATTR_MAX + 1]) @@ -1297,12 +1374,12 @@ static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info) if (!a[OVS_DP_ATTR_NAME] || !a[OVS_DP_ATTR_UPCALL_PID]) goto err; - rtnl_lock(); + ovs_lock(); err = -ENOMEM; dp = kzalloc(sizeof(*dp), GFP_KERNEL); if (dp == NULL) - goto err_unlock_rtnl; + goto err_unlock_ovs; ovs_dp_set_net(dp, hold_net(sock_net(skb->sk))); @@ -1353,37 +1430,34 @@ static int ovs_dp_cmd_new(struct sk_buff *skb, struct genl_info *info) ovs_net = net_generic(ovs_dp_get_net(dp), ovs_net_id); list_add_tail(&dp->list_node, &ovs_net->dps); - rtnl_unlock(); - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_datapath_multicast_group.id, info->nlhdr, - GFP_KERNEL); + ovs_unlock(); + + ovs_notify(reply, info, &ovs_dp_datapath_multicast_group); return 0; err_destroy_local_port: - ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_LOCAL)); + ovs_dp_detach_port(ovs_vport_ovsl(dp, OVSP_LOCAL)); err_destroy_ports_array: kfree(dp->ports); err_destroy_percpu: free_percpu(dp->stats_percpu); err_destroy_table: - ovs_flow_tbl_destroy(genl_dereference(dp->table)); + ovs_flow_tbl_destroy(ovsl_dereference(dp->table)); err_free_dp: release_net(ovs_dp_get_net(dp)); kfree(dp); -err_unlock_rtnl: - rtnl_unlock(); +err_unlock_ovs: + ovs_unlock(); err: return err; } -/* Called with genl_mutex. */ +/* Called with ovs_mutex. */ static void __dp_destroy(struct datapath *dp) { int i; - rtnl_lock(); - for (i = 0; i < DP_VPORT_HASH_BUCKETS; i++) { struct vport *vport; struct hlist_node *n; @@ -1394,14 +1468,11 @@ static void __dp_destroy(struct datapath *dp) } list_del(&dp->list_node); - ovs_dp_detach_port(ovs_vport_rtnl(dp, OVSP_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. + /* OVSP_LOCAL is datapath internal port. We need to make sure that + * all port in datapath are destroyed first before freeing datapath. */ - rtnl_unlock(); + ovs_dp_detach_port(ovs_vport_ovsl(dp, OVSP_LOCAL)); call_rcu(&dp->rcu, destroy_dp_rcu); } @@ -1412,24 +1483,27 @@ static int ovs_dp_cmd_del(struct sk_buff *skb, struct genl_info *info) struct datapath *dp; int err; + ovs_lock(); dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs); err = PTR_ERR(dp); if (IS_ERR(dp)) - return err; + goto unlock; reply = ovs_dp_cmd_build_info(dp, info->snd_portid, info->snd_seq, OVS_DP_CMD_DEL); err = PTR_ERR(reply); if (IS_ERR(reply)) - return err; + goto unlock; __dp_destroy(dp); + ovs_unlock(); - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_datapath_multicast_group.id, info->nlhdr, - GFP_KERNEL); + ovs_notify(reply, info, &ovs_dp_datapath_multicast_group); return 0; +unlock: + ovs_unlock(); + return err; } static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info) @@ -1438,9 +1512,11 @@ static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info) struct datapath *dp; int err; + ovs_lock(); dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs); + err = PTR_ERR(dp); if (IS_ERR(dp)) - return PTR_ERR(dp); + goto unlock; reply = ovs_dp_cmd_build_info(dp, info->snd_portid, info->snd_seq, OVS_DP_CMD_NEW); @@ -1448,31 +1524,45 @@ static int ovs_dp_cmd_set(struct sk_buff *skb, struct genl_info *info) err = PTR_ERR(reply); netlink_set_err(sock_net(skb->sk)->genl_sock, 0, ovs_dp_datapath_multicast_group.id, err); - return 0; + err = 0; + goto unlock; } - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_datapath_multicast_group.id, info->nlhdr, - GFP_KERNEL); + ovs_unlock(); + ovs_notify(reply, info, &ovs_dp_datapath_multicast_group); return 0; +unlock: + ovs_unlock(); + return err; } static int ovs_dp_cmd_get(struct sk_buff *skb, struct genl_info *info) { struct sk_buff *reply; struct datapath *dp; + int err; + ovs_lock(); dp = lookup_datapath(sock_net(skb->sk), info->userhdr, info->attrs); - if (IS_ERR(dp)) - return PTR_ERR(dp); + if (IS_ERR(dp)) { + err = PTR_ERR(dp); + goto unlock; + } reply = ovs_dp_cmd_build_info(dp, info->snd_portid, info->snd_seq, OVS_DP_CMD_NEW); - if (IS_ERR(reply)) - return PTR_ERR(reply); + if (IS_ERR(reply)) { + err = PTR_ERR(reply); + goto unlock; + } + ovs_unlock(); return genlmsg_reply(reply, info); + +unlock: + ovs_unlock(); + return err; } static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) @@ -1482,6 +1572,7 @@ static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) int skip = cb->args[0]; int i = 0; + ovs_lock(); list_for_each_entry(dp, &ovs_net->dps, list_node) { if (i >= skip && ovs_dp_cmd_fill_info(dp, skb, NETLINK_CB(cb->skb).portid, @@ -1490,6 +1581,7 @@ static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb) break; i++; } + ovs_unlock(); cb->args[0] = i; @@ -1535,14 +1627,15 @@ static struct genl_family dp_vport_genl_family = { .name = OVS_VPORT_FAMILY, .version = OVS_VPORT_VERSION, .maxattr = OVS_VPORT_ATTR_MAX, - .netnsok = true + .netnsok = true, + .parallel_ops = true, }; struct genl_multicast_group ovs_dp_vport_multicast_group = { .name = OVS_VPORT_MCGROUP }; -/* Called with RTNL lock or RCU read lock. */ +/* Called with ovs_mutex or RCU read lock. */ static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb, u32 portid, u32 seq, u32 flags, u8 cmd) { @@ -1581,7 +1674,7 @@ error: return err; } -/* Called with RTNL lock or RCU read lock. */ +/* Called with ovs_mutex or RCU read lock. */ struct sk_buff *ovs_vport_cmd_build_info(struct vport *vport, u32 portid, u32 seq, u8 cmd) { @@ -1598,7 +1691,7 @@ struct sk_buff *ovs_vport_cmd_build_info(struct vport *vport, u32 portid, return skb; } -/* Called with RTNL lock or RCU read lock. */ +/* Called with ovs_mutex or RCU read lock. */ static struct vport *lookup_vport(struct net *net, struct ovs_header *ovs_header, struct nlattr *a[OVS_VPORT_ATTR_MAX + 1]) @@ -1624,9 +1717,9 @@ static struct vport *lookup_vport(struct net *net, if (!dp) return ERR_PTR(-ENODEV); - vport = ovs_vport_rtnl_rcu(dp, port_no); + vport = ovs_vport_ovsl_rcu(dp, port_no); if (!vport) - return ERR_PTR(-ENOENT); + return ERR_PTR(-ENODEV); return vport; } else return ERR_PTR(-EINVAL); @@ -1648,7 +1741,7 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info) !a[OVS_VPORT_ATTR_UPCALL_PID]) goto exit; - rtnl_lock(); + ovs_lock(); dp = get_dp(sock_net(skb->sk), ovs_header->dp_ifindex); err = -ENODEV; if (!dp) @@ -1661,7 +1754,7 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info) if (port_no >= DP_MAX_PORTS) goto exit_unlock; - vport = ovs_vport_rtnl_rcu(dp, port_no); + vport = ovs_vport_ovsl(dp, port_no); err = -EBUSY; if (vport) goto exit_unlock; @@ -1671,7 +1764,7 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info) err = -EFBIG; goto exit_unlock; } - vport = ovs_vport_rtnl(dp, port_no); + vport = ovs_vport_ovsl(dp, port_no); if (!vport) break; } @@ -1697,11 +1790,11 @@ static int ovs_vport_cmd_new(struct sk_buff *skb, struct genl_info *info) ovs_dp_detach_port(vport); goto exit_unlock; } - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL); + + ovs_notify(reply, info, &ovs_dp_vport_multicast_group); exit_unlock: - rtnl_unlock(); + ovs_unlock(); exit: return err; } @@ -1713,7 +1806,7 @@ static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info) struct vport *vport; int err; - rtnl_lock(); + ovs_lock(); vport = lookup_vport(sock_net(skb->sk), info->userhdr, a); err = PTR_ERR(vport); if (IS_ERR(vport)) @@ -1742,8 +1835,9 @@ static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info) info->snd_seq, 0, OVS_VPORT_CMD_NEW); BUG_ON(err < 0); - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL); + ovs_unlock(); + ovs_notify(reply, info, &ovs_dp_vport_multicast_group); + return 0; rtnl_unlock(); return 0; @@ -1751,7 +1845,7 @@ static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info) exit_free: kfree_skb(reply); exit_unlock: - rtnl_unlock(); + ovs_unlock(); return err; } @@ -1762,7 +1856,7 @@ static int ovs_vport_cmd_del(struct sk_buff *skb, struct genl_info *info) struct vport *vport; int err; - rtnl_lock(); + ovs_lock(); vport = lookup_vport(sock_net(skb->sk), info->userhdr, a); err = PTR_ERR(vport); if (IS_ERR(vport)) @@ -1782,11 +1876,10 @@ static int ovs_vport_cmd_del(struct sk_buff *skb, struct genl_info *info) err = 0; ovs_dp_detach_port(vport); - genl_notify(reply, genl_info_net(info), info->snd_portid, - ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL); + ovs_notify(reply, info, &ovs_dp_vport_multicast_group); exit_unlock: - rtnl_unlock(); + ovs_unlock(); return err; } @@ -1946,13 +2039,13 @@ static void rehash_flow_table(struct work_struct *work) struct datapath *dp; struct net *net; - genl_lock(); + ovs_lock(); rtnl_lock(); for_each_net(net) { struct ovs_net *ovs_net = net_generic(net, ovs_net_id); list_for_each_entry(dp, &ovs_net->dps, list_node) { - struct flow_table *old_table = genl_dereference(dp->table); + struct flow_table *old_table = ovsl_dereference(dp->table); struct flow_table *new_table; new_table = ovs_flow_tbl_rehash(old_table); @@ -1963,8 +2056,7 @@ static void rehash_flow_table(struct work_struct *work) } } rtnl_unlock(); - genl_unlock(); - + ovs_unlock(); schedule_delayed_work(&rehash_flow_wq, REHASH_FLOW_INTERVAL); } @@ -1973,18 +2065,21 @@ 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 ovs_net *ovs_net = net_generic(net, ovs_net_id); struct datapath *dp, *dp_next; + struct ovs_net *ovs_net = net_generic(net, ovs_net_id); - genl_lock(); + ovs_lock(); list_for_each_entry_safe(dp, dp_next, &ovs_net->dps, list_node) __dp_destroy(dp); - genl_unlock(); + ovs_unlock(); + + cancel_work_sync(&ovs_net->dp_notify_work); } static struct pernet_operations ovs_net_ops = { diff --git a/net/openvswitch/datapath.h b/net/openvswitch/datapath.h index 031dfbf37c93..16b840695216 100644 --- a/net/openvswitch/datapath.h +++ b/net/openvswitch/datapath.h @@ -57,10 +57,9 @@ struct dp_stats_percpu { * struct datapath - datapath for flow-based packet switching * @rcu: RCU callback head for deferred destruction. * @list_node: Element in global 'dps' list. - * @n_flows: Number of flows currently in flow table. - * @table: Current flow table. Protected by genl_lock and RCU. + * @table: Current flow table. Protected by ovs_mutex and RCU. * @ports: Hash table for ports. %OVSP_LOCAL port always exists. Protected by - * RTNL and RCU. + * ovs_mutex and RCU. * @stats_percpu: Per-CPU datapath statistics. * @net: Reference to net namespace. * @@ -86,26 +85,6 @@ struct datapath { #endif }; -struct vport *ovs_lookup_vport(const struct datapath *dp, u16 port_no); - -static inline struct vport *ovs_vport_rcu(const struct datapath *dp, int port_no) -{ - WARN_ON_ONCE(!rcu_read_lock_held()); - return ovs_lookup_vport(dp, port_no); -} - -static inline struct vport *ovs_vport_rtnl_rcu(const struct datapath *dp, int port_no) -{ - WARN_ON_ONCE(!rcu_read_lock_held() && !rtnl_is_locked()); - return ovs_lookup_vport(dp, port_no); -} - -static inline struct vport *ovs_vport_rtnl(const struct datapath *dp, int port_no) -{ - ASSERT_RTNL(); - return ovs_lookup_vport(dp, port_no); -} - /** * struct ovs_skb_cb - OVS data in skb CB * @flow: The flow associated with this packet. May be %NULL if no flow. @@ -119,7 +98,7 @@ struct ovs_skb_cb { * struct dp_upcall - metadata to include with a packet to send to userspace * @cmd: One of %OVS_PACKET_CMD_*. * @key: Becomes %OVS_PACKET_ATTR_KEY. Must be nonnull. - * @userdata: If nonnull, its u64 value is extracted and passed to userspace as + * @userdata: If nonnull, its variable-length value is passed to userspace as * %OVS_PACKET_ATTR_USERDATA. * @pid: Netlink PID to which packet should be sent. If @pid is 0 then no * packet is sent and the packet is accounted in the datapath's @n_lost @@ -132,6 +111,30 @@ struct dp_upcall_info { u32 portid; }; +/** + * struct ovs_net - Per net-namespace data for ovs. + * @dps: List of datapaths to enable dumping them all out. + * Protected by genl_mutex. + */ +struct ovs_net { + struct list_head dps; + struct work_struct dp_notify_work; +}; + +extern int ovs_net_id; +void ovs_lock(void); +void ovs_unlock(void); + +#ifdef CONFIG_LOCKDEP +int lockdep_ovsl_is_held(void); +#else +#define lockdep_ovsl_is_held() 1 +#endif + +#define ASSERT_OVSL() WARN_ON(unlikely(!lockdep_ovsl_is_held())) +#define ovsl_dereference(p) \ + rcu_dereference_protected(p, lockdep_ovsl_is_held()) + static inline struct net *ovs_dp_get_net(struct datapath *dp) { return read_pnet(&dp->net); @@ -142,6 +145,26 @@ static inline void ovs_dp_set_net(struct datapath *dp, struct net *net) write_pnet(&dp->net, net); } +struct vport *ovs_lookup_vport(const struct datapath *dp, u16 port_no); + +static inline struct vport *ovs_vport_rcu(const struct datapath *dp, int port_no) +{ + WARN_ON_ONCE(!rcu_read_lock_held()); + return ovs_lookup_vport(dp, port_no); +} + +static inline struct vport *ovs_vport_ovsl_rcu(const struct datapath *dp, int port_no) +{ + WARN_ON_ONCE(!rcu_read_lock_held() && !lockdep_ovsl_is_held()); + return ovs_lookup_vport(dp, port_no); +} + +static inline struct vport *ovs_vport_ovsl(const struct datapath *dp, int port_no) +{ + ASSERT_OVSL(); + return ovs_lookup_vport(dp, port_no); +} + extern struct notifier_block ovs_dp_device_notifier; extern struct genl_multicast_group ovs_dp_vport_multicast_group; @@ -155,4 +178,5 @@ struct sk_buff *ovs_vport_cmd_build_info(struct vport *, u32 pid, u32 seq, u8 cmd); int ovs_execute_actions(struct datapath *dp, struct sk_buff *skb); +void ovs_dp_notify_wq(struct work_struct *work); #endif /* datapath.h */ diff --git a/net/openvswitch/dp_notify.c b/net/openvswitch/dp_notify.c index 5558350e0d33..ef4feec6cd84 100644 --- a/net/openvswitch/dp_notify.c +++ b/net/openvswitch/dp_notify.c @@ -18,46 +18,78 @@ #include <linux/netdevice.h> #include <net/genetlink.h> +#include <net/netns/generic.h> #include "datapath.h" #include "vport-internal_dev.h" #include "vport-netdev.h" +static void dp_detach_port_notify(struct vport *vport) +{ + struct sk_buff *notify; + struct datapath *dp; + + dp = vport->dp; + notify = ovs_vport_cmd_build_info(vport, 0, 0, + OVS_VPORT_CMD_DEL); + ovs_dp_detach_port(vport); + if (IS_ERR(notify)) { + netlink_set_err(ovs_dp_get_net(dp)->genl_sock, 0, + ovs_dp_vport_multicast_group.id, + PTR_ERR(notify)); + return; + } + + genlmsg_multicast_netns(ovs_dp_get_net(dp), notify, 0, + ovs_dp_vport_multicast_group.id, + GFP_KERNEL); +} + +void ovs_dp_notify_wq(struct work_struct *work) +{ + struct ovs_net *ovs_net = container_of(work, struct ovs_net, dp_notify_work); + struct datapath *dp; + + ovs_lock(); + list_for_each_entry(dp, &ovs_net->dps, list_node) { + 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) { + struct netdev_vport *netdev_vport; + + if (vport->ops->type != OVS_VPORT_TYPE_NETDEV) + continue; + + netdev_vport = netdev_vport_priv(vport); + if (netdev_vport->dev->reg_state == NETREG_UNREGISTERED || + netdev_vport->dev->reg_state == NETREG_UNREGISTERING) + dp_detach_port_notify(vport); + } + } + } + ovs_unlock(); +} + static int dp_device_event(struct notifier_block *unused, unsigned long event, void *ptr) { + struct ovs_net *ovs_net; struct net_device *dev = ptr; - struct vport *vport; + struct vport *vport = NULL; - if (ovs_is_internal_dev(dev)) - vport = ovs_internal_dev_get_vport(dev); - else + if (!ovs_is_internal_dev(dev)) vport = ovs_netdev_get_vport(dev); if (!vport) return NOTIFY_DONE; - switch (event) { - case NETDEV_UNREGISTER: - if (!ovs_is_internal_dev(dev)) { - struct sk_buff *notify; - struct datapath *dp = vport->dp; - - notify = ovs_vport_cmd_build_info(vport, 0, 0, - OVS_VPORT_CMD_DEL); - ovs_dp_detach_port(vport); - if (IS_ERR(notify)) { - netlink_set_err(ovs_dp_get_net(dp)->genl_sock, 0, - ovs_dp_vport_multicast_group.id, - PTR_ERR(notify)); - break; - } - - genlmsg_multicast_netns(ovs_dp_get_net(dp), notify, 0, - ovs_dp_vport_multicast_group.id, - GFP_KERNEL); - } - break; + if (event == NETDEV_UNREGISTER) { + ovs_net = net_generic(dev_net(dev), ovs_net_id); + queue_work(system_wq, &ovs_net->dp_notify_work); } return NOTIFY_DONE; diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c index 67a2b783fe70..b15321a2228c 100644 --- a/net/openvswitch/flow.c +++ b/net/openvswitch/flow.c @@ -211,7 +211,7 @@ struct sw_flow_actions *ovs_flow_actions_alloc(const struct nlattr *actions) return ERR_PTR(-ENOMEM); sfa->actions_len = actions_len; - memcpy(sfa->actions, nla_data(actions), actions_len); + nla_memcpy(sfa->actions, actions, actions_len); return sfa; } @@ -466,7 +466,7 @@ static __be16 parse_ethertype(struct sk_buff *skb) proto = *(__be16 *) skb->data; __skb_pull(skb, sizeof(__be16)); - if (ntohs(proto) >= 1536) + if (ntohs(proto) >= ETH_P_802_3_MIN) return proto; if (skb->len < sizeof(struct llc_snap_hdr)) @@ -483,7 +483,7 @@ static __be16 parse_ethertype(struct sk_buff *skb) __skb_pull(skb, sizeof(struct llc_snap_hdr)); - if (ntohs(llc->ethertype) >= 1536) + if (ntohs(llc->ethertype) >= ETH_P_802_3_MIN) return llc->ethertype; return htons(ETH_P_802_2); @@ -1038,7 +1038,7 @@ int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, if (attrs & (1 << OVS_KEY_ATTR_ETHERTYPE)) { swkey->eth.type = nla_get_be16(a[OVS_KEY_ATTR_ETHERTYPE]); - if (ntohs(swkey->eth.type) < 1536) + if (ntohs(swkey->eth.type) < ETH_P_802_3_MIN) return -EINVAL; attrs &= ~(1 << OVS_KEY_ATTR_ETHERTYPE); } else { diff --git a/net/openvswitch/flow.h b/net/openvswitch/flow.h index a7bb60ff3b5b..0875fde65b9c 100644 --- a/net/openvswitch/flow.h +++ b/net/openvswitch/flow.h @@ -138,27 +138,6 @@ int ovs_flow_extract(struct sk_buff *, u16 in_port, struct sw_flow_key *, void ovs_flow_used(struct sw_flow *, struct sk_buff *); u64 ovs_flow_used_time(unsigned long flow_jiffies); -/* Upper bound on the length of a nlattr-formatted flow key. The longest - * nlattr-formatted flow key would be: - * - * struct pad nl hdr total - * ------ --- ------ ----- - * OVS_KEY_ATTR_PRIORITY 4 -- 4 8 - * OVS_KEY_ATTR_IN_PORT 4 -- 4 8 - * OVS_KEY_ATTR_SKB_MARK 4 -- 4 8 - * OVS_KEY_ATTR_ETHERNET 12 -- 4 16 - * OVS_KEY_ATTR_ETHERTYPE 2 2 4 8 (outer VLAN ethertype) - * OVS_KEY_ATTR_8021Q 4 -- 4 8 - * OVS_KEY_ATTR_ENCAP 0 -- 4 4 (VLAN encapsulation) - * OVS_KEY_ATTR_ETHERTYPE 2 2 4 8 (inner VLAN ethertype) - * OVS_KEY_ATTR_IPV6 40 -- 4 44 - * OVS_KEY_ATTR_ICMPV6 2 2 4 8 - * OVS_KEY_ATTR_ND 28 -- 4 32 - * ------------------------------------------------- - * total 152 - */ -#define FLOW_BUFSIZE 152 - int ovs_flow_to_nlattrs(const struct sw_flow_key *, struct sk_buff *); int ovs_flow_from_nlattrs(struct sw_flow_key *swkey, int *key_lenp, const struct nlattr *); diff --git a/net/openvswitch/vport-internal_dev.c b/net/openvswitch/vport-internal_dev.c index 0531de6c7a4a..84e0a0379186 100644 --- a/net/openvswitch/vport-internal_dev.c +++ b/net/openvswitch/vport-internal_dev.c @@ -63,16 +63,6 @@ static struct rtnl_link_stats64 *internal_dev_get_stats(struct net_device *netde return stats; } -static int internal_dev_mac_addr(struct net_device *dev, void *p) -{ - struct sockaddr *addr = p; - - if (!is_valid_ether_addr(addr->sa_data)) - return -EADDRNOTAVAIL; - memcpy(dev->dev_addr, addr->sa_data, dev->addr_len); - return 0; -} - /* Called with rcu_read_lock_bh. */ static int internal_dev_xmit(struct sk_buff *skb, struct net_device *netdev) { @@ -126,7 +116,7 @@ static const struct net_device_ops internal_dev_netdev_ops = { .ndo_open = internal_dev_open, .ndo_stop = internal_dev_stop, .ndo_start_xmit = internal_dev_xmit, - .ndo_set_mac_address = internal_dev_mac_addr, + .ndo_set_mac_address = eth_mac_addr, .ndo_change_mtu = internal_dev_change_mtu, .ndo_get_stats64 = internal_dev_get_stats, }; @@ -138,6 +128,7 @@ static void do_setup(struct net_device *netdev) netdev->netdev_ops = &internal_dev_netdev_ops; netdev->priv_flags &= ~IFF_TX_SKB_SHARING; + netdev->priv_flags |= IFF_LIVE_ADDR_CHANGE; netdev->destructor = internal_dev_destructor; SET_ETHTOOL_OPS(netdev, &internal_dev_ethtool_ops); netdev->tx_queue_len = 0; @@ -146,7 +137,7 @@ static void do_setup(struct net_device *netdev) NETIF_F_HIGHDMA | NETIF_F_HW_CSUM | NETIF_F_TSO; netdev->vlan_features = netdev->features; - netdev->features |= NETIF_F_HW_VLAN_TX; + netdev->features |= NETIF_F_HW_VLAN_CTAG_TX; netdev->hw_features = netdev->features & ~NETIF_F_LLTX; eth_hw_addr_random(netdev); } @@ -182,16 +173,19 @@ static struct vport *internal_dev_create(const struct vport_parms *parms) if (vport->port_no == OVSP_LOCAL) netdev_vport->dev->features |= NETIF_F_NETNS_LOCAL; + rtnl_lock(); err = register_netdevice(netdev_vport->dev); if (err) goto error_free_netdev; dev_set_promiscuity(netdev_vport->dev, 1); + rtnl_unlock(); netif_start_queue(netdev_vport->dev); return vport; error_free_netdev: + rtnl_unlock(); free_netdev(netdev_vport->dev); error_free_vport: ovs_vport_free(vport); @@ -204,10 +198,13 @@ static void internal_dev_destroy(struct vport *vport) struct netdev_vport *netdev_vport = netdev_vport_priv(vport); netif_stop_queue(netdev_vport->dev); + rtnl_lock(); dev_set_promiscuity(netdev_vport->dev, -1); /* unregister_netdevice() waits for an RCU grace period. */ unregister_netdevice(netdev_vport->dev); + + rtnl_unlock(); } static int internal_dev_recv(struct vport *vport, struct sk_buff *skb) @@ -235,7 +232,6 @@ const struct vport_ops ovs_internal_vport_ops = { .create = internal_dev_create, .destroy = internal_dev_destroy, .get_name = ovs_netdev_get_name, - .get_ifindex = ovs_netdev_get_ifindex, .send = internal_dev_recv, }; diff --git a/net/openvswitch/vport-netdev.c b/net/openvswitch/vport-netdev.c index 2130d61c384a..4f01c6d2ffa4 100644 --- a/net/openvswitch/vport-netdev.c +++ b/net/openvswitch/vport-netdev.c @@ -100,16 +100,20 @@ static struct vport *netdev_create(const struct vport_parms *parms) goto error_put; } + rtnl_lock(); err = netdev_rx_handler_register(netdev_vport->dev, netdev_frame_hook, vport); if (err) - goto error_put; + goto error_unlock; dev_set_promiscuity(netdev_vport->dev, 1); netdev_vport->dev->priv_flags |= IFF_OVS_DATAPATH; + rtnl_unlock(); return vport; +error_unlock: + rtnl_unlock(); error_put: dev_put(netdev_vport->dev); error_free_vport: @@ -131,9 +135,11 @@ static void netdev_destroy(struct vport *vport) { struct netdev_vport *netdev_vport = netdev_vport_priv(vport); + rtnl_lock(); netdev_vport->dev->priv_flags &= ~IFF_OVS_DATAPATH; netdev_rx_handler_unregister(netdev_vport->dev); dev_set_promiscuity(netdev_vport->dev, -1); + rtnl_unlock(); call_rcu(&netdev_vport->rcu, free_port_rcu); } @@ -144,12 +150,6 @@ const char *ovs_netdev_get_name(const struct vport *vport) return netdev_vport->dev->name; } -int ovs_netdev_get_ifindex(const struct vport *vport) -{ - const struct netdev_vport *netdev_vport = netdev_vport_priv(vport); - return netdev_vport->dev->ifindex; -} - static unsigned int packet_length(const struct sk_buff *skb) { unsigned int length = skb->len - ETH_HLEN; @@ -200,6 +200,5 @@ const struct vport_ops ovs_netdev_vport_ops = { .create = netdev_create, .destroy = netdev_destroy, .get_name = ovs_netdev_get_name, - .get_ifindex = ovs_netdev_get_ifindex, .send = netdev_send, }; diff --git a/net/openvswitch/vport-netdev.h b/net/openvswitch/vport-netdev.h index 6478079b3417..a3cb3a32cd77 100644 --- a/net/openvswitch/vport-netdev.h +++ b/net/openvswitch/vport-netdev.h @@ -40,6 +40,5 @@ netdev_vport_priv(const struct vport *vport) const char *ovs_netdev_get_name(const struct vport *); const char *ovs_netdev_get_config(const struct vport *); -int ovs_netdev_get_ifindex(const struct vport *); #endif /* vport_netdev.h */ diff --git a/net/openvswitch/vport.c b/net/openvswitch/vport.c index f6b8132ce4cb..720623190eaa 100644 --- a/net/openvswitch/vport.c +++ b/net/openvswitch/vport.c @@ -40,7 +40,7 @@ static const struct vport_ops *vport_ops_list[] = { &ovs_internal_vport_ops, }; -/* Protected by RCU read lock for reading, RTNL lock for writing. */ +/* Protected by RCU read lock for reading, ovs_mutex for writing. */ static struct hlist_head *dev_table; #define VPORT_HASH_BUCKETS 1024 @@ -80,7 +80,7 @@ static struct hlist_head *hash_bucket(struct net *net, const char *name) * * @name: name of port to find * - * Must be called with RTNL or RCU read lock. + * Must be called with ovs or RCU read lock. */ struct vport *ovs_vport_locate(struct net *net, const char *name) { @@ -128,7 +128,7 @@ struct vport *ovs_vport_alloc(int priv_size, const struct vport_ops *ops, vport->ops = ops; INIT_HLIST_NODE(&vport->dp_hash_node); - vport->percpu_stats = alloc_percpu(struct vport_percpu_stats); + vport->percpu_stats = alloc_percpu(struct pcpu_tstats); if (!vport->percpu_stats) { kfree(vport); return ERR_PTR(-ENOMEM); @@ -161,7 +161,7 @@ void ovs_vport_free(struct vport *vport) * @parms: Information about new vport. * * Creates a new vport with the specified configuration (which is dependent on - * device type). RTNL lock must be held. + * device type). ovs_mutex must be held. */ struct vport *ovs_vport_add(const struct vport_parms *parms) { @@ -169,8 +169,6 @@ struct vport *ovs_vport_add(const struct vport_parms *parms) int err = 0; int i; - ASSERT_RTNL(); - for (i = 0; i < ARRAY_SIZE(vport_ops_list); i++) { if (vport_ops_list[i]->type == parms->type) { struct hlist_head *bucket; @@ -201,12 +199,10 @@ out: * @port: New configuration. * * Modifies an existing device with the specified configuration (which is - * dependent on device type). RTNL lock must be held. + * dependent on device type). ovs_mutex must be held. */ int ovs_vport_set_options(struct vport *vport, struct nlattr *options) { - ASSERT_RTNL(); - if (!vport->ops->set_options) return -EOPNOTSUPP; return vport->ops->set_options(vport, options); @@ -218,11 +214,11 @@ int ovs_vport_set_options(struct vport *vport, struct nlattr *options) * @vport: vport to delete. * * Detaches @vport from its datapath and destroys it. It is possible to fail - * for reasons such as lack of memory. RTNL lock must be held. + * for reasons such as lack of memory. ovs_mutex must be held. */ void ovs_vport_del(struct vport *vport) { - ASSERT_RTNL(); + ASSERT_OVSL(); hlist_del_rcu(&vport->hash_node); @@ -237,7 +233,7 @@ void ovs_vport_del(struct vport *vport) * * Retrieves transmit, receive, and error stats for the given device. * - * Must be called with RTNL lock or rcu_read_lock. + * Must be called with ovs_mutex or rcu_read_lock. */ void ovs_vport_get_stats(struct vport *vport, struct ovs_vport_stats *stats) { @@ -264,16 +260,16 @@ void ovs_vport_get_stats(struct vport *vport, struct ovs_vport_stats *stats) spin_unlock_bh(&vport->stats_lock); for_each_possible_cpu(i) { - const struct vport_percpu_stats *percpu_stats; - struct vport_percpu_stats local_stats; + const struct pcpu_tstats *percpu_stats; + struct pcpu_tstats local_stats; unsigned int start; percpu_stats = per_cpu_ptr(vport->percpu_stats, i); do { - start = u64_stats_fetch_begin_bh(&percpu_stats->sync); + start = u64_stats_fetch_begin_bh(&percpu_stats->syncp); local_stats = *percpu_stats; - } while (u64_stats_fetch_retry_bh(&percpu_stats->sync, start)); + } while (u64_stats_fetch_retry_bh(&percpu_stats->syncp, start)); stats->rx_bytes += local_stats.rx_bytes; stats->rx_packets += local_stats.rx_packets; @@ -296,22 +292,24 @@ void ovs_vport_get_stats(struct vport *vport, struct ovs_vport_stats *stats) * negative error code if a real error occurred. If an error occurs, @skb is * left unmodified. * - * Must be called with RTNL lock or rcu_read_lock. + * Must be called with ovs_mutex or rcu_read_lock. */ int ovs_vport_get_options(const struct vport *vport, struct sk_buff *skb) { struct nlattr *nla; + int err; + + if (!vport->ops->get_options) + return 0; nla = nla_nest_start(skb, OVS_VPORT_ATTR_OPTIONS); if (!nla) return -EMSGSIZE; - if (vport->ops->get_options) { - int err = vport->ops->get_options(vport, skb); - if (err) { - nla_nest_cancel(skb, nla); - return err; - } + err = vport->ops->get_options(vport, skb); + if (err) { + nla_nest_cancel(skb, nla); + return err; } nla_nest_end(skb, nla); @@ -329,13 +327,13 @@ int ovs_vport_get_options(const struct vport *vport, struct sk_buff *skb) */ void ovs_vport_receive(struct vport *vport, struct sk_buff *skb) { - struct vport_percpu_stats *stats; + struct pcpu_tstats *stats; stats = this_cpu_ptr(vport->percpu_stats); - u64_stats_update_begin(&stats->sync); + u64_stats_update_begin(&stats->syncp); stats->rx_packets++; stats->rx_bytes += skb->len; - u64_stats_update_end(&stats->sync); + u64_stats_update_end(&stats->syncp); ovs_dp_process_received_packet(vport, skb); } @@ -346,7 +344,7 @@ void ovs_vport_receive(struct vport *vport, struct sk_buff *skb) * @vport: vport on which to send the packet * @skb: skb to send * - * Sends the given packet and returns the length of data sent. Either RTNL + * Sends the given packet and returns the length of data sent. Either ovs * lock or rcu_read_lock must be held. */ int ovs_vport_send(struct vport *vport, struct sk_buff *skb) @@ -354,14 +352,14 @@ int ovs_vport_send(struct vport *vport, struct sk_buff *skb) int sent = vport->ops->send(vport, skb); if (likely(sent)) { - struct vport_percpu_stats *stats; + struct pcpu_tstats *stats; stats = this_cpu_ptr(vport->percpu_stats); - u64_stats_update_begin(&stats->sync); + u64_stats_update_begin(&stats->syncp); stats->tx_packets++; stats->tx_bytes += sent; - u64_stats_update_end(&stats->sync); + u64_stats_update_end(&stats->syncp); } return sent; } diff --git a/net/openvswitch/vport.h b/net/openvswitch/vport.h index 3f7961ea3c56..68a377bc0841 100644 --- a/net/openvswitch/vport.h +++ b/net/openvswitch/vport.h @@ -19,6 +19,7 @@ #ifndef VPORT_H #define VPORT_H 1 +#include <linux/if_tunnel.h> #include <linux/list.h> #include <linux/netlink.h> #include <linux/openvswitch.h> @@ -50,14 +51,6 @@ int ovs_vport_send(struct vport *, struct sk_buff *); /* The following definitions are for implementers of vport devices: */ -struct vport_percpu_stats { - u64 rx_bytes; - u64 rx_packets; - u64 tx_bytes; - u64 tx_packets; - struct u64_stats_sync sync; -}; - struct vport_err_stats { u64 rx_dropped; u64 rx_errors; @@ -68,10 +61,10 @@ struct vport_err_stats { /** * struct vport - one port within a datapath * @rcu: RCU callback head for deferred destruction. - * @port_no: Index into @dp's @ports array. * @dp: Datapath to which this port belongs. * @upcall_portid: The Netlink port to use for packets received on this port that * miss the flow table. + * @port_no: Index into @dp's @ports array. * @hash_node: Element in @dev_table hash table in vport.c. * @dp_hash_node: Element in @datapath->ports hash table in datapath.c. * @ops: Class structure. @@ -81,15 +74,15 @@ struct vport_err_stats { */ struct vport { struct rcu_head rcu; - u16 port_no; struct datapath *dp; u32 upcall_portid; + u16 port_no; struct hlist_node hash_node; struct hlist_node dp_hash_node; const struct vport_ops *ops; - struct vport_percpu_stats __percpu *percpu_stats; + struct pcpu_tstats __percpu *percpu_stats; spinlock_t stats_lock; struct vport_err_stats err_stats; @@ -131,24 +124,22 @@ struct vport_parms { * have any configuration. * @get_name: Get the device's name. * @get_config: Get the device's configuration. - * @get_ifindex: Get the system interface index associated with the device. * May be null if the device does not have an ifindex. * @send: Send a packet on the device. Returns the length of the packet sent. */ struct vport_ops { enum ovs_vport_type type; - /* Called with RTNL lock. */ + /* Called with ovs_mutex. */ struct vport *(*create)(const struct vport_parms *); void (*destroy)(struct vport *); int (*set_options)(struct vport *, struct nlattr *); int (*get_options)(const struct vport *, struct sk_buff *); - /* Called with rcu_read_lock or RTNL lock. */ + /* Called with rcu_read_lock or ovs_mutex. */ const char *(*get_name)(const struct vport *); void (*get_config)(const struct vport *, void *); - int (*get_ifindex)(const struct vport *); int (*send)(struct vport *, struct sk_buff *); }; |