2009-07-08 13:19:16 -07:00
|
|
|
|
/*
|
2015-01-11 13:25:24 -08:00
|
|
|
|
* Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015 Nicira, Inc.
|
2009-07-08 13:19:16 -07:00
|
|
|
|
*
|
2009-06-15 15:11:30 -07:00
|
|
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
|
|
|
* you may not use this file except in compliance with the License.
|
|
|
|
|
* You may obtain a copy of the License at:
|
2009-07-08 13:19:16 -07:00
|
|
|
|
*
|
2009-06-15 15:11:30 -07:00
|
|
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
|
|
|
*
|
|
|
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
|
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
|
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
|
|
|
* See the License for the specific language governing permissions and
|
|
|
|
|
* limitations under the License.
|
2009-07-08 13:19:16 -07:00
|
|
|
|
*/
|
|
|
|
|
#ifndef FLOW_H
|
|
|
|
|
#define FLOW_H 1
|
|
|
|
|
|
2010-02-12 12:51:36 -08:00
|
|
|
|
#include <sys/types.h>
|
2009-07-08 13:19:16 -07:00
|
|
|
|
#include <netinet/in.h>
|
|
|
|
|
#include <stdbool.h>
|
|
|
|
|
#include <stdint.h>
|
|
|
|
|
#include <string.h>
|
2013-09-25 15:07:21 -07:00
|
|
|
|
#include "byte-order.h"
|
2010-04-12 11:49:16 -04:00
|
|
|
|
#include "openflow/nicira-ext.h"
|
2009-07-08 13:19:16 -07:00
|
|
|
|
#include "openflow/openflow.h"
|
2014-04-18 08:26:56 -07:00
|
|
|
|
#include "packets.h"
|
2009-07-08 13:19:16 -07:00
|
|
|
|
#include "hash.h"
|
|
|
|
|
#include "util.h"
|
|
|
|
|
|
2011-01-26 07:11:50 -08:00
|
|
|
|
struct dpif_flow_stats;
|
2009-07-08 13:19:16 -07:00
|
|
|
|
struct ds;
|
2010-11-08 10:37:35 -08:00
|
|
|
|
struct flow_wildcards;
|
2012-09-04 12:43:53 -07:00
|
|
|
|
struct minimask;
|
2009-07-08 13:19:16 -07:00
|
|
|
|
struct ofpbuf;
|
2014-02-26 18:08:04 -08:00
|
|
|
|
struct pkt_metadata;
|
2009-07-08 13:19:16 -07:00
|
|
|
|
|
2011-07-29 13:15:09 -07:00
|
|
|
|
/* This sequence number should be incremented whenever anything involving flows
|
|
|
|
|
* or the wildcarding of flows changes. This will cause build assertion
|
|
|
|
|
* failures in places which likely need to be updated. */
|
2015-02-14 15:13:17 +01:00
|
|
|
|
#define FLOW_WC_SEQ 31
|
2011-07-29 13:15:09 -07:00
|
|
|
|
|
2014-07-28 09:50:37 -07:00
|
|
|
|
/* Number of Open vSwitch extension 32-bit registers. */
|
2012-03-08 14:44:54 -08:00
|
|
|
|
#define FLOW_N_REGS 8
|
2010-11-11 10:41:33 -08:00
|
|
|
|
BUILD_ASSERT_DECL(FLOW_N_REGS <= NXM_NX_MAX_REGS);
|
2015-01-06 11:10:42 -08:00
|
|
|
|
BUILD_ASSERT_DECL(FLOW_N_REGS % 2 == 0); /* Even. */
|
2010-11-11 10:41:33 -08:00
|
|
|
|
|
2014-07-28 09:50:37 -07:00
|
|
|
|
/* Number of OpenFlow 1.5+ 64-bit registers.
|
|
|
|
|
*
|
|
|
|
|
* Each of these overlays a pair of Open vSwitch 32-bit registers, so there
|
|
|
|
|
* are half as many of them.*/
|
|
|
|
|
#define FLOW_N_XREGS (FLOW_N_REGS / 2)
|
|
|
|
|
|
2011-01-23 18:44:44 -08:00
|
|
|
|
/* Used for struct flow's dl_type member for frames that have no Ethernet
|
|
|
|
|
* type, that is, pure 802.2 frames. */
|
|
|
|
|
#define FLOW_DL_TYPE_NONE 0x5ff
|
|
|
|
|
|
Implement new fragment handling policy.
Until now, OVS has handled IP fragments more awkwardly than necessary. It
has not been possible to match on L4 headers, even in fragments with offset
0 where they are actually present. This means that there was no way to
implement ACLs that treat, say, different TCP ports differently, on
fragmented traffic; instead, all decisions for fragment forwarding had to
be made on the basis of L2 and L3 headers alone.
This commit improves the situation significantly. It is still not possible
to match on L4 headers in fragments with nonzero offset, because that
information is simply not present in such fragments, but this commit adds
the ability to match on L4 headers for fragments with zero offset. This
means that it becomes possible to implement ACLs that drop such "first
fragments" on the basis of L4 headers. In practice, that effectively
blocks even fragmented traffic on an L4 basis, because the receiving IP
stack cannot reassemble a full packet when the first fragment is missing.
This commit works by adding a new "fragment type" to the kernel flow match
and making it available through OpenFlow as a new NXM field named
NXM_NX_IP_FRAG. Because OpenFlow 1.0 explicitly says that the L4 fields
are always 0 for IP fragments, it adds a new OpenFlow fragment handling
mode that fills in the L4 fields for "first fragments". It also enhances
ovs-ofctl to allow users to configure this new fragment handling mode and
to parse the new field.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Bug #7557.
2011-10-19 21:33:44 -07:00
|
|
|
|
/* Fragment bits, used for IPv4 and IPv6, always zero for non-IP flows. */
|
2011-11-09 17:10:27 -08:00
|
|
|
|
#define FLOW_NW_FRAG_ANY (1 << 0) /* Set for any IP frag. */
|
|
|
|
|
#define FLOW_NW_FRAG_LATER (1 << 1) /* Set for IP frag with nonzero offset. */
|
|
|
|
|
#define FLOW_NW_FRAG_MASK (FLOW_NW_FRAG_ANY | FLOW_NW_FRAG_LATER)
|
Implement new fragment handling policy.
Until now, OVS has handled IP fragments more awkwardly than necessary. It
has not been possible to match on L4 headers, even in fragments with offset
0 where they are actually present. This means that there was no way to
implement ACLs that treat, say, different TCP ports differently, on
fragmented traffic; instead, all decisions for fragment forwarding had to
be made on the basis of L2 and L3 headers alone.
This commit improves the situation significantly. It is still not possible
to match on L4 headers in fragments with nonzero offset, because that
information is simply not present in such fragments, but this commit adds
the ability to match on L4 headers for fragments with zero offset. This
means that it becomes possible to implement ACLs that drop such "first
fragments" on the basis of L4 headers. In practice, that effectively
blocks even fragmented traffic on an L4 basis, because the receiving IP
stack cannot reassemble a full packet when the first fragment is missing.
This commit works by adding a new "fragment type" to the kernel flow match
and making it available through OpenFlow as a new NXM field named
NXM_NX_IP_FRAG. Because OpenFlow 1.0 explicitly says that the L4 fields
are always 0 for IP fragments, it adds a new OpenFlow fragment handling
mode that fills in the L4 fields for "first fragments". It also enhances
ovs-ofctl to allow users to configure this new fragment handling mode and
to parse the new field.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Bug #7557.
2011-10-19 21:33:44 -07:00
|
|
|
|
|
2011-11-09 17:10:27 -08:00
|
|
|
|
BUILD_ASSERT_DECL(FLOW_NW_FRAG_ANY == NX_IP_FRAG_ANY);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_NW_FRAG_LATER == NX_IP_FRAG_LATER);
|
Implement new fragment handling policy.
Until now, OVS has handled IP fragments more awkwardly than necessary. It
has not been possible to match on L4 headers, even in fragments with offset
0 where they are actually present. This means that there was no way to
implement ACLs that treat, say, different TCP ports differently, on
fragmented traffic; instead, all decisions for fragment forwarding had to
be made on the basis of L2 and L3 headers alone.
This commit improves the situation significantly. It is still not possible
to match on L4 headers in fragments with nonzero offset, because that
information is simply not present in such fragments, but this commit adds
the ability to match on L4 headers for fragments with zero offset. This
means that it becomes possible to implement ACLs that drop such "first
fragments" on the basis of L4 headers. In practice, that effectively
blocks even fragmented traffic on an L4 basis, because the receiving IP
stack cannot reassemble a full packet when the first fragment is missing.
This commit works by adding a new "fragment type" to the kernel flow match
and making it available through OpenFlow as a new NXM field named
NXM_NX_IP_FRAG. Because OpenFlow 1.0 explicitly says that the L4 fields
are always 0 for IP fragments, it adds a new OpenFlow fragment handling
mode that fills in the L4 fields for "first fragments". It also enhances
ovs-ofctl to allow users to configure this new fragment handling mode and
to parse the new field.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Bug #7557.
2011-10-19 21:33:44 -07:00
|
|
|
|
|
2012-09-13 20:11:08 -07:00
|
|
|
|
#define FLOW_TNL_F_DONT_FRAGMENT (1 << 0)
|
|
|
|
|
#define FLOW_TNL_F_CSUM (1 << 1)
|
|
|
|
|
#define FLOW_TNL_F_KEY (1 << 2)
|
2014-05-27 21:50:35 -07:00
|
|
|
|
#define FLOW_TNL_F_OAM (1 << 3)
|
2012-11-21 18:51:36 -08:00
|
|
|
|
|
2014-09-05 15:44:20 -07:00
|
|
|
|
#define FLOW_TNL_F_MASK ((1 << 4) - 1)
|
|
|
|
|
|
2012-11-21 18:51:36 -08:00
|
|
|
|
const char *flow_tun_flag_to_string(uint32_t flags);
|
|
|
|
|
|
2014-02-04 10:32:35 -08:00
|
|
|
|
/* Maximum number of supported MPLS labels. */
|
|
|
|
|
#define FLOW_MAX_MPLS_LABELS 3
|
|
|
|
|
|
2012-09-25 15:25:51 -07:00
|
|
|
|
/*
|
2013-10-17 14:28:20 -07:00
|
|
|
|
* A flow in the network.
|
|
|
|
|
*
|
|
|
|
|
* Must be initialized to all zeros to make any compiler-induced padding
|
|
|
|
|
* zeroed. Helps also in keeping unused fields (such as mutually exclusive
|
|
|
|
|
* IPv4 and IPv6 addresses) zeroed out.
|
|
|
|
|
*
|
|
|
|
|
* The meaning of 'in_port' is context-dependent. In most cases, it is a
|
|
|
|
|
* 16-bit OpenFlow 1.0 port number. In the software datapath interface (dpif)
|
2014-09-18 04:17:54 -07:00
|
|
|
|
* layer and its implementations (e.g. dpif-netlink, dpif-netdev), it is
|
|
|
|
|
* instead a 32-bit datapath port number.
|
2013-11-19 17:31:29 -08:00
|
|
|
|
*
|
|
|
|
|
* The fields are organized in four segments to facilitate staged lookup, where
|
|
|
|
|
* lower layer fields are first used to determine if the later fields need to
|
|
|
|
|
* be looked at. This enables better wildcarding for datapath flows.
|
2014-04-18 08:26:56 -07:00
|
|
|
|
*
|
|
|
|
|
* NOTE: Order of the fields is significant, any change in the order must be
|
|
|
|
|
* reflected in miniflow_extract()!
|
2013-10-17 14:28:20 -07:00
|
|
|
|
*/
|
2010-10-11 13:31:35 -07:00
|
|
|
|
struct flow {
|
2014-10-17 09:37:11 -07:00
|
|
|
|
/* Metadata */
|
2012-09-13 20:11:08 -07:00
|
|
|
|
struct flow_tnl tunnel; /* Encapsulating tunnel parameters. */
|
2012-06-27 01:09:44 +12:00
|
|
|
|
ovs_be64 metadata; /* OpenFlow Metadata. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
uint32_t regs[FLOW_N_REGS]; /* Registers. */
|
|
|
|
|
uint32_t skb_priority; /* Packet priority for QoS. */
|
|
|
|
|
uint32_t pkt_mark; /* Packet mark. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint32_t dp_hash; /* Datapath computed hash value. The exact
|
|
|
|
|
* computation is opaque to the user space. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
union flow_in_port in_port; /* Input port.*/
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint32_t recirc_id; /* Must be exact match. */
|
2015-01-11 13:25:24 -08:00
|
|
|
|
uint32_t conj_id; /* Conjunction ID. */
|
2014-11-03 14:24:01 -08:00
|
|
|
|
ofp_port_t actset_output; /* Output port in action set. */
|
2015-01-11 13:25:24 -08:00
|
|
|
|
uint8_t pad1[6]; /* Pad to 64 bits. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* L2, Order the same as in the Ethernet header! (64-bit aligned) */
|
2014-10-22 14:58:43 +08:00
|
|
|
|
uint8_t dl_dst[ETH_ADDR_LEN]; /* Ethernet destination address. */
|
|
|
|
|
uint8_t dl_src[ETH_ADDR_LEN]; /* Ethernet source address. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
ovs_be16 dl_type; /* Ethernet frame type. */
|
|
|
|
|
ovs_be16 vlan_tci; /* If 802.1Q, TCI | VLAN_CFI; otherwise 0. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
ovs_be32 mpls_lse[ROUND_UP(FLOW_MAX_MPLS_LABELS, 2)]; /* MPLS label stack
|
|
|
|
|
(with padding). */
|
|
|
|
|
/* L3 (64-bit aligned) */
|
|
|
|
|
ovs_be32 nw_src; /* IPv4 source address. */
|
|
|
|
|
ovs_be32 nw_dst; /* IPv4 destination address. */
|
2011-11-02 18:22:22 -07:00
|
|
|
|
struct in6_addr ipv6_src; /* IPv6 source address. */
|
|
|
|
|
struct in6_addr ipv6_dst; /* IPv6 destination address. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
ovs_be32 ipv6_label; /* IPv6 flow label. */
|
|
|
|
|
uint8_t nw_frag; /* FLOW_FRAG_* flags. */
|
2011-11-09 17:10:27 -08:00
|
|
|
|
uint8_t nw_tos; /* IP ToS (including DSCP and ECN). */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
uint8_t nw_ttl; /* IP TTL/Hop Limit. */
|
|
|
|
|
uint8_t nw_proto; /* IP protocol or low 8 bits of ARP opcode. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
struct in6_addr nd_target; /* IPv6 neighbor discovery (ND) target. */
|
2014-10-22 14:58:43 +08:00
|
|
|
|
uint8_t arp_sha[ETH_ADDR_LEN]; /* ARP/ND source hardware address. */
|
|
|
|
|
uint8_t arp_tha[ETH_ADDR_LEN]; /* ARP/ND target hardware address. */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
ovs_be16 tcp_flags; /* TCP flags. With L3 to avoid matching L4. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
ovs_be16 pad2; /* Pad to 64 bits. */
|
2014-02-04 10:32:35 -08:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* L4 (64-bit aligned) */
|
2013-11-19 17:31:29 -08:00
|
|
|
|
ovs_be16 tp_src; /* TCP/UDP/SCTP source port. */
|
2014-06-18 22:14:30 -03:00
|
|
|
|
ovs_be16 tp_dst; /* TCP/UDP/SCTP destination port. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
ovs_be32 igmp_group_ip4; /* IGMP group IPv4 address.
|
2014-06-18 22:14:30 -03:00
|
|
|
|
* Keep last for BUILD_ASSERT_DECL below. */
|
2010-10-11 13:31:35 -07:00
|
|
|
|
};
|
2015-01-06 11:10:42 -08:00
|
|
|
|
BUILD_ASSERT_DECL(sizeof(struct flow) % sizeof(uint64_t) == 0);
|
2012-06-18 15:12:57 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define FLOW_U64S (sizeof(struct flow) / sizeof(uint64_t))
|
2012-08-07 13:43:18 -07:00
|
|
|
|
|
2014-09-20 07:30:02 +00:00
|
|
|
|
/* Some flow fields are mutually exclusive or only appear within the flow
|
|
|
|
|
* pipeline. IPv6 headers are bigger than IPv4 and MPLS, and IPv6 ND packets
|
|
|
|
|
* are bigger than TCP,UDP and IGMP packets. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define FLOW_MAX_PACKET_U64S (FLOW_U64S \
|
|
|
|
|
/* Unused in datapath */ - FLOW_U64_SIZE(regs) \
|
|
|
|
|
- FLOW_U64_SIZE(metadata) \
|
|
|
|
|
/* L2.5/3 */ - FLOW_U64_SIZE(nw_src) /* incl. nw_dst */ \
|
|
|
|
|
- FLOW_U64_SIZE(mpls_lse) \
|
|
|
|
|
/* L4 */ - FLOW_U64_SIZE(tp_src) \
|
2014-09-20 07:30:02 +00:00
|
|
|
|
)
|
|
|
|
|
|
2012-06-18 15:12:57 -07:00
|
|
|
|
/* Remember to update FLOW_WC_SEQ when changing 'struct flow'. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
BUILD_ASSERT_DECL(offsetof(struct flow, igmp_group_ip4) + sizeof(uint32_t)
|
2015-01-11 13:25:24 -08:00
|
|
|
|
== sizeof(struct flow_tnl) + 192
|
2015-02-14 15:13:17 +01:00
|
|
|
|
&& FLOW_WC_SEQ == 31);
|
2013-11-19 17:31:29 -08:00
|
|
|
|
|
|
|
|
|
/* Incremental points at which flow classification may be performed in
|
|
|
|
|
* segments.
|
|
|
|
|
* This is located here since this is dependent on the structure of the
|
|
|
|
|
* struct flow defined above:
|
2015-01-06 11:10:42 -08:00
|
|
|
|
* Each offset must be on a distinct, successive U64 boundary strictly
|
2013-11-19 17:31:29 -08:00
|
|
|
|
* within the struct flow. */
|
|
|
|
|
enum {
|
2014-04-18 08:26:56 -07:00
|
|
|
|
FLOW_SEGMENT_1_ENDS_AT = offsetof(struct flow, dl_dst),
|
2015-01-06 11:10:42 -08:00
|
|
|
|
FLOW_SEGMENT_2_ENDS_AT = offsetof(struct flow, nw_src),
|
2014-04-18 08:26:56 -07:00
|
|
|
|
FLOW_SEGMENT_3_ENDS_AT = offsetof(struct flow, tp_src),
|
2013-11-19 17:31:29 -08:00
|
|
|
|
};
|
2015-01-06 11:10:42 -08:00
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_1_ENDS_AT % sizeof(uint64_t) == 0);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_2_ENDS_AT % sizeof(uint64_t) == 0);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_3_ENDS_AT % sizeof(uint64_t) == 0);
|
2013-11-19 17:31:29 -08:00
|
|
|
|
BUILD_ASSERT_DECL( 0 < FLOW_SEGMENT_1_ENDS_AT);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_1_ENDS_AT < FLOW_SEGMENT_2_ENDS_AT);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_2_ENDS_AT < FLOW_SEGMENT_3_ENDS_AT);
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_SEGMENT_3_ENDS_AT < sizeof(struct flow));
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
extern const uint8_t flow_segment_u64s[];
|
2010-10-11 13:31:35 -07:00
|
|
|
|
|
Do not include zeroed metadata fields in NXM/OXM packet-in messages.
NXM and OpenFlow 1.2+ allow including the values of arbitrary flow metadata
in "packet-in" messages. Open vSwitch has until now always included all
the values of the metadata fields that it implements in NXT_PACKET_IN
messages.
However, this has at least two disadvantages:
- Most of the metadata fields tend to be zero most of the time, which
wastes space in the message.
- It means that controllers must be very liberal about accepting
fields that they know nothing about in packet-in messages, since any
switch upgrade could cause new fields to appear even if the
controller does nothing to give them nonzero values. (Controllers
have to be prepared to tolerate unknown fields in any case, but this
property makes unknown fields more likely to appear than otherwise.)
This commit changes Open vSwitch so that metadata fields whose values are
zero are not reported in packet-ins, fixing both problems. (This is
explicitly allowed by OpenFlow 1.2+.)
This commit mainly fixes a sort of internal conceptual dissonance centering
around struct flow_metadata. This structure is supposed to report the
metadata for a given flow. If you look at a flow, it has particular
metadata values; it doesn't have masks, and the idea of a mask for a
particular flow doesn't really make sense. However, struct flow_metadata
did have masks. This led to internal confusion; one can see this in, for
example, the following code removed by this commit in ofproto-dpif.c to
handle misses in the OpenFlow flow table:
/* Registers aren't meaningful on a miss. */
memset(pin.fmd.reg_masks, 0, sizeof pin.fmd.reg_masks);
What this code was really trying to say is that on a flow miss, the
registers are zero, so they shouldn't be included in the packet-in message.
It did manage to omit the registers, by marking them as "wild", but it is
conceptually more correct to simply omit them because they are zero (and
that's one effect of this commit).
Bug #12968.
Reported-by: Igor Ganichev <iganichev@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
2012-08-15 10:16:49 -07:00
|
|
|
|
/* Represents the metadata fields of struct flow. */
|
2012-01-04 16:40:13 -08:00
|
|
|
|
struct flow_metadata {
|
2014-03-04 14:20:19 -08:00
|
|
|
|
uint32_t dp_hash; /* Datapath computed hash field. */
|
|
|
|
|
uint32_t recirc_id; /* Recirculation ID. */
|
2012-01-04 16:40:13 -08:00
|
|
|
|
ovs_be64 tun_id; /* Encapsulating tunnel ID. */
|
2013-05-09 15:24:16 +03:00
|
|
|
|
ovs_be32 tun_src; /* Tunnel outer IPv4 src addr */
|
|
|
|
|
ovs_be32 tun_dst; /* Tunnel outer IPv4 dst addr */
|
2015-02-14 15:13:17 +01:00
|
|
|
|
ovs_be16 gbp_id; /* Group policy ID */
|
|
|
|
|
uint8_t gbp_flags; /* Group policy flags */
|
Do not include zeroed metadata fields in NXM/OXM packet-in messages.
NXM and OpenFlow 1.2+ allow including the values of arbitrary flow metadata
in "packet-in" messages. Open vSwitch has until now always included all
the values of the metadata fields that it implements in NXT_PACKET_IN
messages.
However, this has at least two disadvantages:
- Most of the metadata fields tend to be zero most of the time, which
wastes space in the message.
- It means that controllers must be very liberal about accepting
fields that they know nothing about in packet-in messages, since any
switch upgrade could cause new fields to appear even if the
controller does nothing to give them nonzero values. (Controllers
have to be prepared to tolerate unknown fields in any case, but this
property makes unknown fields more likely to appear than otherwise.)
This commit changes Open vSwitch so that metadata fields whose values are
zero are not reported in packet-ins, fixing both problems. (This is
explicitly allowed by OpenFlow 1.2+.)
This commit mainly fixes a sort of internal conceptual dissonance centering
around struct flow_metadata. This structure is supposed to report the
metadata for a given flow. If you look at a flow, it has particular
metadata values; it doesn't have masks, and the idea of a mask for a
particular flow doesn't really make sense. However, struct flow_metadata
did have masks. This led to internal confusion; one can see this in, for
example, the following code removed by this commit in ofproto-dpif.c to
handle misses in the OpenFlow flow table:
/* Registers aren't meaningful on a miss. */
memset(pin.fmd.reg_masks, 0, sizeof pin.fmd.reg_masks);
What this code was really trying to say is that on a flow miss, the
registers are zero, so they shouldn't be included in the packet-in message.
It did manage to omit the registers, by marking them as "wild", but it is
conceptually more correct to simply omit them because they are zero (and
that's one effect of this commit).
Bug #12968.
Reported-by: Igor Ganichev <iganichev@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
2012-08-15 10:16:49 -07:00
|
|
|
|
ovs_be64 metadata; /* OpenFlow 1.1+ metadata field. */
|
2012-01-04 16:40:13 -08:00
|
|
|
|
uint32_t regs[FLOW_N_REGS]; /* Registers. */
|
2013-08-06 12:57:16 -07:00
|
|
|
|
uint32_t pkt_mark; /* Packet mark. */
|
2013-06-19 16:58:44 -07:00
|
|
|
|
ofp_port_t in_port; /* OpenFlow port or zero. */
|
2012-01-04 16:40:13 -08:00
|
|
|
|
};
|
|
|
|
|
|
2014-02-26 18:08:04 -08:00
|
|
|
|
void flow_extract(struct ofpbuf *, const struct pkt_metadata *md,
|
2013-06-19 16:58:44 -07:00
|
|
|
|
struct flow *);
|
2013-01-25 16:22:07 +09:00
|
|
|
|
|
2011-08-19 09:39:16 -07:00
|
|
|
|
void flow_zero_wildcards(struct flow *, const struct flow_wildcards *);
|
2013-12-06 18:53:12 -08:00
|
|
|
|
void flow_unwildcard_tp_ports(const struct flow *, struct flow_wildcards *);
|
2012-01-04 16:40:13 -08:00
|
|
|
|
void flow_get_metadata(const struct flow *, struct flow_metadata *);
|
2011-08-19 09:39:16 -07:00
|
|
|
|
|
2010-09-03 11:30:02 -07:00
|
|
|
|
char *flow_to_string(const struct flow *);
|
2012-11-21 18:51:36 -08:00
|
|
|
|
void format_flags(struct ds *ds, const char *(*bit_to_string)(uint32_t),
|
|
|
|
|
uint32_t flags, char del);
|
2013-12-02 15:14:09 -08:00
|
|
|
|
void format_flags_masked(struct ds *ds, const char *name,
|
|
|
|
|
const char *(*bit_to_string)(uint32_t),
|
|
|
|
|
uint32_t flags, uint32_t mask);
|
2012-11-21 18:51:36 -08:00
|
|
|
|
|
2010-09-03 11:30:02 -07:00
|
|
|
|
void flow_format(struct ds *, const struct flow *);
|
|
|
|
|
void flow_print(FILE *, const struct flow *);
|
2011-10-25 16:33:38 -07:00
|
|
|
|
static inline int flow_compare_3way(const struct flow *, const struct flow *);
|
2010-09-03 11:30:02 -07:00
|
|
|
|
static inline bool flow_equal(const struct flow *, const struct flow *);
|
|
|
|
|
static inline size_t flow_hash(const struct flow *, uint32_t basis);
|
2009-07-08 13:19:16 -07:00
|
|
|
|
|
2012-07-22 22:42:55 -07:00
|
|
|
|
void flow_set_dl_vlan(struct flow *, ovs_be16 vid);
|
2012-07-22 23:20:22 -07:00
|
|
|
|
void flow_set_vlan_vid(struct flow *, ovs_be16 vid);
|
2011-11-21 14:14:02 -08:00
|
|
|
|
void flow_set_vlan_pcp(struct flow *, uint8_t pcp);
|
|
|
|
|
|
2014-02-04 10:32:35 -08:00
|
|
|
|
int flow_count_mpls_labels(const struct flow *, struct flow_wildcards *);
|
|
|
|
|
int flow_count_common_mpls_labels(const struct flow *a, int an,
|
|
|
|
|
const struct flow *b, int bn,
|
|
|
|
|
struct flow_wildcards *wc);
|
|
|
|
|
void flow_push_mpls(struct flow *, int n, ovs_be16 mpls_eth_type,
|
|
|
|
|
struct flow_wildcards *);
|
|
|
|
|
bool flow_pop_mpls(struct flow *, int n, ovs_be16 eth_type,
|
|
|
|
|
struct flow_wildcards *);
|
|
|
|
|
void flow_set_mpls_label(struct flow *, int idx, ovs_be32 label);
|
|
|
|
|
void flow_set_mpls_ttl(struct flow *, int idx, uint8_t ttl);
|
|
|
|
|
void flow_set_mpls_tc(struct flow *, int idx, uint8_t tc);
|
|
|
|
|
void flow_set_mpls_bos(struct flow *, int idx, uint8_t stack);
|
|
|
|
|
void flow_set_mpls_lse(struct flow *, int idx, ovs_be32 lse);
|
2013-01-25 16:22:07 +09:00
|
|
|
|
|
2011-09-08 14:32:13 -07:00
|
|
|
|
void flow_compose(struct ofpbuf *, const struct flow *);
|
|
|
|
|
|
2014-07-28 09:50:37 -07:00
|
|
|
|
static inline uint64_t
|
|
|
|
|
flow_get_xreg(const struct flow *flow, int idx)
|
|
|
|
|
{
|
|
|
|
|
return ((uint64_t) flow->regs[idx * 2] << 32) | flow->regs[idx * 2 + 1];
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline void
|
|
|
|
|
flow_set_xreg(struct flow *flow, int idx, uint64_t value)
|
|
|
|
|
{
|
|
|
|
|
flow->regs[idx * 2] = value >> 32;
|
|
|
|
|
flow->regs[idx * 2 + 1] = value;
|
|
|
|
|
}
|
|
|
|
|
|
2009-07-08 13:19:16 -07:00
|
|
|
|
static inline int
|
2011-10-25 16:33:38 -07:00
|
|
|
|
flow_compare_3way(const struct flow *a, const struct flow *b)
|
2009-07-08 13:19:16 -07:00
|
|
|
|
{
|
2012-06-18 15:12:57 -07:00
|
|
|
|
return memcmp(a, b, sizeof *a);
|
2009-07-08 13:19:16 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool
|
2010-09-03 11:30:02 -07:00
|
|
|
|
flow_equal(const struct flow *a, const struct flow *b)
|
2009-07-08 13:19:16 -07:00
|
|
|
|
{
|
2011-10-25 16:33:38 -07:00
|
|
|
|
return !flow_compare_3way(a, b);
|
2009-07-08 13:19:16 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline size_t
|
2010-09-03 11:30:02 -07:00
|
|
|
|
flow_hash(const struct flow *flow, uint32_t basis)
|
2009-07-08 13:19:16 -07:00
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return hash_words64((const uint64_t *)flow,
|
|
|
|
|
sizeof *flow / sizeof(uint64_t), basis);
|
2009-07-08 13:19:16 -07:00
|
|
|
|
}
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
2013-06-19 16:58:44 -07:00
|
|
|
|
static inline uint16_t
|
|
|
|
|
ofp_to_u16(ofp_port_t ofp_port)
|
|
|
|
|
{
|
|
|
|
|
return (OVS_FORCE uint16_t) ofp_port;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline uint32_t
|
|
|
|
|
odp_to_u32(odp_port_t odp_port)
|
|
|
|
|
{
|
|
|
|
|
return (OVS_FORCE uint32_t) odp_port;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline uint32_t
|
|
|
|
|
ofp11_to_u32(ofp11_port_t ofp11_port)
|
|
|
|
|
{
|
|
|
|
|
return (OVS_FORCE uint32_t) ofp11_port;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline ofp_port_t
|
|
|
|
|
u16_to_ofp(uint16_t port)
|
|
|
|
|
{
|
|
|
|
|
return OFP_PORT_C(port);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline odp_port_t
|
|
|
|
|
u32_to_odp(uint32_t port)
|
|
|
|
|
{
|
|
|
|
|
return ODP_PORT_C(port);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline ofp11_port_t
|
|
|
|
|
u32_to_ofp11(uint32_t port)
|
|
|
|
|
{
|
|
|
|
|
return OFP11_PORT_C(port);
|
|
|
|
|
}
|
|
|
|
|
|
2013-06-22 10:33:27 -07:00
|
|
|
|
static inline uint32_t
|
|
|
|
|
hash_ofp_port(ofp_port_t ofp_port)
|
|
|
|
|
{
|
|
|
|
|
return hash_int(ofp_to_u16(ofp_port), 0);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline uint32_t
|
|
|
|
|
hash_odp_port(odp_port_t odp_port)
|
|
|
|
|
{
|
|
|
|
|
return hash_int(odp_to_u32(odp_port), 0);
|
|
|
|
|
}
|
2012-08-07 13:43:18 -07:00
|
|
|
|
|
2012-08-07 13:38:38 -07:00
|
|
|
|
/* Wildcards for a flow.
|
|
|
|
|
*
|
|
|
|
|
* A 1-bit in each bit in 'masks' indicates that the corresponding bit of
|
|
|
|
|
* the flow is significant (must match). A 0-bit indicates that the
|
|
|
|
|
* corresponding bit of the flow is wildcarded (need not match). */
|
2009-07-08 13:19:16 -07:00
|
|
|
|
struct flow_wildcards {
|
2012-08-07 13:38:38 -07:00
|
|
|
|
struct flow masks;
|
2009-07-08 13:19:16 -07:00
|
|
|
|
};
|
2011-07-29 13:15:09 -07:00
|
|
|
|
|
2014-10-01 15:35:45 -07:00
|
|
|
|
#define WC_MASK_FIELD(WC, FIELD) \
|
|
|
|
|
memset(&(WC)->masks.FIELD, 0xff, sizeof (WC)->masks.FIELD)
|
|
|
|
|
#define WC_UNMASK_FIELD(WC, FIELD) \
|
|
|
|
|
memset(&(WC)->masks.FIELD, 0, sizeof (WC)->masks.FIELD)
|
|
|
|
|
|
2010-11-10 14:39:54 -08:00
|
|
|
|
void flow_wildcards_init_catchall(struct flow_wildcards *);
|
2010-10-27 20:15:56 -07:00
|
|
|
|
|
2014-10-01 15:35:45 -07:00
|
|
|
|
void flow_wildcards_init_for_packet(struct flow_wildcards *,
|
|
|
|
|
const struct flow *);
|
2014-10-17 09:37:11 -07:00
|
|
|
|
uint64_t flow_wc_map(const struct flow *);
|
2014-10-01 15:35:45 -07:00
|
|
|
|
|
2013-12-10 23:32:51 -08:00
|
|
|
|
void flow_wildcards_clear_non_packet_fields(struct flow_wildcards *);
|
|
|
|
|
|
2011-09-12 16:38:52 -07:00
|
|
|
|
bool flow_wildcards_is_catchall(const struct flow_wildcards *);
|
2010-11-08 16:45:00 -08:00
|
|
|
|
|
2010-11-11 10:41:33 -08:00
|
|
|
|
void flow_wildcards_set_reg_mask(struct flow_wildcards *,
|
|
|
|
|
int idx, uint32_t mask);
|
2014-07-28 09:50:37 -07:00
|
|
|
|
void flow_wildcards_set_xreg_mask(struct flow_wildcards *,
|
|
|
|
|
int idx, uint64_t mask);
|
2009-07-08 13:19:16 -07:00
|
|
|
|
|
2013-06-10 22:48:58 -07:00
|
|
|
|
void flow_wildcards_and(struct flow_wildcards *dst,
|
|
|
|
|
const struct flow_wildcards *src1,
|
|
|
|
|
const struct flow_wildcards *src2);
|
|
|
|
|
void flow_wildcards_or(struct flow_wildcards *dst,
|
|
|
|
|
const struct flow_wildcards *src1,
|
|
|
|
|
const struct flow_wildcards *src2);
|
2010-11-03 11:00:58 -07:00
|
|
|
|
bool flow_wildcards_has_extra(const struct flow_wildcards *,
|
|
|
|
|
const struct flow_wildcards *);
|
2011-05-26 16:23:21 -07:00
|
|
|
|
uint32_t flow_wildcards_hash(const struct flow_wildcards *, uint32_t basis);
|
2010-11-03 11:00:58 -07:00
|
|
|
|
bool flow_wildcards_equal(const struct flow_wildcards *,
|
|
|
|
|
const struct flow_wildcards *);
|
2014-02-26 10:07:38 -08:00
|
|
|
|
uint32_t flow_hash_5tuple(const struct flow *flow, uint32_t basis);
|
2011-02-01 18:50:25 -08:00
|
|
|
|
uint32_t flow_hash_symmetric_l4(const struct flow *flow, uint32_t basis);
|
2010-11-03 11:00:58 -07:00
|
|
|
|
|
2013-10-17 14:28:20 -07:00
|
|
|
|
/* Initialize a flow with random fields that matter for nx_hash_fields. */
|
|
|
|
|
void flow_random_hash_fields(struct flow *);
|
2013-06-26 16:37:16 -07:00
|
|
|
|
void flow_mask_hash_fields(const struct flow *, struct flow_wildcards *,
|
|
|
|
|
enum nx_hash_fields);
|
2011-07-13 16:20:24 -07:00
|
|
|
|
uint32_t flow_hash_fields(const struct flow *, enum nx_hash_fields,
|
|
|
|
|
uint16_t basis);
|
|
|
|
|
const char *flow_hash_fields_to_str(enum nx_hash_fields);
|
|
|
|
|
bool flow_hash_fields_valid(enum nx_hash_fields);
|
2011-06-06 14:21:40 -07:00
|
|
|
|
|
2013-06-10 22:48:58 -07:00
|
|
|
|
uint32_t flow_hash_in_wildcards(const struct flow *,
|
|
|
|
|
const struct flow_wildcards *,
|
|
|
|
|
uint32_t basis);
|
|
|
|
|
|
2012-08-07 13:43:18 -07:00
|
|
|
|
bool flow_equal_except(const struct flow *a, const struct flow *b,
|
|
|
|
|
const struct flow_wildcards *);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
|
|
|
|
/* Compressed flow. */
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Number of 64-bit words present in struct miniflow. */
|
|
|
|
|
#define MINI_N_INLINE 4
|
2014-08-26 15:11:39 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Maximum number of 64-bit words supported. */
|
|
|
|
|
BUILD_ASSERT_DECL(FLOW_U64S <= 63);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
|
|
|
|
/* A sparse representation of a "struct flow".
|
|
|
|
|
*
|
|
|
|
|
* A "struct flow" is fairly large and tends to be mostly zeros. Sparse
|
|
|
|
|
* representation has two advantages. First, it saves memory. Second, it
|
|
|
|
|
* saves time when the goal is to iterate over only the nonzero parts of the
|
|
|
|
|
* struct.
|
|
|
|
|
*
|
2015-01-06 11:10:42 -08:00
|
|
|
|
* The 'map' member holds one bit for each uint64_t in a "struct flow". Each
|
|
|
|
|
* 0-bit indicates that the corresponding uint64_t is zero, each 1-bit that it
|
2014-04-29 15:50:39 -07:00
|
|
|
|
* *may* be nonzero (see below how this applies to minimasks).
|
2012-09-04 12:43:53 -07:00
|
|
|
|
*
|
2014-04-29 15:50:39 -07:00
|
|
|
|
* The 'values_inline' boolean member indicates that the values are at
|
|
|
|
|
* 'inline_values'. If 'values_inline' is zero, then the values are
|
|
|
|
|
* offline at 'offline_values'. In either case, values is an array that has
|
|
|
|
|
* one element for each 1-bit in 'map'. The least-numbered 1-bit is in
|
|
|
|
|
* the first element of the values array, the next 1-bit is in the next array
|
|
|
|
|
* element, and so on.
|
2012-09-04 12:43:53 -07:00
|
|
|
|
*
|
2014-08-26 15:11:39 -07:00
|
|
|
|
* MINI_N_INLINE is the default number of inline words. When a miniflow is
|
|
|
|
|
* dynamically allocated the actual amount of inline storage may be different.
|
|
|
|
|
* In that case 'inline_values' contains storage at least for the number
|
2015-01-06 11:10:42 -08:00
|
|
|
|
* of words indicated by 'map' (one uint64_t for each 1-bit in the map).
|
2014-08-26 15:11:39 -07:00
|
|
|
|
*
|
2014-04-29 15:50:39 -07:00
|
|
|
|
* Elements in values array are allowed to be zero. This is useful for "struct
|
2013-02-06 16:13:19 -08:00
|
|
|
|
* minimatch", for which ensuring that the miniflow and minimask members have
|
2013-12-20 08:16:31 -08:00
|
|
|
|
* same 'map' allows optimization. This allowance applies only to a miniflow
|
|
|
|
|
* that is not a mask. That is, a minimask may NOT have zero elements in
|
|
|
|
|
* its 'values'.
|
2012-09-04 12:43:53 -07:00
|
|
|
|
*/
|
|
|
|
|
struct miniflow {
|
2014-04-29 15:50:39 -07:00
|
|
|
|
uint64_t map:63;
|
2014-08-27 08:21:45 -07:00
|
|
|
|
uint64_t values_inline:1;
|
2014-04-29 15:50:39 -07:00
|
|
|
|
union {
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint64_t *offline_values;
|
|
|
|
|
uint64_t inline_values[MINI_N_INLINE]; /* Minimum inline size. */
|
2014-04-29 15:50:39 -07:00
|
|
|
|
};
|
2012-09-04 12:43:53 -07:00
|
|
|
|
};
|
2014-08-26 15:11:39 -07:00
|
|
|
|
BUILD_ASSERT_DECL(sizeof(struct miniflow)
|
2015-01-06 11:10:42 -08:00
|
|
|
|
== sizeof(uint64_t) + MINI_N_INLINE * sizeof(uint64_t));
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_VALUES_SIZE(COUNT) ((COUNT) * sizeof(uint64_t))
|
2014-04-29 15:50:39 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline uint64_t *miniflow_values(struct miniflow *mf)
|
2014-04-29 15:50:39 -07:00
|
|
|
|
{
|
2014-04-29 15:50:39 -07:00
|
|
|
|
return OVS_LIKELY(mf->values_inline)
|
|
|
|
|
? mf->inline_values : mf->offline_values;
|
2014-04-29 15:50:39 -07:00
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline const uint64_t *miniflow_get_values(const struct miniflow *mf)
|
2014-04-29 15:50:39 -07:00
|
|
|
|
{
|
2014-04-29 15:50:39 -07:00
|
|
|
|
return OVS_LIKELY(mf->values_inline)
|
|
|
|
|
? mf->inline_values : mf->offline_values;
|
2014-04-29 15:50:39 -07:00
|
|
|
|
}
|
|
|
|
|
|
2014-04-18 08:26:56 -07:00
|
|
|
|
/* This is useful for initializing a miniflow for a miniflow_extract() call. */
|
|
|
|
|
static inline void miniflow_initialize(struct miniflow *mf,
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint64_t buf[FLOW_U64S])
|
2014-04-18 08:26:56 -07:00
|
|
|
|
{
|
|
|
|
|
mf->map = 0;
|
2015-01-06 11:10:42 -08:00
|
|
|
|
mf->values_inline = (buf == (uint64_t *)(mf + 1));
|
2014-04-29 15:50:39 -07:00
|
|
|
|
if (!mf->values_inline) {
|
|
|
|
|
mf->offline_values = buf;
|
|
|
|
|
}
|
2014-04-18 08:26:56 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
struct pkt_metadata;
|
|
|
|
|
|
|
|
|
|
/* The 'dst->values' must be initialized with a buffer with space for
|
2015-01-06 11:10:42 -08:00
|
|
|
|
* FLOW_U64S. 'dst->map' is ignored on input and set on output to
|
2014-04-18 08:26:56 -07:00
|
|
|
|
* indicate which fields were extracted. */
|
|
|
|
|
void miniflow_extract(struct ofpbuf *packet, const struct pkt_metadata *,
|
|
|
|
|
struct miniflow *dst);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
void miniflow_init(struct miniflow *, const struct flow *);
|
2013-02-06 16:13:19 -08:00
|
|
|
|
void miniflow_init_with_minimask(struct miniflow *, const struct flow *,
|
|
|
|
|
const struct minimask *);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
void miniflow_clone(struct miniflow *, const struct miniflow *);
|
2014-04-29 15:50:39 -07:00
|
|
|
|
void miniflow_clone_inline(struct miniflow *, const struct miniflow *,
|
|
|
|
|
size_t n_values);
|
2013-08-27 12:25:48 -07:00
|
|
|
|
void miniflow_move(struct miniflow *dst, struct miniflow *);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
void miniflow_destroy(struct miniflow *);
|
|
|
|
|
|
|
|
|
|
void miniflow_expand(const struct miniflow *, struct flow *);
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline uint64_t flow_u64_value(const struct flow *flow, size_t index)
|
2014-06-13 10:38:05 -07:00
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return ((uint64_t *)(flow))[index];
|
2014-06-13 10:38:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline uint64_t *flow_u64_lvalue(struct flow *flow, size_t index)
|
2014-06-13 10:38:05 -07:00
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return &((uint64_t *)(flow))[index];
|
2014-06-13 10:38:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
2014-05-28 16:56:29 -07:00
|
|
|
|
static inline bool
|
2015-01-06 11:10:42 -08:00
|
|
|
|
flow_get_next_in_map(const struct flow *flow, uint64_t map, uint64_t *value)
|
2014-04-29 15:50:39 -07:00
|
|
|
|
{
|
|
|
|
|
if (map) {
|
2015-01-06 11:10:42 -08:00
|
|
|
|
*value = flow_u64_value(flow, raw_ctz(map));
|
2014-04-29 15:50:39 -07:00
|
|
|
|
return true;
|
|
|
|
|
}
|
|
|
|
|
return false;
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Iterate through all flow u64 values specified by 'MAP'. */
|
2014-05-28 16:56:29 -07:00
|
|
|
|
#define FLOW_FOR_EACH_IN_MAP(VALUE, FLOW, MAP) \
|
|
|
|
|
for (uint64_t map__ = (MAP); \
|
|
|
|
|
flow_get_next_in_map(FLOW, map__, &(VALUE)); \
|
|
|
|
|
map__ = zero_rightmost_1bit(map__))
|
2014-04-29 15:50:39 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Iterate through all struct flow u64 indices specified by 'MAP'. */
|
|
|
|
|
#define MAP_FOR_EACH_INDEX(U64IDX, MAP) \
|
2014-11-21 11:23:25 -08:00
|
|
|
|
for (uint64_t map__ = (MAP); \
|
2015-01-06 11:10:42 -08:00
|
|
|
|
map__ && ((U64IDX) = raw_ctz(map__), true); \
|
2014-06-13 10:38:05 -07:00
|
|
|
|
map__ = zero_rightmost_1bit(map__))
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define FLOW_U64_SIZE(FIELD) \
|
|
|
|
|
DIV_ROUND_UP(sizeof(((struct flow *)0)->FIELD), sizeof(uint64_t))
|
2014-04-29 15:50:38 -07:00
|
|
|
|
|
|
|
|
|
#define MINIFLOW_MAP(FIELD) \
|
2015-01-06 11:10:42 -08:00
|
|
|
|
(((UINT64_C(1) << FLOW_U64_SIZE(FIELD)) - 1) \
|
|
|
|
|
<< (offsetof(struct flow, FIELD) / sizeof(uint64_t)))
|
2014-04-29 15:50:38 -07:00
|
|
|
|
|
2014-10-07 12:59:14 -07:00
|
|
|
|
struct mf_for_each_in_map_aux {
|
2015-01-06 11:10:42 -08:00
|
|
|
|
const uint64_t *values;
|
2014-10-07 12:59:14 -07:00
|
|
|
|
uint64_t fmap;
|
|
|
|
|
uint64_t map;
|
|
|
|
|
};
|
2014-04-18 08:26:56 -07:00
|
|
|
|
|
2014-10-07 12:59:14 -07:00
|
|
|
|
static inline bool
|
2015-01-06 11:10:42 -08:00
|
|
|
|
mf_get_next_in_map(struct mf_for_each_in_map_aux *aux, uint64_t *value)
|
2014-10-07 12:59:14 -07:00
|
|
|
|
{
|
|
|
|
|
if (aux->map) {
|
|
|
|
|
uint64_t rm1bit = rightmost_1bit(aux->map);
|
|
|
|
|
aux->map -= rm1bit;
|
|
|
|
|
|
|
|
|
|
if (aux->fmap & rm1bit) {
|
|
|
|
|
/* Advance 'aux->values' to point to the value for 'rm1bit'. */
|
|
|
|
|
uint64_t trash = aux->fmap & (rm1bit - 1);
|
|
|
|
|
if (trash) {
|
|
|
|
|
aux->fmap -= trash;
|
|
|
|
|
aux->values += count_1bits(trash);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Retrieve the value for 'rm1bit' then advance past it. */
|
|
|
|
|
aux->fmap -= rm1bit;
|
|
|
|
|
*value = *aux->values++;
|
|
|
|
|
} else {
|
|
|
|
|
*value = 0;
|
2014-04-18 08:26:56 -07:00
|
|
|
|
}
|
2014-10-07 12:59:14 -07:00
|
|
|
|
return true;
|
|
|
|
|
} else {
|
|
|
|
|
return false;
|
2014-04-18 08:26:56 -07:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Iterate through all miniflow u64 values specified by 'MAP'. */
|
2014-04-18 08:26:56 -07:00
|
|
|
|
#define MINIFLOW_FOR_EACH_IN_MAP(VALUE, FLOW, MAP) \
|
2014-10-07 12:59:14 -07:00
|
|
|
|
for (struct mf_for_each_in_map_aux aux__ \
|
2015-01-06 11:10:42 -08:00
|
|
|
|
= { miniflow_get_values(FLOW), (FLOW)->map, MAP }; \
|
2014-10-07 12:59:14 -07:00
|
|
|
|
mf_get_next_in_map(&aux__, &(VALUE)); \
|
|
|
|
|
)
|
2014-04-18 08:26:56 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* This can be used when it is known that 'u64_idx' is set in 'map'. */
|
|
|
|
|
static inline uint64_t
|
|
|
|
|
miniflow_values_get__(const uint64_t *values, uint64_t map, int u64_idx)
|
2014-11-26 15:17:26 -08:00
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return values[count_1bits(map & ((UINT64_C(1) << u64_idx) - 1))];
|
2014-11-26 15:17:26 -08:00
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* This can be used when it is known that 'u64_idx' is set in
|
2014-11-26 15:17:26 -08:00
|
|
|
|
* the map of 'mf'. */
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline uint64_t
|
|
|
|
|
miniflow_get__(const struct miniflow *mf, int u64_idx)
|
2014-11-26 15:17:26 -08:00
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return miniflow_values_get__(miniflow_get_values(mf), mf->map, u64_idx);
|
2014-11-26 15:17:26 -08:00
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Get the value of 'FIELD' of an up to 8 byte wide integer type 'TYPE' of
|
2014-04-29 15:50:38 -07:00
|
|
|
|
* a miniflow. */
|
|
|
|
|
#define MINIFLOW_GET_TYPE(MF, TYPE, OFS) \
|
2015-01-06 11:10:42 -08:00
|
|
|
|
(((MF)->map & (UINT64_C(1) << (OFS) / sizeof(uint64_t))) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
? ((OVS_FORCE const TYPE *) \
|
2015-01-06 11:10:42 -08:00
|
|
|
|
(miniflow_get_values(MF) \
|
|
|
|
|
+ count_1bits((MF)->map & \
|
|
|
|
|
((UINT64_C(1) << (OFS) / sizeof(uint64_t)) - 1)))) \
|
|
|
|
|
[(OFS) % sizeof(uint64_t) / sizeof(TYPE)] \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
: 0) \
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_U8(FLOW, FIELD) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, uint8_t, offsetof(struct flow, FIELD))
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_U16(FLOW, FIELD) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, uint16_t, offsetof(struct flow, FIELD))
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_BE16(FLOW, FIELD) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, ovs_be16, offsetof(struct flow, FIELD))
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_U32(FLOW, FIELD) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, uint32_t, offsetof(struct flow, FIELD))
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_BE32(FLOW, FIELD) \
|
2014-04-29 15:50:38 -07:00
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, ovs_be32, offsetof(struct flow, FIELD))
|
2015-01-06 11:10:42 -08:00
|
|
|
|
#define MINIFLOW_GET_U64(FLOW, FIELD) \
|
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, uint64_t, offsetof(struct flow, FIELD))
|
|
|
|
|
#define MINIFLOW_GET_BE64(FLOW, FIELD) \
|
|
|
|
|
MINIFLOW_GET_TYPE(FLOW, ovs_be64, offsetof(struct flow, FIELD))
|
|
|
|
|
|
|
|
|
|
static inline uint64_t miniflow_get(const struct miniflow *,
|
|
|
|
|
unsigned int u64_ofs);
|
|
|
|
|
static inline uint32_t miniflow_get_u32(const struct miniflow *,
|
|
|
|
|
unsigned int u32_ofs);
|
|
|
|
|
static inline ovs_be32 miniflow_get_be32(const struct miniflow *,
|
|
|
|
|
unsigned int be32_ofs);
|
2014-04-18 08:26:56 -07:00
|
|
|
|
static inline uint16_t miniflow_get_vid(const struct miniflow *);
|
|
|
|
|
static inline uint16_t miniflow_get_tcp_flags(const struct miniflow *);
|
2013-09-25 15:07:21 -07:00
|
|
|
|
static inline ovs_be64 miniflow_get_metadata(const struct miniflow *);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
|
|
|
|
bool miniflow_equal(const struct miniflow *a, const struct miniflow *b);
|
|
|
|
|
bool miniflow_equal_in_minimask(const struct miniflow *a,
|
|
|
|
|
const struct miniflow *b,
|
|
|
|
|
const struct minimask *);
|
|
|
|
|
bool miniflow_equal_flow_in_minimask(const struct miniflow *a,
|
|
|
|
|
const struct flow *b,
|
|
|
|
|
const struct minimask *);
|
2014-04-18 08:26:57 -07:00
|
|
|
|
uint32_t miniflow_hash_5tuple(const struct miniflow *flow, uint32_t basis);
|
2013-12-20 08:16:31 -08:00
|
|
|
|
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
|
|
|
|
/* Compressed flow wildcards. */
|
|
|
|
|
|
|
|
|
|
/* A sparse representation of a "struct flow_wildcards".
|
|
|
|
|
*
|
2013-12-20 08:16:31 -08:00
|
|
|
|
* See the large comment on struct miniflow for details.
|
|
|
|
|
*
|
|
|
|
|
* Note: While miniflow can have zero data for a 1-bit in the map,
|
|
|
|
|
* a minimask may not! We rely on this in the implementation. */
|
2012-09-04 12:43:53 -07:00
|
|
|
|
struct minimask {
|
|
|
|
|
struct miniflow masks;
|
|
|
|
|
};
|
|
|
|
|
|
|
|
|
|
void minimask_init(struct minimask *, const struct flow_wildcards *);
|
|
|
|
|
void minimask_clone(struct minimask *, const struct minimask *);
|
2013-08-27 12:25:48 -07:00
|
|
|
|
void minimask_move(struct minimask *dst, struct minimask *src);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
void minimask_combine(struct minimask *dst,
|
|
|
|
|
const struct minimask *a, const struct minimask *b,
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint64_t storage[FLOW_U64S]);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
void minimask_destroy(struct minimask *);
|
|
|
|
|
|
|
|
|
|
void minimask_expand(const struct minimask *, struct flow_wildcards *);
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
static inline uint32_t minimask_get_u32(const struct minimask *,
|
|
|
|
|
unsigned int u32_ofs);
|
|
|
|
|
static inline ovs_be32 minimask_get_be32(const struct minimask *,
|
|
|
|
|
unsigned int be32_ofs);
|
2014-04-18 08:26:56 -07:00
|
|
|
|
static inline uint16_t minimask_get_vid_mask(const struct minimask *);
|
2013-09-25 15:07:21 -07:00
|
|
|
|
static inline ovs_be64 minimask_get_metadata_mask(const struct minimask *);
|
2012-09-04 12:43:53 -07:00
|
|
|
|
|
|
|
|
|
bool minimask_equal(const struct minimask *a, const struct minimask *b);
|
|
|
|
|
bool minimask_has_extra(const struct minimask *, const struct minimask *);
|
2014-04-29 15:50:38 -07:00
|
|
|
|
|
2014-04-29 15:50:39 -07:00
|
|
|
|
|
2014-04-29 15:50:38 -07:00
|
|
|
|
/* Returns true if 'mask' matches every packet, false if 'mask' fixes any bits
|
|
|
|
|
* or fields. */
|
|
|
|
|
static inline bool
|
|
|
|
|
minimask_is_catchall(const struct minimask *mask)
|
|
|
|
|
{
|
|
|
|
|
/* For every 1-bit in mask's map, the corresponding value is non-zero,
|
|
|
|
|
* so the only way the mask can not fix any bits or fields is for the
|
|
|
|
|
* map the be zero. */
|
|
|
|
|
return mask->masks.map == 0;
|
|
|
|
|
}
|
2014-04-29 15:50:38 -07:00
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Returns the uint64_t that would be at byte offset '8 * u64_ofs' if 'flow'
|
|
|
|
|
* were expanded into a "struct flow". */
|
|
|
|
|
static inline uint64_t miniflow_get(const struct miniflow *flow,
|
|
|
|
|
unsigned int u64_ofs)
|
|
|
|
|
{
|
|
|
|
|
return flow->map & (UINT64_C(1) << u64_ofs)
|
|
|
|
|
? miniflow_get__(flow, u64_ofs) : 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline uint32_t miniflow_get_u32(const struct miniflow *flow,
|
|
|
|
|
unsigned int u32_ofs)
|
|
|
|
|
{
|
|
|
|
|
uint64_t value = miniflow_get(flow, u32_ofs / 2);
|
|
|
|
|
|
|
|
|
|
#if WORDS_BIGENDIAN
|
|
|
|
|
return (u32_ofs & 1) ? value : value >> 32;
|
|
|
|
|
#else
|
|
|
|
|
return (u32_ofs & 1) ? value >> 32 : value;
|
|
|
|
|
#endif
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline ovs_be32 miniflow_get_be32(const struct miniflow *flow,
|
|
|
|
|
unsigned int be32_ofs)
|
|
|
|
|
{
|
|
|
|
|
return (OVS_FORCE ovs_be32)miniflow_get_u32(flow, be32_ofs);
|
|
|
|
|
}
|
|
|
|
|
|
2014-04-18 08:26:56 -07:00
|
|
|
|
/* Returns the VID within the vlan_tci member of the "struct flow" represented
|
|
|
|
|
* by 'flow'. */
|
|
|
|
|
static inline uint16_t
|
|
|
|
|
miniflow_get_vid(const struct miniflow *flow)
|
|
|
|
|
{
|
2014-04-29 15:50:38 -07:00
|
|
|
|
ovs_be16 tci = MINIFLOW_GET_BE16(flow, vlan_tci);
|
2014-04-18 08:26:56 -07:00
|
|
|
|
return vlan_tci_to_vid(tci);
|
|
|
|
|
}
|
|
|
|
|
|
2015-01-06 11:10:42 -08:00
|
|
|
|
/* Returns the uint32_t that would be at byte offset '4 * u32_ofs' if 'mask'
|
|
|
|
|
* were expanded into a "struct flow_wildcards". */
|
|
|
|
|
static inline uint32_t
|
|
|
|
|
minimask_get_u32(const struct minimask *mask, unsigned int u32_ofs)
|
|
|
|
|
{
|
|
|
|
|
return miniflow_get_u32(&mask->masks, u32_ofs);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline ovs_be32
|
|
|
|
|
minimask_get_be32(const struct minimask *mask, unsigned int be32_ofs)
|
|
|
|
|
{
|
|
|
|
|
return (OVS_FORCE ovs_be32)minimask_get_u32(mask, be32_ofs);
|
|
|
|
|
}
|
|
|
|
|
|
2014-04-18 08:26:56 -07:00
|
|
|
|
/* Returns the VID mask within the vlan_tci member of the "struct
|
|
|
|
|
* flow_wildcards" represented by 'mask'. */
|
|
|
|
|
static inline uint16_t
|
|
|
|
|
minimask_get_vid_mask(const struct minimask *mask)
|
|
|
|
|
{
|
|
|
|
|
return miniflow_get_vid(&mask->masks);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Returns the value of the "tcp_flags" field in 'flow'. */
|
|
|
|
|
static inline uint16_t
|
|
|
|
|
miniflow_get_tcp_flags(const struct miniflow *flow)
|
|
|
|
|
{
|
2014-04-29 15:50:38 -07:00
|
|
|
|
return ntohs(MINIFLOW_GET_BE16(flow, tcp_flags));
|
2014-04-18 08:26:56 -07:00
|
|
|
|
}
|
|
|
|
|
|
2013-09-25 15:07:21 -07:00
|
|
|
|
/* Returns the value of the OpenFlow 1.1+ "metadata" field in 'flow'. */
|
|
|
|
|
static inline ovs_be64
|
|
|
|
|
miniflow_get_metadata(const struct miniflow *flow)
|
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return MINIFLOW_GET_BE64(flow, metadata);
|
2013-09-25 15:07:21 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* Returns the mask for the OpenFlow 1.1+ "metadata" field in 'mask'.
|
|
|
|
|
*
|
|
|
|
|
* The return value is all-1-bits if 'mask' matches on the whole value of the
|
|
|
|
|
* metadata field, all-0-bits if 'mask' entirely wildcards the metadata field,
|
|
|
|
|
* or some other value if the metadata field is partially matched, partially
|
|
|
|
|
* wildcarded. */
|
|
|
|
|
static inline ovs_be64
|
|
|
|
|
minimask_get_metadata_mask(const struct minimask *mask)
|
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
return MINIFLOW_GET_BE64(&mask->masks, metadata);
|
2013-09-25 15:07:21 -07:00
|
|
|
|
}
|
2012-08-07 13:43:18 -07:00
|
|
|
|
|
2014-04-29 15:50:38 -07:00
|
|
|
|
/* Perform a bitwise OR of miniflow 'src' flow data with the equivalent
|
|
|
|
|
* fields in 'dst', storing the result in 'dst'. */
|
|
|
|
|
static inline void
|
|
|
|
|
flow_union_with_miniflow(struct flow *dst, const struct miniflow *src)
|
|
|
|
|
{
|
2015-01-06 11:10:42 -08:00
|
|
|
|
uint64_t *dst_u64 = (uint64_t *) dst;
|
|
|
|
|
const uint64_t *p = miniflow_get_values(src);
|
2014-11-26 15:30:33 -08:00
|
|
|
|
int idx;
|
2014-04-29 15:50:38 -07:00
|
|
|
|
|
2014-11-26 15:30:33 -08:00
|
|
|
|
MAP_FOR_EACH_INDEX(idx, src->map) {
|
2015-01-06 11:10:42 -08:00
|
|
|
|
dst_u64[idx] |= *p++;
|
2014-04-29 15:50:38 -07:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2014-04-18 08:26:56 -07:00
|
|
|
|
static inline struct pkt_metadata
|
|
|
|
|
pkt_metadata_from_flow(const struct flow *flow)
|
|
|
|
|
{
|
|
|
|
|
struct pkt_metadata md;
|
|
|
|
|
|
|
|
|
|
md.recirc_id = flow->recirc_id;
|
|
|
|
|
md.dp_hash = flow->dp_hash;
|
|
|
|
|
md.tunnel = flow->tunnel;
|
|
|
|
|
md.skb_priority = flow->skb_priority;
|
|
|
|
|
md.pkt_mark = flow->pkt_mark;
|
|
|
|
|
md.in_port = flow->in_port;
|
|
|
|
|
|
|
|
|
|
return md;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool is_ip_any(const struct flow *flow)
|
|
|
|
|
{
|
|
|
|
|
return dl_type_is_ip_any(flow->dl_type);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool is_icmpv4(const struct flow *flow)
|
|
|
|
|
{
|
|
|
|
|
return (flow->dl_type == htons(ETH_TYPE_IP)
|
|
|
|
|
&& flow->nw_proto == IPPROTO_ICMP);
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static inline bool is_icmpv6(const struct flow *flow)
|
|
|
|
|
{
|
|
|
|
|
return (flow->dl_type == htons(ETH_TYPE_IPV6)
|
|
|
|
|
&& flow->nw_proto == IPPROTO_ICMPV6);
|
|
|
|
|
}
|
|
|
|
|
|
2014-04-24 13:18:18 -07:00
|
|
|
|
static inline bool is_stp(const struct flow *flow)
|
|
|
|
|
{
|
|
|
|
|
return (eth_addr_equals(flow->dl_dst, eth_addr_stp)
|
|
|
|
|
&& flow->dl_type == htons(FLOW_DL_TYPE_NONE));
|
|
|
|
|
}
|
|
|
|
|
|
2009-07-08 13:19:16 -07:00
|
|
|
|
#endif /* flow.h */
|