Commit 92cf1f23 authored by David S. Miller's avatar David S. Miller

Merge branch 'master' of git://git.kernel.org/pub/scm/linux/kernel/git/jesse/openvswitch

Jesse Gross says:

====================
A number of improvements for net-next/3.10.

Highlights include:

 * Properly exposing linux/openvswitch.h to userspace after the uapi
   changes.

 * Simplification of locking. It immediately makes things simpler to
   reason about and avoids holding RTNL mutex for longer than
   necessary. In the near future it will also enable tunnel
   registration and more fine-grained locking.

 * Miscellaneous cleanups and simplifications.
====================
Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
parents 98d2f0e6 e0f0ecf3
......@@ -19,436 +19,6 @@
#ifndef _LINUX_OPENVSWITCH_H
#define _LINUX_OPENVSWITCH_H 1
#include <linux/types.h>
/**
* struct ovs_header - header for OVS Generic Netlink messages.
* @dp_ifindex: ifindex of local port for datapath (0 to make a request not
* specific to a datapath).
*
* Attributes following the header are specific to a particular OVS Generic
* Netlink family, but all of the OVS families use this header.
*/
struct ovs_header {
int dp_ifindex;
};
/* Datapaths. */
#define OVS_DATAPATH_FAMILY "ovs_datapath"
#define OVS_DATAPATH_MCGROUP "ovs_datapath"
#define OVS_DATAPATH_VERSION 0x1
enum ovs_datapath_cmd {
OVS_DP_CMD_UNSPEC,
OVS_DP_CMD_NEW,
OVS_DP_CMD_DEL,
OVS_DP_CMD_GET,
OVS_DP_CMD_SET
};
/**
* enum ovs_datapath_attr - attributes for %OVS_DP_* commands.
* @OVS_DP_ATTR_NAME: Name of the network device that serves as the "local
* port". This is the name of the network device whose dp_ifindex is given in
* the &struct ovs_header. Always present in notifications. Required in
* %OVS_DP_NEW requests. May be used as an alternative to specifying
* dp_ifindex in other requests (with a dp_ifindex of 0).
* @OVS_DP_ATTR_UPCALL_PID: The Netlink socket in userspace that is initially
* set on the datapath port (for OVS_ACTION_ATTR_MISS). Only valid on
* %OVS_DP_CMD_NEW requests. A value of zero indicates that upcalls should
* not be sent.
* @OVS_DP_ATTR_STATS: Statistics about packets that have passed through the
* datapath. Always present in notifications.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_DP_* commands.
*/
enum ovs_datapath_attr {
OVS_DP_ATTR_UNSPEC,
OVS_DP_ATTR_NAME, /* name of dp_ifindex netdev */
OVS_DP_ATTR_UPCALL_PID, /* Netlink PID to receive upcalls */
OVS_DP_ATTR_STATS, /* struct ovs_dp_stats */
__OVS_DP_ATTR_MAX
};
#define OVS_DP_ATTR_MAX (__OVS_DP_ATTR_MAX - 1)
struct ovs_dp_stats {
__u64 n_hit; /* Number of flow table matches. */
__u64 n_missed; /* Number of flow table misses. */
__u64 n_lost; /* Number of misses not sent to userspace. */
__u64 n_flows; /* Number of flows present */
};
struct ovs_vport_stats {
__u64 rx_packets; /* total packets received */
__u64 tx_packets; /* total packets transmitted */
__u64 rx_bytes; /* total bytes received */
__u64 tx_bytes; /* total bytes transmitted */
__u64 rx_errors; /* bad packets received */
__u64 tx_errors; /* packet transmit problems */
__u64 rx_dropped; /* no space in linux buffers */
__u64 tx_dropped; /* no space available in linux */
};
/* Fixed logical ports. */
#define OVSP_LOCAL ((__u32)0)
/* Packet transfer. */
#define OVS_PACKET_FAMILY "ovs_packet"
#define OVS_PACKET_VERSION 0x1
enum ovs_packet_cmd {
OVS_PACKET_CMD_UNSPEC,
/* Kernel-to-user notifications. */
OVS_PACKET_CMD_MISS, /* Flow table miss. */
OVS_PACKET_CMD_ACTION, /* OVS_ACTION_ATTR_USERSPACE action. */
/* Userspace commands. */
OVS_PACKET_CMD_EXECUTE /* Apply actions to a packet. */
};
/**
* enum ovs_packet_attr - attributes for %OVS_PACKET_* commands.
* @OVS_PACKET_ATTR_PACKET: Present for all notifications. Contains the entire
* packet as received, from the start of the Ethernet header onward. For
* %OVS_PACKET_CMD_ACTION, %OVS_PACKET_ATTR_PACKET reflects changes made by
* actions preceding %OVS_ACTION_ATTR_USERSPACE, but %OVS_PACKET_ATTR_KEY is
* the flow key extracted from the packet as originally received.
* @OVS_PACKET_ATTR_KEY: Present for all notifications. Contains the flow key
* extracted from the packet as nested %OVS_KEY_ATTR_* attributes. This allows
* userspace to adapt its flow setup strategy by comparing its notion of the
* flow key against the kernel's.
* @OVS_PACKET_ATTR_ACTIONS: Contains actions for the packet. Used
* for %OVS_PACKET_CMD_EXECUTE. It has nested %OVS_ACTION_ATTR_* attributes.
* @OVS_PACKET_ATTR_USERDATA: Present for an %OVS_PACKET_CMD_ACTION
* notification if the %OVS_ACTION_ATTR_USERSPACE action specified an
* %OVS_USERSPACE_ATTR_USERDATA attribute, with the same length and content
* specified there.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_PACKET_* commands.
*/
enum ovs_packet_attr {
OVS_PACKET_ATTR_UNSPEC,
OVS_PACKET_ATTR_PACKET, /* Packet data. */
OVS_PACKET_ATTR_KEY, /* Nested OVS_KEY_ATTR_* attributes. */
OVS_PACKET_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
OVS_PACKET_ATTR_USERDATA, /* OVS_ACTION_ATTR_USERSPACE arg. */
__OVS_PACKET_ATTR_MAX
};
#define OVS_PACKET_ATTR_MAX (__OVS_PACKET_ATTR_MAX - 1)
/* Virtual ports. */
#define OVS_VPORT_FAMILY "ovs_vport"
#define OVS_VPORT_MCGROUP "ovs_vport"
#define OVS_VPORT_VERSION 0x1
enum ovs_vport_cmd {
OVS_VPORT_CMD_UNSPEC,
OVS_VPORT_CMD_NEW,
OVS_VPORT_CMD_DEL,
OVS_VPORT_CMD_GET,
OVS_VPORT_CMD_SET
};
enum ovs_vport_type {
OVS_VPORT_TYPE_UNSPEC,
OVS_VPORT_TYPE_NETDEV, /* network device */
OVS_VPORT_TYPE_INTERNAL, /* network device implemented by datapath */
__OVS_VPORT_TYPE_MAX
};
#define OVS_VPORT_TYPE_MAX (__OVS_VPORT_TYPE_MAX - 1)
/**
* enum ovs_vport_attr - attributes for %OVS_VPORT_* commands.
* @OVS_VPORT_ATTR_PORT_NO: 32-bit port number within datapath.
* @OVS_VPORT_ATTR_TYPE: 32-bit %OVS_VPORT_TYPE_* constant describing the type
* of vport.
* @OVS_VPORT_ATTR_NAME: Name of vport. For a vport based on a network device
* this is the name of the network device. Maximum length %IFNAMSIZ-1 bytes
* plus a null terminator.
* @OVS_VPORT_ATTR_OPTIONS: Vport-specific configuration information.
* @OVS_VPORT_ATTR_UPCALL_PID: The Netlink socket in userspace that
* OVS_PACKET_CMD_MISS upcalls will be directed to for packets received on
* this port. A value of zero indicates that upcalls should not be sent.
* @OVS_VPORT_ATTR_STATS: A &struct ovs_vport_stats giving statistics for
* packets sent or received through the vport.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_VPORT_* commands.
*
* For %OVS_VPORT_CMD_NEW requests, the %OVS_VPORT_ATTR_TYPE and
* %OVS_VPORT_ATTR_NAME attributes are required. %OVS_VPORT_ATTR_PORT_NO is
* optional; if not specified a free port number is automatically selected.
* Whether %OVS_VPORT_ATTR_OPTIONS is required or optional depends on the type
* of vport.
* and other attributes are ignored.
*
* For other requests, if %OVS_VPORT_ATTR_NAME is specified then it is used to
* look up the vport to operate on; otherwise dp_idx from the &struct
* ovs_header plus %OVS_VPORT_ATTR_PORT_NO determine the vport.
*/
enum ovs_vport_attr {
OVS_VPORT_ATTR_UNSPEC,
OVS_VPORT_ATTR_PORT_NO, /* u32 port number within datapath */
OVS_VPORT_ATTR_TYPE, /* u32 OVS_VPORT_TYPE_* constant. */
OVS_VPORT_ATTR_NAME, /* string name, up to IFNAMSIZ bytes long */
OVS_VPORT_ATTR_OPTIONS, /* nested attributes, varies by vport type */
OVS_VPORT_ATTR_UPCALL_PID, /* u32 Netlink PID to receive upcalls */
OVS_VPORT_ATTR_STATS, /* struct ovs_vport_stats */
__OVS_VPORT_ATTR_MAX
};
#define OVS_VPORT_ATTR_MAX (__OVS_VPORT_ATTR_MAX - 1)
/* Flows. */
#define OVS_FLOW_FAMILY "ovs_flow"
#define OVS_FLOW_MCGROUP "ovs_flow"
#define OVS_FLOW_VERSION 0x1
enum ovs_flow_cmd {
OVS_FLOW_CMD_UNSPEC,
OVS_FLOW_CMD_NEW,
OVS_FLOW_CMD_DEL,
OVS_FLOW_CMD_GET,
OVS_FLOW_CMD_SET
};
struct ovs_flow_stats {
__u64 n_packets; /* Number of matched packets. */
__u64 n_bytes; /* Number of matched bytes. */
};
enum ovs_key_attr {
OVS_KEY_ATTR_UNSPEC,
OVS_KEY_ATTR_ENCAP, /* Nested set of encapsulated attributes. */
OVS_KEY_ATTR_PRIORITY, /* u32 skb->priority */
OVS_KEY_ATTR_IN_PORT, /* u32 OVS dp port number */
OVS_KEY_ATTR_ETHERNET, /* struct ovs_key_ethernet */
OVS_KEY_ATTR_VLAN, /* be16 VLAN TCI */
OVS_KEY_ATTR_ETHERTYPE, /* be16 Ethernet type */
OVS_KEY_ATTR_IPV4, /* struct ovs_key_ipv4 */
OVS_KEY_ATTR_IPV6, /* struct ovs_key_ipv6 */
OVS_KEY_ATTR_TCP, /* struct ovs_key_tcp */
OVS_KEY_ATTR_UDP, /* struct ovs_key_udp */
OVS_KEY_ATTR_ICMP, /* struct ovs_key_icmp */
OVS_KEY_ATTR_ICMPV6, /* struct ovs_key_icmpv6 */
OVS_KEY_ATTR_ARP, /* struct ovs_key_arp */
OVS_KEY_ATTR_ND, /* struct ovs_key_nd */
OVS_KEY_ATTR_SKB_MARK, /* u32 skb mark */
__OVS_KEY_ATTR_MAX
};
#define OVS_KEY_ATTR_MAX (__OVS_KEY_ATTR_MAX - 1)
/**
* enum ovs_frag_type - IPv4 and IPv6 fragment type
* @OVS_FRAG_TYPE_NONE: Packet is not a fragment.
* @OVS_FRAG_TYPE_FIRST: Packet is a fragment with offset 0.
* @OVS_FRAG_TYPE_LATER: Packet is a fragment with nonzero offset.
*
* Used as the @ipv4_frag in &struct ovs_key_ipv4 and as @ipv6_frag &struct
* ovs_key_ipv6.
*/
enum ovs_frag_type {
OVS_FRAG_TYPE_NONE,
OVS_FRAG_TYPE_FIRST,
OVS_FRAG_TYPE_LATER,
__OVS_FRAG_TYPE_MAX
};
#define OVS_FRAG_TYPE_MAX (__OVS_FRAG_TYPE_MAX - 1)
struct ovs_key_ethernet {
__u8 eth_src[6];
__u8 eth_dst[6];
};
struct ovs_key_ipv4 {
__be32 ipv4_src;
__be32 ipv4_dst;
__u8 ipv4_proto;
__u8 ipv4_tos;
__u8 ipv4_ttl;
__u8 ipv4_frag; /* One of OVS_FRAG_TYPE_*. */
};
struct ovs_key_ipv6 {
__be32 ipv6_src[4];
__be32 ipv6_dst[4];
__be32 ipv6_label; /* 20-bits in least-significant bits. */
__u8 ipv6_proto;
__u8 ipv6_tclass;
__u8 ipv6_hlimit;
__u8 ipv6_frag; /* One of OVS_FRAG_TYPE_*. */
};
struct ovs_key_tcp {
__be16 tcp_src;
__be16 tcp_dst;
};
struct ovs_key_udp {
__be16 udp_src;
__be16 udp_dst;
};
struct ovs_key_icmp {
__u8 icmp_type;
__u8 icmp_code;
};
struct ovs_key_icmpv6 {
__u8 icmpv6_type;
__u8 icmpv6_code;
};
struct ovs_key_arp {
__be32 arp_sip;
__be32 arp_tip;
__be16 arp_op;
__u8 arp_sha[6];
__u8 arp_tha[6];
};
struct ovs_key_nd {
__u32 nd_target[4];
__u8 nd_sll[6];
__u8 nd_tll[6];
};
/**
* enum ovs_flow_attr - attributes for %OVS_FLOW_* commands.
* @OVS_FLOW_ATTR_KEY: Nested %OVS_KEY_ATTR_* attributes specifying the flow
* key. Always present in notifications. Required for all requests (except
* dumps).
* @OVS_FLOW_ATTR_ACTIONS: Nested %OVS_ACTION_ATTR_* attributes specifying
* the actions to take for packets that match the key. Always present in
* notifications. Required for %OVS_FLOW_CMD_NEW requests, optional for
* %OVS_FLOW_CMD_SET requests.
* @OVS_FLOW_ATTR_STATS: &struct ovs_flow_stats giving statistics for this
* flow. Present in notifications if the stats would be nonzero. Ignored in
* requests.
* @OVS_FLOW_ATTR_TCP_FLAGS: An 8-bit value giving the OR'd value of all of the
* TCP flags seen on packets in this flow. Only present in notifications for
* TCP flows, and only if it would be nonzero. Ignored in requests.
* @OVS_FLOW_ATTR_USED: A 64-bit integer giving the time, in milliseconds on
* the system monotonic clock, at which a packet was last processed for this
* flow. Only present in notifications if a packet has been processed for this
* flow. Ignored in requests.
* @OVS_FLOW_ATTR_CLEAR: If present in a %OVS_FLOW_CMD_SET request, clears the
* last-used time, accumulated TCP flags, and statistics for this flow.
* Otherwise ignored in requests. Never present in notifications.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_FLOW_* commands.
*/
enum ovs_flow_attr {
OVS_FLOW_ATTR_UNSPEC,
OVS_FLOW_ATTR_KEY, /* Sequence of OVS_KEY_ATTR_* attributes. */
OVS_FLOW_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
OVS_FLOW_ATTR_STATS, /* struct ovs_flow_stats. */
OVS_FLOW_ATTR_TCP_FLAGS, /* 8-bit OR'd TCP flags. */
OVS_FLOW_ATTR_USED, /* u64 msecs last used in monotonic time. */
OVS_FLOW_ATTR_CLEAR, /* Flag to clear stats, tcp_flags, used. */
__OVS_FLOW_ATTR_MAX
};
#define OVS_FLOW_ATTR_MAX (__OVS_FLOW_ATTR_MAX - 1)
/**
* enum ovs_sample_attr - Attributes for %OVS_ACTION_ATTR_SAMPLE action.
* @OVS_SAMPLE_ATTR_PROBABILITY: 32-bit fraction of packets to sample with
* @OVS_ACTION_ATTR_SAMPLE. A value of 0 samples no packets, a value of
* %UINT32_MAX samples all packets and intermediate values sample intermediate
* fractions of packets.
* @OVS_SAMPLE_ATTR_ACTIONS: Set of actions to execute in sampling event.
* Actions are passed as nested attributes.
*
* Executes the specified actions with the given probability on a per-packet
* basis.
*/
enum ovs_sample_attr {
OVS_SAMPLE_ATTR_UNSPEC,
OVS_SAMPLE_ATTR_PROBABILITY, /* u32 number */
OVS_SAMPLE_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
__OVS_SAMPLE_ATTR_MAX,
};
#define OVS_SAMPLE_ATTR_MAX (__OVS_SAMPLE_ATTR_MAX - 1)
/**
* enum ovs_userspace_attr - Attributes for %OVS_ACTION_ATTR_USERSPACE action.
* @OVS_USERSPACE_ATTR_PID: u32 Netlink PID to which the %OVS_PACKET_CMD_ACTION
* message should be sent. Required.
* @OVS_USERSPACE_ATTR_USERDATA: If present, its variable-length argument is
* copied to the %OVS_PACKET_CMD_ACTION message as %OVS_PACKET_ATTR_USERDATA.
*/
enum ovs_userspace_attr {
OVS_USERSPACE_ATTR_UNSPEC,
OVS_USERSPACE_ATTR_PID, /* u32 Netlink PID to receive upcalls. */
OVS_USERSPACE_ATTR_USERDATA, /* Optional user-specified cookie. */
__OVS_USERSPACE_ATTR_MAX
};
#define OVS_USERSPACE_ATTR_MAX (__OVS_USERSPACE_ATTR_MAX - 1)
/**
* struct ovs_action_push_vlan - %OVS_ACTION_ATTR_PUSH_VLAN action argument.
* @vlan_tpid: Tag protocol identifier (TPID) to push.
* @vlan_tci: Tag control identifier (TCI) to push. The CFI bit must be set
* (but it will not be set in the 802.1Q header that is pushed).
*
* The @vlan_tpid value is typically %ETH_P_8021Q. The only acceptable TPID
* values are those that the kernel module also parses as 802.1Q headers, to
* prevent %OVS_ACTION_ATTR_PUSH_VLAN followed by %OVS_ACTION_ATTR_POP_VLAN
* from having surprising results.
*/
struct ovs_action_push_vlan {
__be16 vlan_tpid; /* 802.1Q TPID. */
__be16 vlan_tci; /* 802.1Q TCI (VLAN ID and priority). */
};
/**
* enum ovs_action_attr - Action types.
*
* @OVS_ACTION_ATTR_OUTPUT: Output packet to port.
* @OVS_ACTION_ATTR_USERSPACE: Send packet to userspace according to nested
* %OVS_USERSPACE_ATTR_* attributes.
* @OVS_ACTION_ATTR_SET: Replaces the contents of an existing header. The
* single nested %OVS_KEY_ATTR_* attribute specifies a header to modify and its
* value.
* @OVS_ACTION_ATTR_PUSH_VLAN: Push a new outermost 802.1Q header onto the
* packet.
* @OVS_ACTION_ATTR_POP_VLAN: Pop the outermost 802.1Q header off the packet.
* @OVS_ACTION_ATTR_SAMPLE: Probabilitically executes actions, as specified in
* the nested %OVS_SAMPLE_ATTR_* attributes.
*
* Only a single header can be set with a single %OVS_ACTION_ATTR_SET. Not all
* fields within a header are modifiable, e.g. the IPv4 protocol and fragment
* type may not be changed.
*/
enum ovs_action_attr {
OVS_ACTION_ATTR_UNSPEC,
OVS_ACTION_ATTR_OUTPUT, /* u32 port number. */
OVS_ACTION_ATTR_USERSPACE, /* Nested OVS_USERSPACE_ATTR_*. */
OVS_ACTION_ATTR_SET, /* One nested OVS_KEY_ATTR_*. */
OVS_ACTION_ATTR_PUSH_VLAN, /* struct ovs_action_push_vlan. */
OVS_ACTION_ATTR_POP_VLAN, /* No argument. */
OVS_ACTION_ATTR_SAMPLE, /* Nested OVS_SAMPLE_ATTR_*. */
__OVS_ACTION_ATTR_MAX
};
#define OVS_ACTION_ATTR_MAX (__OVS_ACTION_ATTR_MAX - 1)
#include <uapi/linux/openvswitch.h>
#endif /* _LINUX_OPENVSWITCH_H */
......@@ -285,6 +285,7 @@ header-y += nvram.h
header-y += omap3isp.h
header-y += omapfb.h
header-y += oom.h
header-y += openvswitch.h
header-y += packet_diag.h
header-y += param.h
header-y += parport.h
......
/*
* Copyright (c) 2007-2011 Nicira Networks.
*
* 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
*/
#ifndef _UAPI__LINUX_OPENVSWITCH_H
#define _UAPI__LINUX_OPENVSWITCH_H 1
#include <linux/types.h>
#include <linux/if_ether.h>
/**
* struct ovs_header - header for OVS Generic Netlink messages.
* @dp_ifindex: ifindex of local port for datapath (0 to make a request not
* specific to a datapath).
*
* Attributes following the header are specific to a particular OVS Generic
* Netlink family, but all of the OVS families use this header.
*/
struct ovs_header {
int dp_ifindex;
};
/* Datapaths. */
#define OVS_DATAPATH_FAMILY "ovs_datapath"
#define OVS_DATAPATH_MCGROUP "ovs_datapath"
#define OVS_DATAPATH_VERSION 0x1
enum ovs_datapath_cmd {
OVS_DP_CMD_UNSPEC,
OVS_DP_CMD_NEW,
OVS_DP_CMD_DEL,
OVS_DP_CMD_GET,
OVS_DP_CMD_SET
};
/**
* enum ovs_datapath_attr - attributes for %OVS_DP_* commands.
* @OVS_DP_ATTR_NAME: Name of the network device that serves as the "local
* port". This is the name of the network device whose dp_ifindex is given in
* the &struct ovs_header. Always present in notifications. Required in
* %OVS_DP_NEW requests. May be used as an alternative to specifying
* dp_ifindex in other requests (with a dp_ifindex of 0).
* @OVS_DP_ATTR_UPCALL_PID: The Netlink socket in userspace that is initially
* set on the datapath port (for OVS_ACTION_ATTR_MISS). Only valid on
* %OVS_DP_CMD_NEW requests. A value of zero indicates that upcalls should
* not be sent.
* @OVS_DP_ATTR_STATS: Statistics about packets that have passed through the
* datapath. Always present in notifications.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_DP_* commands.
*/
enum ovs_datapath_attr {
OVS_DP_ATTR_UNSPEC,
OVS_DP_ATTR_NAME, /* name of dp_ifindex netdev */
OVS_DP_ATTR_UPCALL_PID, /* Netlink PID to receive upcalls */
OVS_DP_ATTR_STATS, /* struct ovs_dp_stats */
__OVS_DP_ATTR_MAX
};
#define OVS_DP_ATTR_MAX (__OVS_DP_ATTR_MAX - 1)
struct ovs_dp_stats {
__u64 n_hit; /* Number of flow table matches. */
__u64 n_missed; /* Number of flow table misses. */
__u64 n_lost; /* Number of misses not sent to userspace. */
__u64 n_flows; /* Number of flows present */
};
struct ovs_vport_stats {
__u64 rx_packets; /* total packets received */
__u64 tx_packets; /* total packets transmitted */
__u64 rx_bytes; /* total bytes received */
__u64 tx_bytes; /* total bytes transmitted */
__u64 rx_errors; /* bad packets received */
__u64 tx_errors; /* packet transmit problems */
__u64 rx_dropped; /* no space in linux buffers */
__u64 tx_dropped; /* no space available in linux */
};
/* Fixed logical ports. */
#define OVSP_LOCAL ((__u32)0)
/* Packet transfer. */
#define OVS_PACKET_FAMILY "ovs_packet"
#define OVS_PACKET_VERSION 0x1
enum ovs_packet_cmd {
OVS_PACKET_CMD_UNSPEC,
/* Kernel-to-user notifications. */
OVS_PACKET_CMD_MISS, /* Flow table miss. */
OVS_PACKET_CMD_ACTION, /* OVS_ACTION_ATTR_USERSPACE action. */
/* Userspace commands. */
OVS_PACKET_CMD_EXECUTE /* Apply actions to a packet. */
};
/**
* enum ovs_packet_attr - attributes for %OVS_PACKET_* commands.
* @OVS_PACKET_ATTR_PACKET: Present for all notifications. Contains the entire
* packet as received, from the start of the Ethernet header onward. For
* %OVS_PACKET_CMD_ACTION, %OVS_PACKET_ATTR_PACKET reflects changes made by
* actions preceding %OVS_ACTION_ATTR_USERSPACE, but %OVS_PACKET_ATTR_KEY is
* the flow key extracted from the packet as originally received.
* @OVS_PACKET_ATTR_KEY: Present for all notifications. Contains the flow key
* extracted from the packet as nested %OVS_KEY_ATTR_* attributes. This allows
* userspace to adapt its flow setup strategy by comparing its notion of the
* flow key against the kernel's.
* @OVS_PACKET_ATTR_ACTIONS: Contains actions for the packet. Used
* for %OVS_PACKET_CMD_EXECUTE. It has nested %OVS_ACTION_ATTR_* attributes.
* @OVS_PACKET_ATTR_USERDATA: Present for an %OVS_PACKET_CMD_ACTION
* notification if the %OVS_ACTION_ATTR_USERSPACE action specified an
* %OVS_USERSPACE_ATTR_USERDATA attribute, with the same length and content
* specified there.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_PACKET_* commands.
*/
enum ovs_packet_attr {
OVS_PACKET_ATTR_UNSPEC,
OVS_PACKET_ATTR_PACKET, /* Packet data. */
OVS_PACKET_ATTR_KEY, /* Nested OVS_KEY_ATTR_* attributes. */
OVS_PACKET_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
OVS_PACKET_ATTR_USERDATA, /* OVS_ACTION_ATTR_USERSPACE arg. */
__OVS_PACKET_ATTR_MAX
};
#define OVS_PACKET_ATTR_MAX (__OVS_PACKET_ATTR_MAX - 1)
/* Virtual ports. */
#define OVS_VPORT_FAMILY "ovs_vport"
#define OVS_VPORT_MCGROUP "ovs_vport"
#define OVS_VPORT_VERSION 0x1
enum ovs_vport_cmd {
OVS_VPORT_CMD_UNSPEC,
OVS_VPORT_CMD_NEW,
OVS_VPORT_CMD_DEL,
OVS_VPORT_CMD_GET,
OVS_VPORT_CMD_SET
};
enum ovs_vport_type {
OVS_VPORT_TYPE_UNSPEC,
OVS_VPORT_TYPE_NETDEV, /* network device */
OVS_VPORT_TYPE_INTERNAL, /* network device implemented by datapath */
__OVS_VPORT_TYPE_MAX
};
#define OVS_VPORT_TYPE_MAX (__OVS_VPORT_TYPE_MAX - 1)
/**
* enum ovs_vport_attr - attributes for %OVS_VPORT_* commands.
* @OVS_VPORT_ATTR_PORT_NO: 32-bit port number within datapath.
* @OVS_VPORT_ATTR_TYPE: 32-bit %OVS_VPORT_TYPE_* constant describing the type
* of vport.
* @OVS_VPORT_ATTR_NAME: Name of vport. For a vport based on a network device
* this is the name of the network device. Maximum length %IFNAMSIZ-1 bytes
* plus a null terminator.
* @OVS_VPORT_ATTR_OPTIONS: Vport-specific configuration information.
* @OVS_VPORT_ATTR_UPCALL_PID: The Netlink socket in userspace that
* OVS_PACKET_CMD_MISS upcalls will be directed to for packets received on
* this port. A value of zero indicates that upcalls should not be sent.
* @OVS_VPORT_ATTR_STATS: A &struct ovs_vport_stats giving statistics for
* packets sent or received through the vport.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_VPORT_* commands.
*
* For %OVS_VPORT_CMD_NEW requests, the %OVS_VPORT_ATTR_TYPE and
* %OVS_VPORT_ATTR_NAME attributes are required. %OVS_VPORT_ATTR_PORT_NO is
* optional; if not specified a free port number is automatically selected.
* Whether %OVS_VPORT_ATTR_OPTIONS is required or optional depends on the type
* of vport.
* and other attributes are ignored.
*
* For other requests, if %OVS_VPORT_ATTR_NAME is specified then it is used to
* look up the vport to operate on; otherwise dp_idx from the &struct
* ovs_header plus %OVS_VPORT_ATTR_PORT_NO determine the vport.
*/
enum ovs_vport_attr {
OVS_VPORT_ATTR_UNSPEC,
OVS_VPORT_ATTR_PORT_NO, /* u32 port number within datapath */
OVS_VPORT_ATTR_TYPE, /* u32 OVS_VPORT_TYPE_* constant. */
OVS_VPORT_ATTR_NAME, /* string name, up to IFNAMSIZ bytes long */
OVS_VPORT_ATTR_OPTIONS, /* nested attributes, varies by vport type */
OVS_VPORT_ATTR_UPCALL_PID, /* u32 Netlink PID to receive upcalls */
OVS_VPORT_ATTR_STATS, /* struct ovs_vport_stats */
__OVS_VPORT_ATTR_MAX
};
#define OVS_VPORT_ATTR_MAX (__OVS_VPORT_ATTR_MAX - 1)
/* Flows. */
#define OVS_FLOW_FAMILY "ovs_flow"
#define OVS_FLOW_MCGROUP "ovs_flow"
#define OVS_FLOW_VERSION 0x1
enum ovs_flow_cmd {
OVS_FLOW_CMD_UNSPEC,
OVS_FLOW_CMD_NEW,
OVS_FLOW_CMD_DEL,
OVS_FLOW_CMD_GET,
OVS_FLOW_CMD_SET
};
struct ovs_flow_stats {
__u64 n_packets; /* Number of matched packets. */
__u64 n_bytes; /* Number of matched bytes. */
};
enum ovs_key_attr {
OVS_KEY_ATTR_UNSPEC,
OVS_KEY_ATTR_ENCAP, /* Nested set of encapsulated attributes. */
OVS_KEY_ATTR_PRIORITY, /* u32 skb->priority */
OVS_KEY_ATTR_IN_PORT, /* u32 OVS dp port number */
OVS_KEY_ATTR_ETHERNET, /* struct ovs_key_ethernet */
OVS_KEY_ATTR_VLAN, /* be16 VLAN TCI */
OVS_KEY_ATTR_ETHERTYPE, /* be16 Ethernet type */
OVS_KEY_ATTR_IPV4, /* struct ovs_key_ipv4 */
OVS_KEY_ATTR_IPV6, /* struct ovs_key_ipv6 */
OVS_KEY_ATTR_TCP, /* struct ovs_key_tcp */
OVS_KEY_ATTR_UDP, /* struct ovs_key_udp */
OVS_KEY_ATTR_ICMP, /* struct ovs_key_icmp */
OVS_KEY_ATTR_ICMPV6, /* struct ovs_key_icmpv6 */
OVS_KEY_ATTR_ARP, /* struct ovs_key_arp */
OVS_KEY_ATTR_ND, /* struct ovs_key_nd */
OVS_KEY_ATTR_SKB_MARK, /* u32 skb mark */
__OVS_KEY_ATTR_MAX
};
#define OVS_KEY_ATTR_MAX (__OVS_KEY_ATTR_MAX - 1)
/**
* enum ovs_frag_type - IPv4 and IPv6 fragment type
* @OVS_FRAG_TYPE_NONE: Packet is not a fragment.
* @OVS_FRAG_TYPE_FIRST: Packet is a fragment with offset 0.
* @OVS_FRAG_TYPE_LATER: Packet is a fragment with nonzero offset.
*
* Used as the @ipv4_frag in &struct ovs_key_ipv4 and as @ipv6_frag &struct
* ovs_key_ipv6.
*/
enum ovs_frag_type {
OVS_FRAG_TYPE_NONE,
OVS_FRAG_TYPE_FIRST,
OVS_FRAG_TYPE_LATER,
__OVS_FRAG_TYPE_MAX
};
#define OVS_FRAG_TYPE_MAX (__OVS_FRAG_TYPE_MAX - 1)
struct ovs_key_ethernet {
__u8 eth_src[ETH_ALEN];
__u8 eth_dst[ETH_ALEN];
};
struct ovs_key_ipv4 {
__be32 ipv4_src;
__be32 ipv4_dst;
__u8 ipv4_proto;
__u8 ipv4_tos;
__u8 ipv4_ttl;
__u8 ipv4_frag; /* One of OVS_FRAG_TYPE_*. */
};
struct ovs_key_ipv6 {
__be32 ipv6_src[4];
__be32 ipv6_dst[4];
__be32 ipv6_label; /* 20-bits in least-significant bits. */
__u8 ipv6_proto;
__u8 ipv6_tclass;
__u8 ipv6_hlimit;
__u8 ipv6_frag; /* One of OVS_FRAG_TYPE_*. */
};
struct ovs_key_tcp {
__be16 tcp_src;
__be16 tcp_dst;
};
struct ovs_key_udp {
__be16 udp_src;
__be16 udp_dst;
};
struct ovs_key_icmp {
__u8 icmp_type;
__u8 icmp_code;
};
struct ovs_key_icmpv6 {
__u8 icmpv6_type;
__u8 icmpv6_code;
};
struct ovs_key_arp {
__be32 arp_sip;
__be32 arp_tip;
__be16 arp_op;
__u8 arp_sha[ETH_ALEN];
__u8 arp_tha[ETH_ALEN];
};
struct ovs_key_nd {
__u32 nd_target[4];
__u8 nd_sll[ETH_ALEN];
__u8 nd_tll[ETH_ALEN];
};
/**
* enum ovs_flow_attr - attributes for %OVS_FLOW_* commands.
* @OVS_FLOW_ATTR_KEY: Nested %OVS_KEY_ATTR_* attributes specifying the flow
* key. Always present in notifications. Required for all requests (except
* dumps).
* @OVS_FLOW_ATTR_ACTIONS: Nested %OVS_ACTION_ATTR_* attributes specifying
* the actions to take for packets that match the key. Always present in
* notifications. Required for %OVS_FLOW_CMD_NEW requests, optional for
* %OVS_FLOW_CMD_SET requests.
* @OVS_FLOW_ATTR_STATS: &struct ovs_flow_stats giving statistics for this
* flow. Present in notifications if the stats would be nonzero. Ignored in
* requests.
* @OVS_FLOW_ATTR_TCP_FLAGS: An 8-bit value giving the OR'd value of all of the
* TCP flags seen on packets in this flow. Only present in notifications for
* TCP flows, and only if it would be nonzero. Ignored in requests.
* @OVS_FLOW_ATTR_USED: A 64-bit integer giving the time, in milliseconds on
* the system monotonic clock, at which a packet was last processed for this
* flow. Only present in notifications if a packet has been processed for this
* flow. Ignored in requests.
* @OVS_FLOW_ATTR_CLEAR: If present in a %OVS_FLOW_CMD_SET request, clears the
* last-used time, accumulated TCP flags, and statistics for this flow.
* Otherwise ignored in requests. Never present in notifications.
*
* These attributes follow the &struct ovs_header within the Generic Netlink
* payload for %OVS_FLOW_* commands.
*/
enum ovs_flow_attr {
OVS_FLOW_ATTR_UNSPEC,
OVS_FLOW_ATTR_KEY, /* Sequence of OVS_KEY_ATTR_* attributes. */
OVS_FLOW_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
OVS_FLOW_ATTR_STATS, /* struct ovs_flow_stats. */
OVS_FLOW_ATTR_TCP_FLAGS, /* 8-bit OR'd TCP flags. */
OVS_FLOW_ATTR_USED, /* u64 msecs last used in monotonic time. */
OVS_FLOW_ATTR_CLEAR, /* Flag to clear stats, tcp_flags, used. */
__OVS_FLOW_ATTR_MAX
};
#define OVS_FLOW_ATTR_MAX (__OVS_FLOW_ATTR_MAX - 1)
/**
* enum ovs_sample_attr - Attributes for %OVS_ACTION_ATTR_SAMPLE action.
* @OVS_SAMPLE_ATTR_PROBABILITY: 32-bit fraction of packets to sample with
* @OVS_ACTION_ATTR_SAMPLE. A value of 0 samples no packets, a value of
* %UINT32_MAX samples all packets and intermediate values sample intermediate
* fractions of packets.
* @OVS_SAMPLE_ATTR_ACTIONS: Set of actions to execute in sampling event.
* Actions are passed as nested attributes.
*
* Executes the specified actions with the given probability on a per-packet
* basis.
*/
enum ovs_sample_attr {
OVS_SAMPLE_ATTR_UNSPEC,
OVS_SAMPLE_ATTR_PROBABILITY, /* u32 number */
OVS_SAMPLE_ATTR_ACTIONS, /* Nested OVS_ACTION_ATTR_* attributes. */
__OVS_SAMPLE_ATTR_MAX,
};
#define OVS_SAMPLE_ATTR_MAX (__OVS_SAMPLE_ATTR_MAX - 1)
/**
* enum ovs_userspace_attr - Attributes for %OVS_ACTION_ATTR_USERSPACE action.
* @OVS_USERSPACE_ATTR_PID: u32 Netlink PID to which the %OVS_PACKET_CMD_ACTION
* message should be sent. Required.
* @OVS_USERSPACE_ATTR_USERDATA: If present, its variable-length argument is
* copied to the %OVS_PACKET_CMD_ACTION message as %OVS_PACKET_ATTR_USERDATA.
*/
enum ovs_userspace_attr {
OVS_USERSPACE_ATTR_UNSPEC,
OVS_USERSPACE_ATTR_PID, /* u32 Netlink PID to receive upcalls. */
OVS_USERSPACE_ATTR_USERDATA, /* Optional user-specified cookie. */
__OVS_USERSPACE_ATTR_MAX
};
#define OVS_USERSPACE_ATTR_MAX (__OVS_USERSPACE_ATTR_MAX - 1)
/**
* struct ovs_action_push_vlan - %OVS_ACTION_ATTR_PUSH_VLAN action argument.
* @vlan_tpid: Tag protocol identifier (TPID) to push.
* @vlan_tci: Tag control identifier (TCI) to push. The CFI bit must be set
* (but it will not be set in the 802.1Q header that is pushed).
*
* The @vlan_tpid value is typically %ETH_P_8021Q. The only acceptable TPID
* values are those that the kernel module also parses as 802.1Q headers, to
* prevent %OVS_ACTION_ATTR_PUSH_VLAN followed by %OVS_ACTION_ATTR_POP_VLAN
* from having surprising results.
*/
struct ovs_action_push_vlan {
__be16 vlan_tpid; /* 802.1Q TPID. */
__be16 vlan_tci; /* 802.1Q TCI (VLAN ID and priority). */
};
/**
* enum ovs_action_attr - Action types.
*
* @OVS_ACTION_ATTR_OUTPUT: Output packet to port.
* @OVS_ACTION_ATTR_USERSPACE: Send packet to userspace according to nested
* %OVS_USERSPACE_ATTR_* attributes.
* @OVS_ACTION_ATTR_SET: Replaces the contents of an existing header. The
* single nested %OVS_KEY_ATTR_* attribute specifies a header to modify and its
* value.
* @OVS_ACTION_ATTR_PUSH_VLAN: Push a new outermost 802.1Q header onto the
* packet.
* @OVS_ACTION_ATTR_POP_VLAN: Pop the outermost 802.1Q header off the packet.
* @OVS_ACTION_ATTR_SAMPLE: Probabilitically executes actions, as specified in
* the nested %OVS_SAMPLE_ATTR_* attributes.
*
* Only a single header can be set with a single %OVS_ACTION_ATTR_SET. Not all
* fields within a header are modifiable, e.g. the IPv4 protocol and fragment
* type may not be changed.
*/
enum ovs_action_attr {
OVS_ACTION_ATTR_UNSPEC,
OVS_ACTION_ATTR_OUTPUT, /* u32 port number. */
OVS_ACTION_ATTR_USERSPACE, /* Nested OVS_USERSPACE_ATTR_*. */
OVS_ACTION_ATTR_SET, /* One nested OVS_KEY_ATTR_*. */
OVS_ACTION_ATTR_PUSH_VLAN, /* struct ovs_action_push_vlan. */
OVS_ACTION_ATTR_POP_VLAN, /* No argument. */
OVS_ACTION_ATTR_SAMPLE, /* Nested OVS_SAMPLE_ATTR_*. */
__OVS_ACTION_ATTR_MAX
};
#define OVS_ACTION_ATTR_MAX (__OVS_ACTION_ATTR_MAX - 1)
#endif /* _LINUX_OPENVSWITCH_H */
......@@ -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>
......@@ -56,38 +57,59 @@
#include "flow.h"
#include "vport-internal_dev.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 +117,7 @@ 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 +135,10 @@ 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);
}
......@@ -168,7 +190,7 @@ 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 +202,12 @@ 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);
......@@ -337,6 +357,35 @@ 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 +394,6 @@ 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)) {
......@@ -366,13 +414,7 @@ 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->userdata)
len += NLA_ALIGN(upcall_info->userdata->nla_len);
user_skb = genlmsg_new(len, GFP_ATOMIC);
user_skb = genlmsg_new(upcall_msg_size(skb, upcall_info->userdata), GFP_ATOMIC);
if (!user_skb) {
err = -ENOMEM;
goto out;
......@@ -403,13 +445,13 @@ static int queue_userspace_packet(struct net *net, int dp_ifindex,
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;
......@@ -662,8 +704,7 @@ 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]);
......@@ -673,7 +714,7 @@ 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);
......@@ -744,7 +785,7 @@ static int ovs_packet_cmd_execute(struct sk_buff *skb, struct genl_info *info)
}
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 },
};
......@@ -760,7 +801,7 @@ 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);
......@@ -802,7 +843,17 @@ 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)
......@@ -816,8 +867,7 @@ 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)
......@@ -880,25 +930,10 @@ static int ovs_flow_cmd_fill_info(struct sw_flow *flow, struct datapath *dp,
static struct sk_buff *ovs_flow_cmd_alloc_info(struct sw_flow *flow)
{
const struct sw_flow_actions *sf_acts;
int len;
sf_acts = rcu_dereference_protected(flow->sf_acts,
lockdep_genl_is_held());
/* OVS_FLOW_ATTR_KEY */
len = nla_total_size(FLOW_BUFSIZE);
/* OVS_FLOW_ATTR_ACTIONS */
len += nla_total_size(sf_acts->actions_len);
/* OVS_FLOW_ATTR_STATS */
len += nla_total_size(sizeof(struct ovs_flow_stats));
/* OVS_FLOW_ATTR_TCP_FLAGS */
len += nla_total_size(1);
/* OVS_FLOW_ATTR_USED */
len += nla_total_size(8);
len += NLMSG_ALIGN(sizeof(struct ovs_header));
sf_acts = ovsl_dereference(flow->sf_acts);
return genlmsg_new(len, GFP_KERNEL);
return genlmsg_new(ovs_flow_cmd_msg_size(sf_acts), GFP_KERNEL);
}
static struct sk_buff *ovs_flow_cmd_build_info(struct sw_flow *flow,
......@@ -947,12 +982,13 @@ 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;
......@@ -960,7 +996,7 @@ 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)) {
......@@ -970,7 +1006,7 @@ 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);
}
}
......@@ -978,7 +1014,7 @@ 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);
......@@ -1011,11 +1047,10 @@ 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) ||
......@@ -1026,7 +1061,7 @@ 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);
......@@ -1042,11 +1077,10 @@ 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));
......@@ -1054,6 +1088,8 @@ 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;
}
......@@ -1076,21 +1112,32 @@ 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)
......@@ -1105,25 +1152,33 @@ 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);
......@@ -1132,10 +1187,13 @@ 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)
......@@ -1144,11 +1202,14 @@ 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;
......@@ -1169,6 +1230,7 @@ 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;
}
......@@ -1214,6 +1276,16 @@ 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)
{
......@@ -1252,7 +1324,7 @@ 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);
......@@ -1264,7 +1336,7 @@ 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])
......@@ -1298,12 +1370,12 @@ 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)));
......@@ -1354,37 +1426,34 @@ 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;
......@@ -1395,14 +1464,11 @@ 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);
}
......@@ -1413,24 +1479,27 @@ 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)
......@@ -1439,9 +1508,11 @@ 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);
......@@ -1449,31 +1520,45 @@ 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)
......@@ -1483,6 +1568,7 @@ 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,
......@@ -1491,6 +1577,7 @@ static int ovs_dp_cmd_dump(struct sk_buff *skb, struct netlink_callback *cb)
break;
i++;
}
ovs_unlock();
cb->args[0] = i;
......@@ -1543,7 +1630,7 @@ 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)
{
......@@ -1582,7 +1669,7 @@ static int ovs_vport_cmd_fill_info(struct vport *vport, struct sk_buff *skb,
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)
{
......@@ -1601,7 +1688,7 @@ 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])
......@@ -1627,7 +1714,7 @@ 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(-ENODEV);
return vport;
......@@ -1651,7 +1738,7 @@ 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)
......@@ -1664,7 +1751,7 @@ 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;
......@@ -1674,7 +1761,7 @@ 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;
}
......@@ -1700,11 +1787,11 @@ 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;
}
......@@ -1716,7 +1803,7 @@ 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))
......@@ -1742,11 +1829,12 @@ static int ovs_vport_cmd_set(struct sk_buff *skb, struct genl_info *info)
goto exit_unlock;
}
genl_notify(reply, genl_info_net(info), info->snd_portid,
ovs_dp_vport_multicast_group.id, info->nlhdr, GFP_KERNEL);
ovs_unlock();
ovs_notify(reply, info, &ovs_dp_vport_multicast_group);
return 0;
exit_unlock:
rtnl_unlock();
ovs_unlock();
return err;
}
......@@ -1757,7 +1845,7 @@ 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))
......@@ -1777,11 +1865,10 @@ 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;
}
......@@ -1941,13 +2028,13 @@ 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);
......@@ -1958,8 +2045,7 @@ static void rehash_flow_table(struct work_struct *work)
}
}
rtnl_unlock();
genl_unlock();
ovs_unlock();
schedule_delayed_work(&rehash_flow_wq, REHASH_FLOW_INTERVAL);
}
......@@ -1968,18 +2054,21 @@ 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 = {
......
......@@ -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.
......@@ -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 */
......@@ -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;
......
......@@ -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;
}
......
......@@ -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 *);
......
......@@ -173,16 +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);
......@@ -195,10 +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)
......
......@@ -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);
}
......
......@@ -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 @@ struct vport *ovs_vport_add(const struct vport_parms *parms)
* @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;
}
......
......@@ -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;
......@@ -89,7 +82,7 @@ struct vport {
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;
......@@ -138,14 +131,14 @@ struct vport_parms {
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 *);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment