/*
- * Copyright (c) 2008, 2009, 2010, 2011, 2012 Nicira, Inc.
+ * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
#ifndef OPENFLOW_NICIRA_EXT_H
#define OPENFLOW_NICIRA_EXT_H 1
-#include "openflow/openflow.h"
-#include "openvswitch/types.h"
+#include <openflow/openflow.h>
+#include <openvswitch/types.h>
/* The following vendor extensions, proposed by Nicira, are not yet
* standardized, so they are not included in openflow.h. Some of them may be
* suitable for standardization; others we never expect to standardize. */
-#define NX_VENDOR_ID 0x00002320
\f
/* Nicira vendor-specific error messages extension.
*
\f
/* Nicira vendor requests and replies. */
-/* Header for Nicira vendor requests and replies. */
-struct nicira_header {
- struct ofp_header header;
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be32 subtype; /* See the NXT numbers in ofp-msgs.h. */
-};
-OFP_ASSERT(sizeof(struct nicira_header) == 16);
-
-/* Header for Nicira vendor stats request and reply messages in OpenFlow
- * 1.0. */
-struct nicira10_stats_msg {
- struct ofp10_vendor_stats_msg vsm; /* Vendor NX_VENDOR_ID. */
- ovs_be32 subtype; /* One of NXST_* below. */
- uint8_t pad[4]; /* Align to 64-bits. */
-};
-OFP_ASSERT(sizeof(struct nicira10_stats_msg) == 24);
-
-/* Header for Nicira vendor stats request and reply messages in OpenFlow
- * 1.1. */
-struct nicira11_stats_msg {
- struct ofp11_vendor_stats_msg vsm; /* Vendor NX_VENDOR_ID. */
- ovs_be32 subtype; /* One of NXST_* below. */
-};
-OFP_ASSERT(sizeof(struct nicira11_stats_msg) == 24);
-
/* Fields to use when hashing flows. */
enum nx_hash_fields {
/* Ethernet source address (NXM_OF_ETH_SRC) only. */
* - NXM_OF_IP_SRC / NXM_OF_IP_DST
* - NXM_OF_TCP_SRC / NXM_OF_TCP_DST
*/
- NX_HASH_FIELDS_SYMMETRIC_L4
+ NX_HASH_FIELDS_SYMMETRIC_L4,
+
+ /* L3+L4 only, including the following fields:
+ *
+ * - NXM_OF_IP_PROTO
+ * - NXM_OF_IP_SRC / NXM_OF_IP_DST
+ * - NXM_OF_SCTP_SRC / NXM_OF_SCTP_DST
+ * - NXM_OF_TCP_SRC / NXM_OF_TCP_DST
+ */
+ NX_HASH_FIELDS_SYMMETRIC_L3L4,
+
+ /* L3+L4 only with UDP ports, including the following fields:
+ *
+ * - NXM_OF_IP_PROTO
+ * - NXM_OF_IP_SRC / NXM_OF_IP_DST
+ * - NXM_OF_SCTP_SRC / NXM_OF_SCTP_DST
+ * - NXM_OF_TCP_SRC / NXM_OF_TCP_DST
+ * - NXM_OF_UDP_SRC / NXM_OF_UDP_DST
+ */
+ NX_HASH_FIELDS_SYMMETRIC_L3L4_UDP
+
+
};
/* This command enables or disables an Open vSwitch extension that allows a
* controller to specify the OpenFlow table to which a flow should be added,
* instead of having the switch decide which table is most appropriate as
- * required by OpenFlow 1.0. By default, the extension is disabled.
+ * required by OpenFlow 1.0. Because NXM was designed as an extension to
+ * OpenFlow 1.0, the extension applies equally to ofp10_flow_mod and
+ * nx_flow_mod. By default, the extension is disabled.
*
* When this feature is enabled, Open vSwitch treats struct ofp10_flow_mod's
- * 16-bit 'command' member as two separate fields. The upper 8 bits are used
- * as the table ID, the lower 8 bits specify the command as usual. A table ID
- * of 0xff is treated like a wildcarded table ID.
+ * and struct nx_flow_mod's 16-bit 'command' member as two separate fields.
+ * The upper 8 bits are used as the table ID, the lower 8 bits specify the
+ * command as usual. A table ID of 0xff is treated like a wildcarded table ID.
*
* The specific treatment of the table ID depends on the type of flow mod:
*
OFP_ASSERT(sizeof(struct nx_flow_mod_table_id) == 8);
enum nx_packet_in_format {
- NXPIF_OPENFLOW10 = 0, /* Standard OpenFlow 1.0 compatible. */
- NXPIF_NXM = 1 /* Nicira Extended. */
+ NXPIF_STANDARD = 0, /* OFPT_PACKET_IN for this OpenFlow version. */
+ NXPIF_NXT_PACKET_IN = 1, /* NXT_PACKET_IN (since OVS v1.1). */
+ NXPIF_NXT_PACKET_IN2 = 2, /* NXT_PACKET_IN2 (since OVS v2.6). */
};
-/* NXT_SET_PACKET_IN_FORMAT request. */
+/* NXT_SET_PACKET_IN_FORMAT request.
+ *
+ * For any given OpenFlow version, Open vSwitch supports multiple formats for
+ * "packet-in" messages. The default is always the standard format for the
+ * OpenFlow version in question, but NXT_SET_PACKET_IN_FORMAT can be used to
+ * set an alternative format.
+ *
+ * From OVS v1.1 to OVS v2.5, this request was only honored for OpenFlow 1.0.
+ * Requests to set format NXPIF_NXT_PACKET_IN were accepted for OF1.1+ but they
+ * had no effect. (Requests to set formats other than NXPIF_STANDARD or
+ * NXPIF_NXT_PACKET_IN were rejected with OFPBRC_EPERM.)
+ *
+ * From OVS v2.6 onward, this request is honored for all OpenFlow versions.
+ */
struct nx_set_packet_in_format {
ovs_be32 format; /* One of NXPIF_*. */
};
* might support fields (new registers, new protocols, etc.) that the
* controller does not. The controller must prepared to tolerate these.
*
- * The 'cookie' and 'table_id' fields have no meaning when 'reason' is
- * OFPR_NO_MATCH. In this case they should be set to 0. */
+ * The 'cookie' field has no meaning when 'reason' is OFPR_NO_MATCH. In this
+ * case it should be UINT64_MAX. */
struct nx_packet_in {
ovs_be32 buffer_id; /* ID assigned by datapath. */
ovs_be16 total_len; /* Full length of frame. */
};
OFP_ASSERT(sizeof(struct nx_packet_in) == 24);
+/* NXT_PACKET_IN2
+ * ==============
+ *
+ * NXT_PACKET_IN2 is conceptually similar to OFPT_PACKET_IN but it is expressed
+ * as an extensible set of properties instead of using a fixed structure.
+ *
+ * Added in Open vSwitch 2.6
+ *
+ *
+ * Continuations
+ * -------------
+ *
+ * When a "controller" action specifies the "pause" flag, the controller action
+ * freezes the packet's trip through Open vSwitch flow tables and serializes
+ * that state into the packet-in message as a "continuation". The controller
+ * can later send the continuation back to the switch, which will restart the
+ * packet's traversal from the point where it was interrupted. This permits an
+ * OpenFlow controller to interpose on a packet midway through processing in
+ * Open vSwitch.
+ *
+ * Continuations fit into packet processing this way:
+ *
+ * 1. A packet ingresses into Open vSwitch, which runs it through the OpenFlow
+ * tables.
+ *
+ * 2. An OpenFlow flow executes a "controller" action that includes the "pause"
+ * flag. Open vSwitch serializes the packet processing state and sends it,
+ * as an NXT_PACKET_IN2 that includes an additional NXPINT_CONTINUATION
+ * property (the continuation), to the OpenFlow controller.
+ *
+ * (The controller must use NXAST_CONTROLLER2 to generate the packet-in,
+ * because only this form of the "controller" action has a "pause" flag.
+ * Similarly, the controller must use NXT_SET_PACKET_IN_FORMAT to select
+ * NXT_PACKET_IN2 as the packet-in format, because this is the only format
+ * that supports continuation passing.)
+ *
+ * 3. The controller receives the NXT_PACKET_IN2 and processes it. The
+ * controller can interpret and, if desired, modify some of the contents of
+ * the packet-in, such as the packet and the metadata being processed.
+ *
+ * 4. The controller sends the continuation back to the switch, using an
+ * NXT_RESUME message. Packet processing resumes where it left off.
+ *
+ * The controller might change the pipeline configuration concurrently with
+ * steps 2 through 4. For example, it might add or remove OpenFlow flows. If
+ * that happens, then the packet will experience a mix of processing from the
+ * two configurations, that is, the initial processing (before
+ * NXAST_CONTROLLER2) uses the initial flow table, and the later processing
+ * (after NXT_RESUME) uses the later flow table. This means that the
+ * controller needs to take care to avoid incompatible pipeline changes while
+ * processing continuations.
+ *
+ * External side effects (e.g. "output") of OpenFlow actions processed before
+ * NXAST_CONTROLLER2 is encountered might be executed during step 2 or step 4,
+ * and the details may vary among Open vSwitch features and versions. Thus, a
+ * controller that wants to make sure that side effects are executed must pass
+ * the continuation back to the switch, that is, must not skip step 4.
+ *
+ * Architecturally, continuations may be "stateful" or "stateless", that is,
+ * they may or may not refer to buffered state maintained in Open vSwitch.
+ * This means that a controller should not attempt to resume a given
+ * continuations more than once (because the switch might have discarded the
+ * buffered state after the first use). For the same reason, continuations
+ * might become "stale" if the controller takes too long to resume them
+ * (because the switch might have discarded old buffered state). Taken
+ * together with the previous note, this means that a controller should resume
+ * each continuation exactly once (and promptly).
+ *
+ * Without the information in NXPINT_CONTINUATION, the controller can (with
+ * careful design, and help from the flow cookie) determine where the packet is
+ * in the pipeline, but in the general case it can't determine what nested
+ * "resubmit"s that may be in progress, or what data is on the stack maintained
+ * by NXAST_STACK_PUSH and NXAST_STACK_POP actions, what is in the OpenFlow
+ * action set, etc.
+ *
+ * Continuations are expensive because they require a round trip between the
+ * switch and the controller. Thus, they should not be used to implement
+ * processing that needs to happen at "line rate".
+ *
+ * The contents of NXPINT_CONTINUATION are private to the switch, may change
+ * unpredictably from one version of Open vSwitch to another, and are not
+ * documented here. The contents are also tied to a given Open vSwitch process
+ * and bridge, so that restarting Open vSwitch or deleting and recreating a
+ * bridge will cause the corresponding NXT_RESUME to be rejected.
+ *
+ * In the current implementation, Open vSwitch forks the packet processing
+ * pipeline across patch ports. Suppose, for example, that the pipeline for
+ * br0 outputs to a patch port whose peer belongs to br1, and that the pipeline
+ * for br1 executes a controller action with the "pause" flag. This only
+ * pauses processing within br1, and processing in br0 continues and possibly
+ * completes with visible side effects, such as outputting to ports, before
+ * br1's controller receives or processes the continuation. This
+ * implementation maintains the independence of separate bridges and, since
+ * processing in br1 cannot affect the behavior of br0 anyway, should not cause
+ * visible behavioral changes.
+ *
+ * A stateless implementation of continuations may ignore the "controller"
+ * action max_len, always sending the whole packet, because the full packet is
+ * required to continue traversal.
+ */
+enum nx_packet_in2_prop_type {
+ /* Packet. */
+ NXPINT_PACKET, /* Raw packet data. */
+ NXPINT_FULL_LEN, /* ovs_be32: Full packet len, if truncated. */
+ NXPINT_BUFFER_ID, /* ovs_be32: Buffer ID, if buffered. */
+
+ /* Information about the flow that triggered the packet-in. */
+ NXPINT_TABLE_ID, /* uint8_t: Table ID. */
+ NXPINT_COOKIE, /* ovs_be64: Flow cookie. */
+
+ /* Other. */
+ NXPINT_REASON, /* uint8_t, one of OFPR_*. */
+ NXPINT_METADATA, /* NXM or OXM for metadata fields. */
+ NXPINT_USERDATA, /* From NXAST_CONTROLLER2 userdata. */
+ NXPINT_CONTINUATION, /* Private data for continuing processing. */
+};
+
/* Configures the "role" of the sending controller. The default role is:
*
* - Other (NX_ROLE_OTHER), which allows the controller access to all
};
OFP_ASSERT(sizeof(struct nx_async_config) == 24);
\f
-/* Nicira vendor flow actions. */
-
-enum nx_action_subtype {
- NXAST_SNAT__OBSOLETE, /* No longer used. */
- NXAST_RESUBMIT, /* struct nx_action_resubmit */
- NXAST_SET_TUNNEL, /* struct nx_action_set_tunnel */
- NXAST_DROP_SPOOFED_ARP__OBSOLETE,
- NXAST_SET_QUEUE, /* struct nx_action_set_queue */
- NXAST_POP_QUEUE, /* struct nx_action_pop_queue */
- NXAST_REG_MOVE, /* struct nx_action_reg_move */
- NXAST_REG_LOAD, /* struct nx_action_reg_load */
- NXAST_NOTE, /* struct nx_action_note */
- NXAST_SET_TUNNEL64, /* struct nx_action_set_tunnel64 */
- NXAST_MULTIPATH, /* struct nx_action_multipath */
- NXAST_AUTOPATH__DEPRECATED, /* struct nx_action_autopath */
- NXAST_BUNDLE, /* struct nx_action_bundle */
- NXAST_BUNDLE_LOAD, /* struct nx_action_bundle */
- NXAST_RESUBMIT_TABLE, /* struct nx_action_resubmit */
- NXAST_OUTPUT_REG, /* struct nx_action_output_reg */
- NXAST_LEARN, /* struct nx_action_learn */
- NXAST_EXIT, /* struct nx_action_header */
- NXAST_DEC_TTL, /* struct nx_action_header */
- NXAST_FIN_TIMEOUT, /* struct nx_action_fin_timeout */
- NXAST_CONTROLLER, /* struct nx_action_controller */
- NXAST_DEC_TTL_CNT_IDS, /* struct nx_action_cnt_ids */
-};
-
-/* Header for Nicira-defined actions. */
-struct nx_action_header {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_*. */
- uint8_t pad[6];
-};
-OFP_ASSERT(sizeof(struct nx_action_header) == 16);
-
-/* Action structures for NXAST_RESUBMIT and NXAST_RESUBMIT_TABLE.
- *
- * These actions search one of the switch's flow tables:
- *
- * - For NXAST_RESUBMIT_TABLE only, if the 'table' member is not 255, then
- * it specifies the table to search.
- *
- * - Otherwise (for NXAST_RESUBMIT_TABLE with a 'table' of 255, or for
- * NXAST_RESUBMIT regardless of 'table'), it searches the current flow
- * table, that is, the OpenFlow flow table that contains the flow from
- * which this action was obtained. If this action did not come from a
- * flow table (e.g. it came from an OFPT_PACKET_OUT message), then table 0
- * is the current table.
- *
- * The flow table lookup uses a flow that may be slightly modified from the
- * original lookup:
- *
- * - For NXAST_RESUBMIT, the 'in_port' member of struct nx_action_resubmit
- * is used as the flow's in_port.
- *
- * - For NXAST_RESUBMIT_TABLE, if the 'in_port' member is not OFPP_IN_PORT,
- * then its value is used as the flow's in_port. Otherwise, the original
- * in_port is used.
- *
- * - If actions that modify the flow (e.g. OFPAT_SET_VLAN_VID) precede the
- * resubmit action, then the flow is updated with the new values.
- *
- * Following the lookup, the original in_port is restored.
- *
- * If the modified flow matched in the flow table, then the corresponding
- * actions are executed. Afterward, actions following the resubmit in the
- * original set of actions, if any, are executed; any changes made to the
- * packet (e.g. changes to VLAN) by secondary actions persist when those
- * actions are executed, although the original in_port is restored.
- *
- * Resubmit actions may be used any number of times within a set of actions.
- *
- * Resubmit actions may nest to an implementation-defined depth. Beyond this
- * implementation-defined depth, further resubmit actions are simply ignored.
- *
- * NXAST_RESUBMIT ignores 'table' and 'pad'. NXAST_RESUBMIT_TABLE requires
- * 'pad' to be all-bits-zero.
- *
- * Open vSwitch 1.0.1 and earlier did not support recursion. Open vSwitch
- * before 1.2.90 did not support NXAST_RESUBMIT_TABLE.
- */
-struct nx_action_resubmit {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_RESUBMIT. */
- ovs_be16 in_port; /* New in_port for checking flow table. */
- uint8_t table; /* NXAST_RESUBMIT_TABLE: table to use. */
- uint8_t pad[3];
-};
-OFP_ASSERT(sizeof(struct nx_action_resubmit) == 16);
-
-/* Action structure for NXAST_SET_TUNNEL.
- *
- * Sets the encapsulating tunnel ID to a 32-bit value. The most-significant 32
- * bits of the tunnel ID are set to 0. */
-struct nx_action_set_tunnel {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_SET_TUNNEL. */
- uint8_t pad[2];
- ovs_be32 tun_id; /* Tunnel ID. */
-};
-OFP_ASSERT(sizeof(struct nx_action_set_tunnel) == 16);
-
-/* Action structure for NXAST_SET_TUNNEL64.
- *
- * Sets the encapsulating tunnel ID to a 64-bit value. */
-struct nx_action_set_tunnel64 {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_SET_TUNNEL64. */
- uint8_t pad[6];
- ovs_be64 tun_id; /* Tunnel ID. */
-};
-OFP_ASSERT(sizeof(struct nx_action_set_tunnel64) == 24);
-
-/* Action structure for NXAST_SET_QUEUE.
- *
- * Set the queue that should be used when packets are output. This is similar
- * to the OpenFlow OFPAT_ENQUEUE action, but does not take the output port as
- * an argument. This allows the queue to be defined before the port is
- * known. */
-struct nx_action_set_queue {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_SET_QUEUE. */
- uint8_t pad[2];
- ovs_be32 queue_id; /* Where to enqueue packets. */
-};
-OFP_ASSERT(sizeof(struct nx_action_set_queue) == 16);
-
-/* Action structure for NXAST_POP_QUEUE.
- *
- * Restores the queue to the value it was before any NXAST_SET_QUEUE actions
- * were used. Only the original queue can be restored this way; no stack is
- * maintained. */
-struct nx_action_pop_queue {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_POP_QUEUE. */
- uint8_t pad[6];
-};
-OFP_ASSERT(sizeof(struct nx_action_pop_queue) == 16);
-
-/* Action structure for NXAST_REG_MOVE.
- *
- * Copies src[src_ofs:src_ofs+n_bits] to dst[dst_ofs:dst_ofs+n_bits], where
- * a[b:c] denotes the bits within 'a' numbered 'b' through 'c' (not including
- * bit 'c'). Bit numbering starts at 0 for the least-significant bit, 1 for
- * the next most significant bit, and so on.
- *
- * 'src' and 'dst' are nxm_header values with nxm_hasmask=0. (It doesn't make
- * sense to use nxm_hasmask=1 because the action does not do any kind of
- * matching; it uses the actual value of a field.)
- *
- * The following nxm_header values are potentially acceptable as 'src':
- *
- * - NXM_OF_IN_PORT
- * - NXM_OF_ETH_DST
- * - NXM_OF_ETH_SRC
- * - NXM_OF_ETH_TYPE
- * - NXM_OF_VLAN_TCI
- * - NXM_OF_IP_TOS
- * - NXM_OF_IP_PROTO
- * - NXM_OF_IP_SRC
- * - NXM_OF_IP_DST
- * - NXM_OF_TCP_SRC
- * - NXM_OF_TCP_DST
- * - NXM_OF_UDP_SRC
- * - NXM_OF_UDP_DST
- * - NXM_OF_ICMP_TYPE
- * - NXM_OF_ICMP_CODE
- * - NXM_OF_ARP_OP
- * - NXM_OF_ARP_SPA
- * - NXM_OF_ARP_TPA
- * - NXM_NX_TUN_ID
- * - NXM_NX_ARP_SHA
- * - NXM_NX_ARP_THA
- * - NXM_NX_ICMPV6_TYPE
- * - NXM_NX_ICMPV6_CODE
- * - NXM_NX_ND_SLL
- * - NXM_NX_ND_TLL
- * - NXM_NX_REG(idx) for idx in the switch's accepted range.
- *
- * The following nxm_header values are potentially acceptable as 'dst':
- *
- * - NXM_OF_ETH_DST
- * - NXM_OF_ETH_SRC
- * - NXM_OF_IP_TOS
- * - NXM_OF_IP_SRC
- * - NXM_OF_IP_DST
- * - NXM_OF_TCP_SRC
- * - NXM_OF_TCP_DST
- * - NXM_OF_UDP_SRC
- * - NXM_OF_UDP_DST
- * Modifying any of the above fields changes the corresponding packet
- * header.
- *
- * - NXM_NX_REG(idx) for idx in the switch's accepted range.
- *
- * - NXM_OF_VLAN_TCI. Modifying this field's value has side effects on the
- * packet's 802.1Q header. Setting a value with CFI=0 removes the 802.1Q
- * header (if any), ignoring the other bits. Setting a value with CFI=1
- * adds or modifies the 802.1Q header appropriately, setting the TCI field
- * to the field's new value (with the CFI bit masked out).
- *
- * - NXM_NX_TUN_ID. Modifying this value modifies the tunnel ID used for the
- * packet's next tunnel encapsulation.
- *
- * A given nxm_header value may be used as 'src' or 'dst' only on a flow whose
- * nx_match satisfies its prerequisites. For example, NXM_OF_IP_TOS may be
- * used only if the flow's nx_match includes an nxm_entry that specifies
- * nxm_type=NXM_OF_ETH_TYPE, nxm_hasmask=0, and nxm_value=0x0800.
- *
- * The switch will reject actions for which src_ofs+n_bits is greater than the
- * width of 'src' or dst_ofs+n_bits is greater than the width of 'dst' with
- * error type OFPET_BAD_ACTION, code OFPBAC_BAD_ARGUMENT.
- *
- * This action behaves properly when 'src' overlaps with 'dst', that is, it
- * behaves as if 'src' were copied out to a temporary buffer, then the
- * temporary buffer copied to 'dst'.
- */
-struct nx_action_reg_move {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_REG_MOVE. */
- ovs_be16 n_bits; /* Number of bits. */
- ovs_be16 src_ofs; /* Starting bit offset in source. */
- ovs_be16 dst_ofs; /* Starting bit offset in destination. */
- ovs_be32 src; /* Source register. */
- ovs_be32 dst; /* Destination register. */
-};
-OFP_ASSERT(sizeof(struct nx_action_reg_move) == 24);
-
-/* Action structure for NXAST_REG_LOAD.
- *
- * Copies value[0:n_bits] to dst[ofs:ofs+n_bits], where a[b:c] denotes the bits
- * within 'a' numbered 'b' through 'c' (not including bit 'c'). Bit numbering
- * starts at 0 for the least-significant bit, 1 for the next most significant
- * bit, and so on.
- *
- * 'dst' is an nxm_header with nxm_hasmask=0. See the documentation for
- * NXAST_REG_MOVE, above, for the permitted fields and for the side effects of
- * loading them.
- *
- * The 'ofs' and 'n_bits' fields are combined into a single 'ofs_nbits' field
- * to avoid enlarging the structure by another 8 bytes. To allow 'n_bits' to
- * take a value between 1 and 64 (inclusive) while taking up only 6 bits, it is
- * also stored as one less than its true value:
- *
- * 15 6 5 0
- * +------------------------------+------------------+
- * | ofs | n_bits - 1 |
- * +------------------------------+------------------+
- *
- * The switch will reject actions for which ofs+n_bits is greater than the
- * width of 'dst', or in which any bits in 'value' with value 2**n_bits or
- * greater are set to 1, with error type OFPET_BAD_ACTION, code
- * OFPBAC_BAD_ARGUMENT.
- */
-struct nx_action_reg_load {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_REG_LOAD. */
- ovs_be16 ofs_nbits; /* (ofs << 6) | (n_bits - 1). */
- ovs_be32 dst; /* Destination register. */
- ovs_be64 value; /* Immediate value. */
-};
-OFP_ASSERT(sizeof(struct nx_action_reg_load) == 24);
-
-/* Action structure for NXAST_NOTE.
- *
- * This action has no effect. It is variable length. The switch does not
- * attempt to interpret the user-defined 'note' data in any way. A controller
- * can use this action to attach arbitrary metadata to a flow.
- *
- * This action might go away in the future.
- */
-struct nx_action_note {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* A multiple of 8, but at least 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_NOTE. */
- uint8_t note[6]; /* Start of user-defined data. */
- /* Possibly followed by additional user-defined data. */
-};
-OFP_ASSERT(sizeof(struct nx_action_note) == 16);
-
-/* Action structure for NXAST_MULTIPATH.
- *
- * This action performs the following steps in sequence:
- *
- * 1. Hashes the fields designated by 'fields', one of NX_HASH_FIELDS_*.
- * Refer to the definition of "enum nx_mp_fields" for details.
- *
- * The 'basis' value is used as a universal hash parameter, that is,
- * different values of 'basis' yield different hash functions. The
- * particular universal hash function used is implementation-defined.
- *
- * The hashed fields' values are drawn from the current state of the
- * flow, including all modifications that have been made by actions up to
- * this point.
- *
- * 2. Applies the multipath link choice algorithm specified by 'algorithm',
- * one of NX_MP_ALG_*. Refer to the definition of "enum nx_mp_algorithm"
- * for details.
- *
- * The output of the algorithm is 'link', an unsigned integer less than
- * or equal to 'max_link'.
- *
- * Some algorithms use 'arg' as an additional argument.
- *
- * 3. Stores 'link' in dst[ofs:ofs+n_bits]. The format and semantics of
- * 'dst' and 'ofs_nbits' are similar to those for the NXAST_REG_LOAD
- * action.
- *
- * The switch will reject actions that have an unknown 'fields', or an unknown
- * 'algorithm', or in which ofs+n_bits is greater than the width of 'dst', or
- * in which 'max_link' is greater than or equal to 2**n_bits, with error type
- * OFPET_BAD_ACTION, code OFPBAC_BAD_ARGUMENT.
- */
-struct nx_action_multipath {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 32. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_MULTIPATH. */
-
- /* What fields to hash and how. */
- ovs_be16 fields; /* One of NX_HASH_FIELDS_*. */
- ovs_be16 basis; /* Universal hash parameter. */
- ovs_be16 pad0;
-
- /* Multipath link choice algorithm to apply to hash value. */
- ovs_be16 algorithm; /* One of NX_MP_ALG_*. */
- ovs_be16 max_link; /* Number of output links, minus 1. */
- ovs_be32 arg; /* Algorithm-specific argument. */
- ovs_be16 pad1;
-
- /* Where to store the result. */
- ovs_be16 ofs_nbits; /* (ofs << 6) | (n_bits - 1). */
- ovs_be32 dst; /* Destination. */
-};
-OFP_ASSERT(sizeof(struct nx_action_multipath) == 32);
-
-/* NXAST_MULTIPATH: Multipath link choice algorithm to apply.
- *
- * In the descriptions below, 'n_links' is max_link + 1. */
-enum nx_mp_algorithm {
- /* link = hash(flow) % n_links.
- *
- * Redistributes all traffic when n_links changes. O(1) performance. See
- * RFC 2992.
- *
- * Use UINT16_MAX for max_link to get a raw hash value. */
- NX_MP_ALG_MODULO_N,
-
- /* link = hash(flow) / (MAX_HASH / n_links).
- *
- * Redistributes between one-quarter and one-half of traffic when n_links
- * changes. O(1) performance. See RFC 2992.
- */
- NX_MP_ALG_HASH_THRESHOLD,
-
- /* for i in [0,n_links):
- * weights[i] = hash(flow, i)
- * link = { i such that weights[i] >= weights[j] for all j != i }
- *
- * Redistributes 1/n_links of traffic when n_links changes. O(n_links)
- * performance. If n_links is greater than a threshold (currently 64, but
- * subject to change), Open vSwitch will substitute another algorithm
- * automatically. See RFC 2992. */
- NX_MP_ALG_HRW, /* Highest Random Weight. */
-
- /* i = 0
- * repeat:
- * i = i + 1
- * link = hash(flow, i) % arg
- * while link > max_link
- *
- * Redistributes 1/n_links of traffic when n_links changes. O(1)
- * performance when arg/max_link is bounded by a constant.
- *
- * Redistributes all traffic when arg changes.
- *
- * arg must be greater than max_link and for best performance should be no
- * more than approximately max_link * 2. If arg is outside the acceptable
- * range, Open vSwitch will automatically substitute the least power of 2
- * greater than max_link.
- *
- * This algorithm is specific to Open vSwitch.
- */
- NX_MP_ALG_ITER_HASH /* Iterative Hash. */
-};
-\f
-/* Action structure for NXAST_LEARN.
- *
- * This action adds or modifies a flow in an OpenFlow table, similar to
- * OFPT_FLOW_MOD with OFPFC_MODIFY_STRICT as 'command'. The new flow has the
- * specified idle timeout, hard timeout, priority, cookie, and flags. The new
- * flow's match criteria and actions are built by applying each of the series
- * of flow_mod_spec elements included as part of the action.
- *
- * A flow_mod_spec starts with a 16-bit header. A header that is all-bits-0 is
- * a no-op used for padding the action as a whole to a multiple of 8 bytes in
- * length. Otherwise, the flow_mod_spec can be thought of as copying 'n_bits'
- * bits from a source to a destination. In this case, the header contains
- * multiple fields:
- *
- * 15 14 13 12 11 10 0
- * +------+---+------+---------------------------------+
- * | 0 |src| dst | n_bits |
- * +------+---+------+---------------------------------+
- *
- * The meaning and format of a flow_mod_spec depends on 'src' and 'dst'. The
- * following table summarizes the meaning of each possible combination.
- * Details follow the table:
- *
- * src dst meaning
- * --- --- ----------------------------------------------------------
- * 0 0 Add match criteria based on value in a field.
- * 1 0 Add match criteria based on an immediate value.
- * 0 1 Add NXAST_REG_LOAD action to copy field into a different field.
- * 1 1 Add NXAST_REG_LOAD action to load immediate value into a field.
- * 0 2 Add OFPAT_OUTPUT action to output to port from specified field.
- * All other combinations are undefined and not allowed.
- *
- * The flow_mod_spec header is followed by a source specification and a
- * destination specification. The format and meaning of the source
- * specification depends on 'src':
- *
- * - If 'src' is 0, the source bits are taken from a field in the flow to
- * which this action is attached. (This should be a wildcarded field. If
- * its value is fully specified then the source bits being copied have
- * constant values.)
- *
- * The source specification is an ovs_be32 'field' and an ovs_be16 'ofs'.
- * 'field' is an nxm_header with nxm_hasmask=0, and 'ofs' the starting bit
- * offset within that field. The source bits are field[ofs:ofs+n_bits-1].
- * 'field' and 'ofs' are subject to the same restrictions as the source
- * field in NXAST_REG_MOVE.
- *
- * - If 'src' is 1, the source bits are a constant value. The source
- * specification is (n_bits+15)/16*2 bytes long. Taking those bytes as a
- * number in network order, the source bits are the 'n_bits'
- * least-significant bits. The switch will report an error if other bits
- * in the constant are nonzero.
- *
- * The flow_mod_spec destination specification, for 'dst' of 0 or 1, is an
- * ovs_be32 'field' and an ovs_be16 'ofs'. 'field' is an nxm_header with
- * nxm_hasmask=0 and 'ofs' is a starting bit offset within that field. The
- * meaning of the flow_mod_spec depends on 'dst':
- *
- * - If 'dst' is 0, the flow_mod_spec specifies match criteria for the new
- * flow. The new flow matches only if bits field[ofs:ofs+n_bits-1] in a
- * packet equal the source bits. 'field' may be any nxm_header with
- * nxm_hasmask=0 that is allowed in NXT_FLOW_MOD.
- *
- * Order is significant. Earlier flow_mod_specs must satisfy any
- * prerequisites for matching fields specified later, by copying constant
- * values into prerequisite fields.
- *
- * The switch will reject flow_mod_specs that do not satisfy NXM masking
- * restrictions.
- *
- * - If 'dst' is 1, the flow_mod_spec specifies an NXAST_REG_LOAD action for
- * the new flow. The new flow copies the source bits into
- * field[ofs:ofs+n_bits-1]. Actions are executed in the same order as the
- * flow_mod_specs.
- *
- * A single NXAST_REG_LOAD action writes no more than 64 bits, so n_bits
- * greater than 64 yields multiple NXAST_REG_LOAD actions.
- *
- * The flow_mod_spec destination spec for 'dst' of 2 (when 'src' is 0) is
- * empty. It has the following meaning:
- *
- * - The flow_mod_spec specifies an OFPAT_OUTPUT action for the new flow.
- * The new flow outputs to the OpenFlow port specified by the source field.
- * Of the special output ports with value OFPP_MAX or larger, OFPP_IN_PORT,
- * OFPP_FLOOD, OFPP_LOCAL, and OFPP_ALL are supported. Other special ports
- * may not be used.
- *
- * Resource Management
- * -------------------
- *
- * A switch has a finite amount of flow table space available for learning.
- * When this space is exhausted, no new learning table entries will be learned
- * until some existing flow table entries expire. The controller should be
- * prepared to handle this by flooding (which can be implemented as a
- * low-priority flow).
- *
- * If a learned flow matches a single TCP stream with a relatively long
- * timeout, one may make the best of resource constraints by setting
- * 'fin_idle_timeout' or 'fin_hard_timeout' (both measured in seconds), or
- * both, to shorter timeouts. When either of these is specified as a nonzero
- * value, OVS adds a NXAST_FIN_TIMEOUT action, with the specified timeouts, to
- * the learned flow.
- *
- * Examples
- * --------
- *
- * The following examples give a prose description of the flow_mod_specs along
- * with informal notation for how those would be represented and a hex dump of
- * the bytes that would be required.
- *
- * These examples could work with various nx_action_learn parameters. Typical
- * values would be idle_timeout=OFP_FLOW_PERMANENT, hard_timeout=60,
- * priority=OFP_DEFAULT_PRIORITY, flags=0, table_id=10.
- *
- * 1. Learn input port based on the source MAC, with lookup into
- * NXM_NX_REG1[16:31] by resubmit to in_port=99:
- *
- * Match on in_port=99:
- * ovs_be16(src=1, dst=0, n_bits=16), 20 10
- * ovs_be16(99), 00 63
- * ovs_be32(NXM_OF_IN_PORT), ovs_be16(0) 00 00 00 02 00 00
- *
- * Match Ethernet destination on Ethernet source from packet:
- * ovs_be16(src=0, dst=0, n_bits=48), 00 30
- * ovs_be32(NXM_OF_ETH_SRC), ovs_be16(0) 00 00 04 06 00 00
- * ovs_be32(NXM_OF_ETH_DST), ovs_be16(0) 00 00 02 06 00 00
- *
- * Set NXM_NX_REG1[16:31] to the packet's input port:
- * ovs_be16(src=0, dst=1, n_bits=16), 08 10
- * ovs_be32(NXM_OF_IN_PORT), ovs_be16(0) 00 00 00 02 00 00
- * ovs_be32(NXM_NX_REG1), ovs_be16(16) 00 01 02 04 00 10
- *
- * Given a packet that arrived on port A with Ethernet source address B,
- * this would set up the flow "in_port=99, dl_dst=B,
- * actions=load:A->NXM_NX_REG1[16..31]".
- *
- * In syntax accepted by ovs-ofctl, this action is: learn(in_port=99,
- * NXM_OF_ETH_DST[]=NXM_OF_ETH_SRC[],
- * load:NXM_OF_IN_PORT[]->NXM_NX_REG1[16..31])
- *
- * 2. Output to input port based on the source MAC and VLAN VID, with lookup
- * into NXM_NX_REG1[16:31]:
- *
- * Match on same VLAN ID as packet:
- * ovs_be16(src=0, dst=0, n_bits=12), 00 0c
- * ovs_be32(NXM_OF_VLAN_TCI), ovs_be16(0) 00 00 08 02 00 00
- * ovs_be32(NXM_OF_VLAN_TCI), ovs_be16(0) 00 00 08 02 00 00
- *
- * Match Ethernet destination on Ethernet source from packet:
- * ovs_be16(src=0, dst=0, n_bits=48), 00 30
- * ovs_be32(NXM_OF_ETH_SRC), ovs_be16(0) 00 00 04 06 00 00
- * ovs_be32(NXM_OF_ETH_DST), ovs_be16(0) 00 00 02 06 00 00
- *
- * Output to the packet's input port:
- * ovs_be16(src=0, dst=2, n_bits=16), 10 10
- * ovs_be32(NXM_OF_IN_PORT), ovs_be16(0) 00 00 00 02 00 00
- *
- * Given a packet that arrived on port A with Ethernet source address B in
- * VLAN C, this would set up the flow "dl_dst=B, vlan_vid=C,
- * actions=output:A".
- *
- * In syntax accepted by ovs-ofctl, this action is:
- * learn(NXM_OF_VLAN_TCI[0..11], NXM_OF_ETH_DST[]=NXM_OF_ETH_SRC[],
- * output:NXM_OF_IN_PORT[])
- *
- * 3. Here's a recipe for a very simple-minded MAC learning switch. It uses a
- * 10-second MAC expiration time to make it easier to see what's going on
- *
- * ovs-vsctl del-controller br0
- * ovs-ofctl del-flows br0
- * ovs-ofctl add-flow br0 "table=0 actions=learn(table=1, \
- hard_timeout=10, NXM_OF_VLAN_TCI[0..11], \
- NXM_OF_ETH_DST[]=NXM_OF_ETH_SRC[], \
- output:NXM_OF_IN_PORT[]), resubmit(,1)"
- * ovs-ofctl add-flow br0 "table=1 priority=0 actions=flood"
- *
- * You can then dump the MAC learning table with:
- *
- * ovs-ofctl dump-flows br0 table=1
- *
- * Usage Advice
- * ------------
- *
- * For best performance, segregate learned flows into a table that is not used
- * for any other flows except possibly for a lowest-priority "catch-all" flow
- * (a flow with no match criteria). If different learning actions specify
- * different match criteria, use different tables for the learned flows.
- *
- * The meaning of 'hard_timeout' and 'idle_timeout' can be counterintuitive.
- * These timeouts apply to the flow that is added, which means that a flow with
- * an idle timeout will expire when no traffic has been sent *to* the learned
- * address. This is not usually the intent in MAC learning; instead, we want
- * the MAC learn entry to expire when no traffic has been sent *from* the
- * learned address. Use a hard timeout for that.
- */
-struct nx_action_learn {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* At least 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_LEARN. */
- ovs_be16 idle_timeout; /* Idle time before discarding (seconds). */
- ovs_be16 hard_timeout; /* Max time before discarding (seconds). */
- ovs_be16 priority; /* Priority level of flow entry. */
- ovs_be64 cookie; /* Cookie for new flow. */
- ovs_be16 flags; /* Either 0 or OFPFF_SEND_FLOW_REM. */
- uint8_t table_id; /* Table to insert flow entry. */
- uint8_t pad; /* Must be zero. */
- ovs_be16 fin_idle_timeout; /* Idle timeout after FIN, if nonzero. */
- ovs_be16 fin_hard_timeout; /* Hard timeout after FIN, if nonzero. */
- /* Followed by a sequence of flow_mod_spec elements, as described above,
- * until the end of the action is reached. */
-};
-OFP_ASSERT(sizeof(struct nx_action_learn) == 32);
-
-#define NX_LEARN_N_BITS_MASK 0x3ff
-
-#define NX_LEARN_SRC_FIELD (0 << 13) /* Copy from field. */
-#define NX_LEARN_SRC_IMMEDIATE (1 << 13) /* Copy from immediate value. */
-#define NX_LEARN_SRC_MASK (1 << 13)
-
-#define NX_LEARN_DST_MATCH (0 << 11) /* Add match criterion. */
-#define NX_LEARN_DST_LOAD (1 << 11) /* Add NXAST_REG_LOAD action. */
-#define NX_LEARN_DST_OUTPUT (2 << 11) /* Add OFPAT_OUTPUT action. */
-#define NX_LEARN_DST_RESERVED (3 << 11) /* Not yet defined. */
-#define NX_LEARN_DST_MASK (3 << 11)
-
-/* Action structure for NXAST_FIN_TIMEOUT.
- *
- * This action changes the idle timeout or hard timeout, or both, of this
- * OpenFlow rule when the rule matches a TCP packet with the FIN or RST flag.
- * When such a packet is observed, the action reduces the rule's idle timeout
- * to 'fin_idle_timeout' and its hard timeout to 'fin_hard_timeout'. This
- * action has no effect on an existing timeout that is already shorter than the
- * one that the action specifies. A 'fin_idle_timeout' or 'fin_hard_timeout'
- * of zero has no effect on the respective timeout.
- *
- * 'fin_idle_timeout' and 'fin_hard_timeout' are measured in seconds.
- * 'fin_hard_timeout' specifies time since the flow's creation, not since the
- * receipt of the FIN or RST.
- *
- * This is useful for quickly discarding learned TCP flows that otherwise will
- * take a long time to expire.
- *
- * This action is intended for use with an OpenFlow rule that matches only a
- * single TCP flow. If the rule matches multiple TCP flows (e.g. it wildcards
- * all TCP traffic, or all TCP traffic to a particular port), then any FIN or
- * RST in any of those flows will cause the entire OpenFlow rule to expire
- * early, which is not normally desirable.
- */
-struct nx_action_fin_timeout {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_FIN_TIMEOUT. */
- ovs_be16 fin_idle_timeout; /* New idle timeout, if nonzero. */
- ovs_be16 fin_hard_timeout; /* New hard timeout, if nonzero. */
- ovs_be16 pad; /* Must be zero. */
-};
-OFP_ASSERT(sizeof(struct nx_action_fin_timeout) == 16);
-\f
-/* Action structure for NXAST_AUTOPATH.
- *
- * This action performs the following steps in sequence:
- *
- * 1. Hashes the flow using an implementation-defined hash function.
- *
- * The hashed fields' values are drawn from the current state of the
- * flow, including all modifications that have been made by actions up to
- * this point.
- *
- * 2. Selects an OpenFlow 'port'.
- *
- * 'port' is selected in an implementation-defined manner, taking into
- * account 'id' and the hash value calculated in step 1.
- *
- * Generally a switch will have been configured with a set of ports that
- * may be chosen given 'id'. The switch may take into account any number
- * of factors when choosing 'port' from its configured set. Factors may
- * include carrier, load, and the results of configuration protocols such
- * as LACP.
- *
- * 3. Stores 'port' in dst[ofs:ofs+n_bits].
- *
- * The format and semantics of 'dst' and 'ofs_nbits' are similar to those
- * for the NXAST_REG_LOAD action.
- *
- * The switch will reject actions in which ofs+n_bits is greater than the width
- * of 'dst', with error type OFPET_BAD_ACTION, code OFPBAC_BAD_ARGUMENT.
- */
-struct nx_action_autopath {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_AUTOPATH. */
-
- /* Where to store the result. */
- ovs_be16 ofs_nbits; /* (ofs << 6) | (n_bits - 1). */
- ovs_be32 dst; /* Destination. */
-
- ovs_be32 id; /* Autopath ID. */
- ovs_be32 pad;
-};
-OFP_ASSERT(sizeof(struct nx_action_autopath) == 24);
-\f
-/* Action structure for NXAST_BUNDLE and NXAST_BUNDLE_LOAD.
- *
- * The bundle actions choose a slave from a supplied list of options.
- * NXAST_BUNDLE outputs to its selection. NXAST_BUNDLE_LOAD writes its
- * selection to a register.
- *
- * The list of possible slaves follows the nx_action_bundle structure. The size
- * of each slave is governed by its type as indicated by the 'slave_type'
- * parameter. The list of slaves should be padded at its end with zeros to make
- * the total length of the action a multiple of 8.
- *
- * Switches infer from the 'slave_type' parameter the size of each slave. All
- * implementations must support the NXM_OF_IN_PORT 'slave_type' which indicates
- * that the slaves are OpenFlow port numbers with NXM_LENGTH(NXM_OF_IN_PORT) ==
- * 2 byte width. Switches should reject actions which indicate unknown or
- * unsupported slave types.
- *
- * Switches use a strategy dictated by the 'algorithm' parameter to choose a
- * slave. If the switch does not support the specified 'algorithm' parameter,
- * it should reject the action.
- *
- * Several algorithms take into account liveness when selecting slaves. The
- * liveness of a slave is implementation defined (with one exception), but will
- * generally take into account things like its carrier status and the results
- * of any link monitoring protocols which happen to be running on it. In order
- * to give controllers a place-holder value, the OFPP_NONE port is always
- * considered live.
- *
- * Some slave selection strategies require the use of a hash function, in which
- * case the 'fields' and 'basis' parameters should be populated. The 'fields'
- * parameter (one of NX_HASH_FIELDS_*) designates which parts of the flow to
- * hash. Refer to the definition of "enum nx_hash_fields" for details. The
- * 'basis' parameter is used as a universal hash parameter. Different values
- * of 'basis' yield different hash results.
- *
- * The 'zero' parameter at the end of the action structure is reserved for
- * future use. Switches are required to reject actions which have nonzero
- * bytes in the 'zero' field.
- *
- * NXAST_BUNDLE actions should have 'ofs_nbits' and 'dst' zeroed. Switches
- * should reject actions which have nonzero bytes in either of these fields.
- *
- * NXAST_BUNDLE_LOAD stores the OpenFlow port number of the selected slave in
- * dst[ofs:ofs+n_bits]. The format and semantics of 'dst' and 'ofs_nbits' are
- * similar to those for the NXAST_REG_LOAD action. */
-struct nx_action_bundle {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length including slaves. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_BUNDLE or NXAST_BUNDLE_LOAD. */
-
- /* Slave choice algorithm to apply to hash value. */
- ovs_be16 algorithm; /* One of NX_BD_ALG_*. */
-
- /* What fields to hash and how. */
- ovs_be16 fields; /* One of NX_HASH_FIELDS_*. */
- ovs_be16 basis; /* Universal hash parameter. */
-
- ovs_be32 slave_type; /* NXM_OF_IN_PORT. */
- ovs_be16 n_slaves; /* Number of slaves. */
-
- ovs_be16 ofs_nbits; /* (ofs << 6) | (n_bits - 1). */
- ovs_be32 dst; /* Destination. */
-
- uint8_t zero[4]; /* Reserved. Must be zero. */
-};
-OFP_ASSERT(sizeof(struct nx_action_bundle) == 32);
-
-/* NXAST_BUNDLE: Bundle slave choice algorithm to apply.
- *
- * In the descriptions below, 'slaves' is the list of possible slaves in the
- * order they appear in the OpenFlow action. */
-enum nx_bd_algorithm {
- /* Chooses the first live slave listed in the bundle.
- *
- * O(n_slaves) performance. */
- NX_BD_ALG_ACTIVE_BACKUP,
-
- /* for i in [0,n_slaves):
- * weights[i] = hash(flow, i)
- * slave = { slaves[i] such that weights[i] >= weights[j] for all j != i }
- *
- * Redistributes 1/n_slaves of traffic when a slave's liveness changes.
- * O(n_slaves) performance.
- *
- * Uses the 'fields' and 'basis' parameters. */
- NX_BD_ALG_HRW /* Highest Random Weight. */
-};
-
-\f
-/* Action structure for NXAST_DEC_TTL_CNT_IDS.
- *
- * If the packet is not IPv4 or IPv6, does nothing. For IPv4 or IPv6, if the
- * TTL or hop limit is at least 2, decrements it by 1. Otherwise, if TTL or
- * hop limit is 0 or 1, sends a packet-in to the controllers with each of the
- * 'n_controllers' controller IDs specified in 'cnt_ids'.
- *
- * (This differs from NXAST_DEC_TTL in that for NXAST_DEC_TTL the packet-in is
- * sent only to controllers with id 0.)
- */
-struct nx_action_cnt_ids {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length including slaves. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_DEC_TTL_CNT_IDS. */
-
- ovs_be16 n_controllers; /* Number of controllers. */
- uint8_t zeros[4]; /* Must be zero. */
-
- /* Followed by 1 or more controller ids.
- *
- * uint16_t cnt_ids[]; // Controller ids.
- * uint8_t pad[]; // Must be 0 to 8-byte align cnt_ids[].
- */
-};
-OFP_ASSERT(sizeof(struct nx_action_cnt_ids) == 16);
-
-\f
-/* Action structure for NXAST_OUTPUT_REG.
- *
- * Outputs to the OpenFlow port number written to src[ofs:ofs+nbits].
- *
- * The format and semantics of 'src' and 'ofs_nbits' are similar to those for
- * the NXAST_REG_LOAD action.
- *
- * The acceptable nxm_header values for 'src' are the same as the acceptable
- * nxm_header values for the 'src' field of NXAST_REG_MOVE.
- *
- * The 'max_len' field indicates the number of bytes to send when the chosen
- * port is OFPP_CONTROLLER. Its semantics are equivalent to the 'max_len'
- * field of OFPAT_OUTPUT.
- *
- * The 'zero' field is required to be zeroed for forward compatibility. */
-struct nx_action_output_reg {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* 24. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_OUTPUT_REG. */
-
- ovs_be16 ofs_nbits; /* (ofs << 6) | (n_bits - 1). */
- ovs_be32 src; /* Source. */
-
- ovs_be16 max_len; /* Max length to send to controller. */
-
- uint8_t zero[6]; /* Reserved, must be zero. */
-};
-OFP_ASSERT(sizeof(struct nx_action_output_reg) == 24);
-\f
-/* NXAST_EXIT
- *
- * Discontinues action processing.
- *
- * The NXAST_EXIT action causes the switch to immediately halt processing
- * actions for the flow. Any actions which have already been processed are
- * executed by the switch. However, any further actions, including those which
- * may be in different tables, or different levels of the NXAST_RESUBMIT
- * hierarchy, will be ignored.
- *
- * Uses the nx_action_header structure. */
-\f
/* Flexible flow specifications (aka NXM = Nicira Extended Match).
*
* OpenFlow 1.0 has "struct ofp10_match" for specifying flow matches. This
* short, that is also supported by Open vSwitch. This section also defines a
* replacement for each OpenFlow message that includes struct ofp10_match.
*
+ * OpenFlow 1.2+ introduced OpenFlow Extensible Match (OXM), adapting
+ * the design of NXM. The format of NXM and OXM are compatible.
+ *
*
* Format
* ======
* +----------------------------------+---------------+--+------------------+
*
* The most-significant 23 bits of the header are collectively "nxm_type".
- * Bits 16...31 are "nxm_vendor", one of the NXM_VENDOR_* values below. Bits
- * 9...15 are "nxm_field", which is a vendor-specific value. nxm_type normally
- * designates a protocol header, such as the Ethernet type, but it can also
- * refer to packet metadata, such as the switch port on which a packet arrived.
+ * Bits 16...31 are "nxm_vendor", one of OFPXMC12_* values. In case of
+ * NXM, it's either OFPXMC12_NXM_0 or OFPXMC12_NXM_1.
+ * Bits 9...15 are "nxm_field", which is a vendor-specific value. nxm_type
+ * normally designates a protocol header, such as the Ethernet type, but it
+ * can also refer to packet metadata, such as the switch port on which a packet
+ * arrived.
*
* Bit 8 is "nxm_hasmask" (labeled "hm" above for space reasons). The meaning
* of this bit is explained later.
* value, called "nxm_mask". For each 1-bit in position J in nxm_mask, the
* nx_match matches only packets for which bit J in the given field's value
* matches bit J in nxm_value. A 0-bit in nxm_mask causes the
- * corresponding bits in nxm_value and the field's value to be ignored.
- * (The sense of the nxm_mask bits is the opposite of that used by the
- * "wildcards" member of struct ofp10_match.)
+ * corresponding bit in nxm_value is ignored (it should be 0; Open vSwitch
+ * may enforce this someday), as is the corresponding bit in the field's
+ * value. (The sense of the nxm_mask bits is the opposite of that used by
+ * the "wildcards" member of struct ofp10_match.)
*
* When nxm_hasmask is 1, nxm_length is always even.
*
* nx_match error.
*/
-#define NXM_HEADER__(VENDOR, FIELD, HASMASK, LENGTH) \
- (((VENDOR) << 16) | ((FIELD) << 9) | ((HASMASK) << 8) | (LENGTH))
-#define NXM_HEADER(VENDOR, FIELD, LENGTH) \
- NXM_HEADER__(VENDOR, FIELD, 0, LENGTH)
-#define NXM_HEADER_W(VENDOR, FIELD, LENGTH) \
- NXM_HEADER__(VENDOR, FIELD, 1, (LENGTH) * 2)
-#define NXM_VENDOR(HEADER) ((HEADER) >> 16)
-#define NXM_FIELD(HEADER) (((HEADER) >> 9) & 0x7f)
-#define NXM_TYPE(HEADER) (((HEADER) >> 9) & 0x7fffff)
-#define NXM_HASMASK(HEADER) (((HEADER) >> 8) & 1)
-#define NXM_LENGTH(HEADER) ((HEADER) & 0xff)
-
-#define NXM_MAKE_WILD_HEADER(HEADER) \
- NXM_HEADER_W(NXM_VENDOR(HEADER), NXM_FIELD(HEADER), NXM_LENGTH(HEADER))
-
-/* ## ------------------------------- ## */
-/* ## OpenFlow 1.0-compatible fields. ## */
-/* ## ------------------------------- ## */
-
-/* Physical or virtual port on which the packet was received.
- *
- * Prereqs: None.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Not maskable. */
-#define NXM_OF_IN_PORT NXM_HEADER (0x0000, 0, 2)
-
-/* Source or destination address in Ethernet header.
- *
- * Prereqs: None.
- *
- * Format: 48-bit Ethernet MAC address.
- *
- * Masking: Fully maskable, in versions 1.8 and later. Earlier versions only
- * supported the following masks for NXM_OF_ETH_DST_W: 00:00:00:00:00:00,
- * fe:ff:ff:ff:ff:ff, 01:00:00:00:00:00, ff:ff:ff:ff:ff:ff. */
-#define NXM_OF_ETH_DST NXM_HEADER (0x0000, 1, 6)
-#define NXM_OF_ETH_DST_W NXM_HEADER_W(0x0000, 1, 6)
-#define NXM_OF_ETH_SRC NXM_HEADER (0x0000, 2, 6)
-#define NXM_OF_ETH_SRC_W NXM_HEADER_W(0x0000, 2, 6)
-
-/* Packet's Ethernet type.
- *
- * For an Ethernet II packet this is taken from the Ethernet header. For an
- * 802.2 LLC+SNAP header with OUI 00-00-00 this is taken from the SNAP header.
- * A packet that has neither format has value 0x05ff
- * (OFP_DL_TYPE_NOT_ETH_TYPE).
- *
- * For a packet with an 802.1Q header, this is the type of the encapsulated
- * frame.
- *
- * Prereqs: None.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Not maskable. */
-#define NXM_OF_ETH_TYPE NXM_HEADER (0x0000, 3, 2)
-
-/* 802.1Q TCI.
- *
- * For a packet with an 802.1Q header, this is the Tag Control Information
- * (TCI) field, with the CFI bit forced to 1. For a packet with no 802.1Q
- * header, this has value 0.
- *
- * Prereqs: None.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Arbitrary masks.
- *
- * This field can be used in various ways:
- *
- * - If it is not constrained at all, the nx_match matches packets without
- * an 802.1Q header or with an 802.1Q header that has any TCI value.
- *
- * - Testing for an exact match with 0 matches only packets without an
- * 802.1Q header.
- *
- * - Testing for an exact match with a TCI value with CFI=1 matches packets
- * that have an 802.1Q header with a specified VID and PCP.
- *
- * - Testing for an exact match with a nonzero TCI value with CFI=0 does
- * not make sense. The switch may reject this combination.
- *
- * - Testing with a specific VID and CFI=1, with nxm_mask=0x1fff, matches
- * packets that have an 802.1Q header with that VID (and any PCP).
- *
- * - Testing with a specific PCP and CFI=1, with nxm_mask=0xf000, matches
- * packets that have an 802.1Q header with that PCP (and any VID).
- *
- * - Testing with nxm_value=0, nxm_mask=0x0fff matches packets with no 802.1Q
- * header or with an 802.1Q header with a VID of 0.
- *
- * - Testing with nxm_value=0, nxm_mask=0xe000 matches packets with no 802.1Q
- * header or with an 802.1Q header with a PCP of 0.
- *
- * - Testing with nxm_value=0, nxm_mask=0xefff matches packets with no 802.1Q
- * header or with an 802.1Q header with both VID and PCP of 0.
- */
-#define NXM_OF_VLAN_TCI NXM_HEADER (0x0000, 4, 2)
-#define NXM_OF_VLAN_TCI_W NXM_HEADER_W(0x0000, 4, 2)
-
-/* The "type of service" byte of the IP header, with the ECN bits forced to 0.
- *
- * Prereqs: NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- *
- * Format: 8-bit integer with 2 least-significant bits forced to 0.
- *
- * Masking: Not maskable. */
-#define NXM_OF_IP_TOS NXM_HEADER (0x0000, 5, 1)
-
-/* The "protocol" byte in the IP header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- *
- * Format: 8-bit integer.
- *
- * Masking: Not maskable. */
-#define NXM_OF_IP_PROTO NXM_HEADER (0x0000, 6, 1)
-
-/* The source or destination address in the IP header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x0800 exactly.
- *
- * Format: 32-bit integer in network byte order.
- *
- * Masking: Fully maskable, in Open vSwitch 1.8 and later. In earlier
- * versions, only CIDR masks are allowed, that is, masks that consist of N
- * high-order bits set to 1 and the other 32-N bits set to 0. */
-#define NXM_OF_IP_SRC NXM_HEADER (0x0000, 7, 4)
-#define NXM_OF_IP_SRC_W NXM_HEADER_W(0x0000, 7, 4)
-#define NXM_OF_IP_DST NXM_HEADER (0x0000, 8, 4)
-#define NXM_OF_IP_DST_W NXM_HEADER_W(0x0000, 8, 4)
-
-/* The source or destination port in the TCP header.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- * NXM_OF_IP_PROTO must match 6 exactly.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Fully maskable, in Open vSwitch 1.6 and later. Not maskable, in
- * earlier versions. */
-#define NXM_OF_TCP_SRC NXM_HEADER (0x0000, 9, 2)
-#define NXM_OF_TCP_SRC_W NXM_HEADER_W(0x0000, 9, 2)
-#define NXM_OF_TCP_DST NXM_HEADER (0x0000, 10, 2)
-#define NXM_OF_TCP_DST_W NXM_HEADER_W(0x0000, 10, 2)
-
-/* The source or destination port in the UDP header.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match either 0x0800 or 0x86dd.
- * NXM_OF_IP_PROTO must match 17 exactly.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Fully maskable, in Open vSwitch 1.6 and later. Not maskable, in
- * earlier versions. */
-#define NXM_OF_UDP_SRC NXM_HEADER (0x0000, 11, 2)
-#define NXM_OF_UDP_SRC_W NXM_HEADER_W(0x0000, 11, 2)
-#define NXM_OF_UDP_DST NXM_HEADER (0x0000, 12, 2)
-#define NXM_OF_UDP_DST_W NXM_HEADER_W(0x0000, 12, 2)
-
-/* The type or code in the ICMP header.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match 0x0800 exactly.
- * NXM_OF_IP_PROTO must match 1 exactly.
- *
- * Format: 8-bit integer.
- *
- * Masking: Not maskable. */
-#define NXM_OF_ICMP_TYPE NXM_HEADER (0x0000, 13, 1)
-#define NXM_OF_ICMP_CODE NXM_HEADER (0x0000, 14, 1)
-
-/* ARP opcode.
- *
- * For an Ethernet+IP ARP packet, the opcode in the ARP header. Always 0
- * otherwise. Only ARP opcodes between 1 and 255 should be specified for
- * matching.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x0806 exactly.
- *
- * Format: 16-bit integer in network byte order.
- *
- * Masking: Not maskable. */
-#define NXM_OF_ARP_OP NXM_HEADER (0x0000, 15, 2)
-
-/* For an Ethernet+IP ARP packet, the source or target protocol address
- * in the ARP header. Always 0 otherwise.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x0806 exactly.
- *
- * Format: 32-bit integer in network byte order.
- *
- * Masking: Fully maskable, in Open vSwitch 1.8 and later. In earlier
- * versions, only CIDR masks are allowed, that is, masks that consist of N
- * high-order bits set to 1 and the other 32-N bits set to 0. */
-#define NXM_OF_ARP_SPA NXM_HEADER (0x0000, 16, 4)
-#define NXM_OF_ARP_SPA_W NXM_HEADER_W(0x0000, 16, 4)
-#define NXM_OF_ARP_TPA NXM_HEADER (0x0000, 17, 4)
-#define NXM_OF_ARP_TPA_W NXM_HEADER_W(0x0000, 17, 4)
-
-/* ## ------------------------ ## */
-/* ## Nicira match extensions. ## */
-/* ## ------------------------ ## */
-
-/* Metadata registers.
- *
- * Registers initially have value 0. Actions allow register values to be
- * manipulated.
- *
- * Prereqs: None.
- *
- * Format: Array of 32-bit integer registers. Space is reserved for up to
- * NXM_NX_MAX_REGS registers, but switches may implement fewer.
- *
- * Masking: Arbitrary masks. */
+/* Number of registers allocated NXM field IDs. */
#define NXM_NX_MAX_REGS 16
-#define NXM_NX_REG(IDX) NXM_HEADER (0x0001, IDX, 4)
-#define NXM_NX_REG_W(IDX) NXM_HEADER_W(0x0001, IDX, 4)
-#define NXM_NX_REG_IDX(HEADER) NXM_FIELD(HEADER)
-#define NXM_IS_NX_REG(HEADER) (!((((HEADER) ^ NXM_NX_REG0)) & 0xffffe1ff))
-#define NXM_IS_NX_REG_W(HEADER) (!((((HEADER) ^ NXM_NX_REG0_W)) & 0xffffe1ff))
-#define NXM_NX_REG0 NXM_HEADER (0x0001, 0, 4)
-#define NXM_NX_REG0_W NXM_HEADER_W(0x0001, 0, 4)
-#define NXM_NX_REG1 NXM_HEADER (0x0001, 1, 4)
-#define NXM_NX_REG1_W NXM_HEADER_W(0x0001, 1, 4)
-#define NXM_NX_REG2 NXM_HEADER (0x0001, 2, 4)
-#define NXM_NX_REG2_W NXM_HEADER_W(0x0001, 2, 4)
-#define NXM_NX_REG3 NXM_HEADER (0x0001, 3, 4)
-#define NXM_NX_REG3_W NXM_HEADER_W(0x0001, 3, 4)
-#define NXM_NX_REG4 NXM_HEADER (0x0001, 4, 4)
-#define NXM_NX_REG4_W NXM_HEADER_W(0x0001, 4, 4)
-#define NXM_NX_REG5 NXM_HEADER (0x0001, 5, 4)
-#define NXM_NX_REG5_W NXM_HEADER_W(0x0001, 5, 4)
-#define NXM_NX_REG6 NXM_HEADER (0x0001, 6, 4)
-#define NXM_NX_REG6_W NXM_HEADER_W(0x0001, 6, 4)
-#define NXM_NX_REG7 NXM_HEADER (0x0001, 7, 4)
-#define NXM_NX_REG7_W NXM_HEADER_W(0x0001, 7, 4)
-
-/* Tunnel ID.
- *
- * For a packet received via GRE tunnel including a (32-bit) key, the key is
- * stored in the low 32-bits and the high bits are zeroed. For other packets,
- * the value is 0.
- *
- * Prereqs: None.
- *
- * Format: 64-bit integer in network byte order.
- *
- * Masking: Arbitrary masks. */
-#define NXM_NX_TUN_ID NXM_HEADER (0x0001, 16, 8)
-#define NXM_NX_TUN_ID_W NXM_HEADER_W(0x0001, 16, 8)
-
-/* For an Ethernet+IP ARP packet, the source or target hardware address
- * in the ARP header. Always 0 otherwise.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x0806 exactly.
- *
- * Format: 48-bit Ethernet MAC address.
- *
- * Masking: Not maskable. */
-#define NXM_NX_ARP_SHA NXM_HEADER (0x0001, 17, 6)
-#define NXM_NX_ARP_THA NXM_HEADER (0x0001, 18, 6)
-
-/* The source or destination address in the IPv6 header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x86dd exactly.
- *
- * Format: 128-bit IPv6 address.
- *
- * Masking: Fully maskable, in Open vSwitch 1.8 and later. In previous
- * versions, only CIDR masks are allowed, that is, masks that consist of N
- * high-order bits set to 1 and the other 128-N bits set to 0. */
-#define NXM_NX_IPV6_SRC NXM_HEADER (0x0001, 19, 16)
-#define NXM_NX_IPV6_SRC_W NXM_HEADER_W(0x0001, 19, 16)
-#define NXM_NX_IPV6_DST NXM_HEADER (0x0001, 20, 16)
-#define NXM_NX_IPV6_DST_W NXM_HEADER_W(0x0001, 20, 16)
-
-/* The type or code in the ICMPv6 header.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match 0x86dd exactly.
- * NXM_OF_IP_PROTO must match 58 exactly.
- *
- * Format: 8-bit integer.
- *
- * Masking: Not maskable. */
-#define NXM_NX_ICMPV6_TYPE NXM_HEADER (0x0001, 21, 1)
-#define NXM_NX_ICMPV6_CODE NXM_HEADER (0x0001, 22, 1)
-
-/* The target address in an IPv6 Neighbor Discovery message.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match 0x86dd exactly.
- * NXM_OF_IP_PROTO must match 58 exactly.
- * NXM_OF_ICMPV6_TYPE must be either 135 or 136.
- *
- * Format: 128-bit IPv6 address.
- *
- * Masking: Fully maskable, in Open vSwitch 1.8 and later. In previous
- * versions, only CIDR masks are allowed, that is, masks that consist of N
- * high-order bits set to 1 and the other 128-N bits set to 0. */
-#define NXM_NX_ND_TARGET NXM_HEADER (0x0001, 23, 16)
-#define NXM_NX_ND_TARGET_W NXM_HEADER_W (0x0001, 23, 16)
-
-/* The source link-layer address option in an IPv6 Neighbor Discovery
- * message.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match 0x86dd exactly.
- * NXM_OF_IP_PROTO must match 58 exactly.
- * NXM_OF_ICMPV6_TYPE must be exactly 135.
- *
- * Format: 48-bit Ethernet MAC address.
- *
- * Masking: Not maskable. */
-#define NXM_NX_ND_SLL NXM_HEADER (0x0001, 24, 6)
-
-/* The target link-layer address option in an IPv6 Neighbor Discovery
- * message.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must match 0x86dd exactly.
- * NXM_OF_IP_PROTO must match 58 exactly.
- * NXM_OF_ICMPV6_TYPE must be exactly 136.
- *
- * Format: 48-bit Ethernet MAC address.
- *
- * Masking: Not maskable. */
-#define NXM_NX_ND_TLL NXM_HEADER (0x0001, 25, 6)
-
-/* IP fragment information.
- *
- * Prereqs:
- * NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- *
- * Format: 8-bit value with one of the values 0, 1, or 3, as described below.
- *
- * Masking: Fully maskable.
- *
- * This field has three possible values:
- *
- * - A packet that is not an IP fragment has value 0.
- *
- * - A packet that is an IP fragment with offset 0 (the first fragment) has
- * bit 0 set and thus value 1.
- *
- * - A packet that is an IP fragment with nonzero offset has bits 0 and 1 set
- * and thus value 3.
- *
- * NX_IP_FRAG_ANY and NX_IP_FRAG_LATER are declared to symbolically represent
- * the meanings of bits 0 and 1.
- *
- * The switch may reject matches against values that can never appear.
- *
- * It is important to understand how this field interacts with the OpenFlow IP
- * fragment handling mode:
- *
- * - In OFPC_FRAG_DROP mode, the OpenFlow switch drops all IP fragments
- * before they reach the flow table, so every packet that is available for
- * matching will have value 0 in this field.
- *
- * - Open vSwitch does not implement OFPC_FRAG_REASM mode, but if it did then
- * IP fragments would be reassembled before they reached the flow table and
- * again every packet available for matching would always have value 0.
- *
- * - In OFPC_FRAG_NORMAL mode, all three values are possible, but OpenFlow
- * 1.0 says that fragments' transport ports are always 0, even for the
- * first fragment, so this does not provide much extra information.
- *
- * - In OFPC_FRAG_NX_MATCH mode, all three values are possible. For
- * fragments with offset 0, Open vSwitch makes L4 header information
- * available.
- */
-#define NXM_NX_IP_FRAG NXM_HEADER (0x0001, 26, 1)
-#define NXM_NX_IP_FRAG_W NXM_HEADER_W(0x0001, 26, 1)
/* Bits in the value of NXM_NX_IP_FRAG. */
#define NX_IP_FRAG_ANY (1 << 0) /* Is this a fragment? */
#define NX_IP_FRAG_LATER (1 << 1) /* Is this a fragment with nonzero offset? */
-/* The flow label in the IPv6 header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must match 0x86dd exactly.
- *
- * Format: 20-bit IPv6 flow label in least-significant bits.
- *
- * Masking: Not maskable. */
-#define NXM_NX_IPV6_LABEL NXM_HEADER (0x0001, 27, 4)
-
-/* The ECN of the IP header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- *
- * Format: ECN in the low-order 2 bits.
- *
- * Masking: Not maskable. */
-#define NXM_NX_IP_ECN NXM_HEADER (0x0001, 28, 1)
-
-/* The time-to-live/hop limit of the IP header.
- *
- * Prereqs: NXM_OF_ETH_TYPE must be either 0x0800 or 0x86dd.
- *
- * Format: 8-bit integer.
- *
- * Masking: Not maskable. */
-#define NXM_NX_IP_TTL NXM_HEADER (0x0001, 29, 1)
-
-/* Flow cookie.
- *
- * This may be used to gain the OpenFlow 1.1-like ability to restrict
- * certain NXM-based Flow Mod and Flow Stats Request messages to flows
- * with specific cookies. See the "nx_flow_mod" and "nx_flow_stats_request"
- * structure definitions for more details. This match is otherwise not
- * allowed.
- *
- * Prereqs: None.
- *
- * Format: 64-bit integer in network byte order.
- *
- * Masking: Arbitrary masks. */
-#define NXM_NX_COOKIE NXM_HEADER (0x0001, 30, 8)
-#define NXM_NX_COOKIE_W NXM_HEADER_W(0x0001, 30, 8)
+/* Bits in the value of NXM_NX_TUN_FLAGS. */
+#define NX_TUN_FLAG_OAM (1 << 0) /* Is this an OAM packet? */
/* ## --------------------- ## */
/* ## Requests and replies. ## */
enum nx_flow_format {
NXFF_OPENFLOW10 = 0, /* Standard OpenFlow 1.0 compatible. */
- NXFF_NXM = 2, /* Nicira extended match. */
- NXFF_OPENFLOW12 = 3 /* OpenFlow 1.2 format. */
+ NXFF_NXM = 2 /* Nicira extended match. */
};
/* NXT_SET_FLOW_FORMAT request. */
*/
struct nx_flow_mod {
ovs_be64 cookie; /* Opaque controller-issued identifier. */
- ovs_be16 command; /* One of OFPFC_*. */
+ ovs_be16 command; /* OFPFC_* + possibly a table ID (see comment
+ * on struct nx_flow_mod_table_id). */
ovs_be16 idle_timeout; /* Idle time before discarding (seconds). */
ovs_be16 hard_timeout; /* Max time before discarding (seconds). */
ovs_be16 priority; /* Priority level of flow entry. */
};
OFP_ASSERT(sizeof(struct nx_flow_mod) == 32);
-/* NXT_FLOW_REMOVED (analogous to OFPT_FLOW_REMOVED). */
+/* NXT_FLOW_REMOVED (analogous to OFPT_FLOW_REMOVED).
+ *
+ * 'table_id' is present only in Open vSwitch 1.11 and later. In earlier
+ * versions of Open vSwitch, this is a padding byte that is always zeroed.
+ * Therefore, a 'table_id' value of 0 indicates that the table ID is not known,
+ * and other values may be interpreted as one more than the flow's former table
+ * ID. */
struct nx_flow_removed {
ovs_be64 cookie; /* Opaque controller-issued identifier. */
ovs_be16 priority; /* Priority level of flow entry. */
uint8_t reason; /* One of OFPRR_*. */
- uint8_t pad[1]; /* Align to 32-bits. */
+ uint8_t table_id; /* Flow's former table ID, plus one. */
ovs_be32 duration_sec; /* Time flow was alive in seconds. */
ovs_be32 duration_nsec; /* Time flow was alive in nanoseconds beyond
duration_sec. */
ovs_be16 controller_id; /* New controller connection ID. */
};
OFP_ASSERT(sizeof(struct nx_controller_id) == 8);
-
-/* Action structure for NXAST_CONTROLLER.
- *
- * This generalizes using OFPAT_OUTPUT to send a packet to OFPP_CONTROLLER. In
- * addition to the 'max_len' that OFPAT_OUTPUT supports, it also allows
- * specifying:
- *
- * - 'reason': The reason code to use in the ofp_packet_in or nx_packet_in.
- *
- * - 'controller_id': The ID of the controller connection to which the
- * ofp_packet_in should be sent. The ofp_packet_in or nx_packet_in is
- * sent only to controllers that have the specified controller connection
- * ID. See "struct nx_controller_id" for more information. */
-struct nx_action_controller {
- ovs_be16 type; /* OFPAT_VENDOR. */
- ovs_be16 len; /* Length is 16. */
- ovs_be32 vendor; /* NX_VENDOR_ID. */
- ovs_be16 subtype; /* NXAST_CONTROLLER. */
- ovs_be16 max_len; /* Maximum length to send to controller. */
- ovs_be16 controller_id; /* Controller ID to send packet-in. */
- uint8_t reason; /* enum ofp_packet_in_reason (OFPR_*). */
- uint8_t zero; /* Must be zero. */
-};
-OFP_ASSERT(sizeof(struct nx_action_controller) == 16);
\f
/* Flow Table Monitoring
* =====================
* a flow_mod with type OFPFC_MODIFY affects multiple flows, but only some
* of those modifications succeed (e.g. due to hardware limitations).
*
- * This cannot occur with the current implementation of the Open vSwitch
- * software datapath. It could happen with other datapath implementations.
+ * This cannot occur with the Open vSwitch software datapath. This also
+ * cannot occur in Open vSwitch 2.4 and later, because these versions only
+ * execute any flow modifications if all of them will succeed.
*
* - Changes that race with conflicting changes made by other controllers or
* other flow_mods (not separated by barriers) by the same controller.
* (regardless of datapath) because Open vSwitch internally serializes
* potentially conflicting changes.
*
+ * - Changes that occur when flow notification is paused (see "Buffer
+ * Management" above).
+ *
* A flow_mod that does not change the flow table will not trigger any
* notification, even an abbreviated one. For example, a "modify" or "delete"
* flow_mod that does not match any flows will not trigger a notification.
};
OFP_ASSERT(sizeof(struct nx_flow_monitor_cancel) == 4);
+/* Variable-length option TLV table maintenance commands.
+ *
+ * The option in Type-Length-Value format is widely used in tunnel options,
+ * e.g., the base Geneve header is followed by zero or more options in TLV
+ * format. Each option consists of a four byte option header and a variable
+ * amount of option data interpreted according to the type. The generic TLV
+ * format in tunnel options is as following:
+ *
+ * 0 1 2 3
+ * 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Option Class | Type |R|R|R| Length |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ * | Variable Option Data |
+ * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
+ *
+ * In order to work with this variable-length options in TLV format in
+ * tunnel options, we need to maintain a mapping table between an option
+ * TLV (defined by <class, type, length>) and an NXM field that can be
+ * operated on for the purposes of matches, actions, etc. This mapping
+ * must be explicitly specified by the user.
+ *
+ * There are two primary groups of OpenFlow messages that are introduced
+ * as Nicira extensions: modification commands (add, delete, clear mappings)
+ * and table status request/reply to dump the current table along with switch
+ * information.
+ *
+ * Note that mappings should not be changed while they are in active use by
+ * a flow. The result of doing so is undefined. */
+
+/* TLV table commands */
+enum nx_tlv_table_mod_command {
+ NXTTMC_ADD, /* New mappings (fails if an option is already
+ mapped). */
+ NXTTMC_DELETE, /* Delete mappings, identified by index
+ * (unmapped options are ignored). */
+ NXTTMC_CLEAR, /* Clear all mappings. Additional information
+ in this command is ignored. */
+};
+
+/* Map between an option TLV and an NXM field. */
+struct nx_tlv_map {
+ ovs_be16 option_class; /* TLV class. */
+ uint8_t option_type; /* TLV type. */
+ uint8_t option_len; /* TLV length (multiple of 4). */
+ ovs_be16 index; /* NXM_NX_TUN_METADATA<n> index */
+ uint8_t pad[2];
+};
+OFP_ASSERT(sizeof(struct nx_tlv_map) == 8);
+
+/* NXT_TLV_TABLE_MOD.
+ *
+ * Use to configure a mapping between option TLVs (class, type, length)
+ * and NXM fields (NXM_NX_TUN_METADATA<n> where 'index' is <n>).
+ *
+ * This command is atomic: all operations on different options will
+ * either succeed or fail. */
+struct nx_tlv_table_mod {
+ ovs_be16 command; /* One of NTTTMC_* */
+ uint8_t pad[6];
+ /* struct nx_tlv_map[0]; Array of maps between indicies and option
+ TLVs. The number of elements is inferred
+ from the length field in the header. */
+};
+OFP_ASSERT(sizeof(struct nx_tlv_table_mod) == 8);
+
+/* NXT_TLV_TABLE_REPLY.
+ *
+ * Issued in reponse to an NXT_TLV_TABLE_REQUEST to give information
+ * about the current status of the TLV table in the switch. Provides
+ * both static information about the switch's capabilities as well as
+ * the configured TLV table. */
+struct nx_tlv_table_reply {
+ ovs_be32 max_option_space; /* Maximum total of option sizes supported. */
+ ovs_be16 max_fields; /* Maximum number of match fields supported. */
+ uint8_t reserved[10];
+ /* struct nx_tlv_map[0]; Array of maps between indicies and option
+ TLVs. The number of elements is inferred
+ from the length field in the header. */
+};
+OFP_ASSERT(sizeof(struct nx_tlv_table_reply) == 16);
+
#endif /* openflow/nicira-ext.h */