X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=lib%2Fofp-util.h;h=1f7780825981c08cc4ff26f90816b930b33905e6;hb=2431be1b68d386bd616378d2c528242775c4d54a;hp=2f73ec2bf6f91a3479fe97858aaaec59a1cd6b8e;hpb=64626975103668a5662622819536780ba87bef79;p=sliver-openvswitch.git diff --git a/lib/ofp-util.h b/lib/ofp-util.h index 2f73ec2bf..1f7780825 100644 --- a/lib/ofp-util.h +++ b/lib/ofp-util.h @@ -1,5 +1,5 @@ /* - * Copyright (c) 2008, 2009, 2010, 2011, 2012 Nicira, Inc. + * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013 Nicira, Inc. * * Licensed under the Apache License, Version 2.0 (the "License"); * you may not use this file except in compliance with the License. @@ -17,7 +17,6 @@ #ifndef OFP_UTIL_H #define OFP_UTIL_H 1 -#include #include #include #include @@ -28,16 +27,30 @@ #include "netdev.h" #include "openflow/nicira-ext.h" #include "openvswitch/types.h" +#include "type-props.h" struct ofpbuf; +union ofp_action; /* Port numbers. */ -enum ofperr ofputil_port_from_ofp11(ovs_be32 ofp11_port, uint16_t *ofp10_port); -ovs_be32 ofputil_port_to_ofp11(uint16_t ofp10_port); - -enum ofperr ofputil_check_output_port(uint16_t ofp_port, int max_ports); -uint16_t ofputil_port_from_string(const char *); -void ofputil_format_port(uint16_t port, struct ds *); +enum ofperr ofputil_port_from_ofp11(ovs_be32 ofp11_port, + ofp_port_t *ofp10_port); +ovs_be32 ofputil_port_to_ofp11(ofp_port_t ofp10_port); + +enum ofperr ofputil_check_output_port(ofp_port_t ofp_port, + ofp_port_t max_ports); +bool ofputil_port_from_string(const char *, ofp_port_t *portp); +void ofputil_format_port(ofp_port_t port, struct ds *); +void ofputil_port_to_string(ofp_port_t, char namebuf[OFP_MAX_PORT_NAME_LEN], + size_t bufsize); + +/* Group numbers. */ +enum { MAX_GROUP_NAME_LEN = INT_STRLEN(uint32_t) }; +bool ofputil_group_from_string(const char *, uint32_t *group_id); +void ofputil_format_group(uint32_t group_id, struct ds *); +void ofputil_group_to_string(uint32_t group_id, + char namebuf[MAX_GROUP_NAME_LEN + 1], + size_t bufsize); /* Converting OFPFW10_NW_SRC_MASK and OFPFW10_NW_DST_MASK wildcard bit counts * to and from IP bitmasks. */ @@ -45,12 +58,15 @@ ovs_be32 ofputil_wcbits_to_netmask(int wcbits); int ofputil_netmask_to_wcbits(ovs_be32 netmask); /* Protocols. + * + * A "protocol" is an OpenFlow version plus, for some OpenFlow versions, + * a bit extra about the flow match format in use. * * These are arranged from most portable to least portable, or alternatively - * from least powerful to most powerful. Formats earlier on the list are more - * likely to be understood for the purpose of making requests, but formats - * later on the list are more likely to accurately describe a flow within a - * switch. + * from least powerful to most powerful. Protocols earlier on the list are + * more likely to be understood for the purpose of making requests, but + * protocol later on the list are more likely to accurately describe a flow + * within a switch. * * On any given OpenFlow connection, a single protocol is in effect at any * given time. These values use separate bits only because that makes it easy @@ -58,33 +74,71 @@ int ofputil_netmask_to_wcbits(ovs_be32 netmask); * to implement set union and intersection. */ enum ofputil_protocol { - /* OpenFlow 1.0-based protocols. */ - OFPUTIL_P_OF10 = 1 << 0, /* OpenFlow 1.0 flow format. */ - OFPUTIL_P_OF10_TID = 1 << 1, /* OF1.0 + flow_mod_table_id extension. */ -#define OFPUTIL_P_OF10_ANY (OFPUTIL_P_OF10 | OFPUTIL_P_OF10_TID) - - /* OpenFlow 1.0 with NXM-based flow formats. */ - OFPUTIL_P_NXM = 1 << 2, /* Nicira extended match. */ - OFPUTIL_P_NXM_TID = 1 << 3, /* NXM + flow_mod_table_id extension. */ -#define OFPUTIL_P_NXM_ANY (OFPUTIL_P_NXM | OFPUTIL_P_NXM_TID) - - /* OpenFlow 1.2 */ - OFPUTIL_P_OF12 = 1 << 4, /* OpenFlow 1.2 flow format. */ + /* OpenFlow 1.0 protocols. + * + * The "STD" protocols use the standard OpenFlow 1.0 flow format. + * The "NXM" protocols use the Nicira Extensible Match (NXM) flow format. + * + * The protocols with "TID" mean that the nx_flow_mod_table_id Nicira + * extension has been enabled. The other protocols have it disabled. + */ +#define OFPUTIL_P_NONE 0 + OFPUTIL_P_OF10_STD = 1 << 0, + OFPUTIL_P_OF10_STD_TID = 1 << 1, + OFPUTIL_P_OF10_NXM = 1 << 2, + OFPUTIL_P_OF10_NXM_TID = 1 << 3, +#define OFPUTIL_P_OF10_STD_ANY (OFPUTIL_P_OF10_STD | OFPUTIL_P_OF10_STD_TID) +#define OFPUTIL_P_OF10_NXM_ANY (OFPUTIL_P_OF10_NXM | OFPUTIL_P_OF10_NXM_TID) +#define OFPUTIL_P_OF10_ANY (OFPUTIL_P_OF10_STD_ANY | OFPUTIL_P_OF10_NXM_ANY) + + /* OpenFlow 1.1 protocol. + * + * We only support the standard OpenFlow 1.1 flow format. + * + * OpenFlow 1.1 always operates with an equivalent of the + * nx_flow_mod_table_id Nicira extension enabled, so there is no "TID" + * variant. */ + OFPUTIL_P_OF11_STD = 1 << 4, + + /* OpenFlow 1.2+ protocols (only one variant each). + * + * These use the standard OpenFlow Extensible Match (OXM) flow format. + * + * OpenFlow 1.2+ always operates with an equivalent of the + * nx_flow_mod_table_id Nicira extension enabled, so there is no "TID" + * variant. */ + OFPUTIL_P_OF12_OXM = 1 << 5, + OFPUTIL_P_OF13_OXM = 1 << 6, +#define OFPUTIL_P_ANY_OXM (OFPUTIL_P_OF12_OXM | OFPUTIL_P_OF13_OXM) + +#define OFPUTIL_P_NXM_OF11_UP (OFPUTIL_P_OF10_NXM_ANY | OFPUTIL_P_OF11_STD | \ + OFPUTIL_P_ANY_OXM) + +#define OFPUTIL_P_NXM_OXM_ANY (OFPUTIL_P_OF10_NXM_ANY | OFPUTIL_P_ANY_OXM) + +#define OFPUTIL_P_OF11_UP (OFPUTIL_P_OF11_STD | OFPUTIL_P_ANY_OXM) + +#define OFPUTIL_P_OF12_UP (OFPUTIL_P_ANY_OXM) + +#define OFPUTIL_P_OF13_UP (OFPUTIL_P_OF13_OXM) /* All protocols. */ -#define OFPUTIL_P_ANY (OFPUTIL_P_OF10_ANY | OFPUTIL_P_NXM_ANY) +#define OFPUTIL_P_ANY ((1 << 7) - 1) /* Protocols in which a specific table may be specified in flow_mods. */ -#define OFPUTIL_P_TID (OFPUTIL_P_OF10_TID | OFPUTIL_P_NXM_TID) +#define OFPUTIL_P_TID (OFPUTIL_P_OF10_STD_TID | \ + OFPUTIL_P_OF10_NXM_TID | \ + OFPUTIL_P_OF11_STD | \ + OFPUTIL_P_ANY_OXM) }; /* Protocols to use for flow dumps, from most to least preferred. */ extern enum ofputil_protocol ofputil_flow_dump_protocols[]; extern size_t ofputil_n_flow_dump_protocols; -enum ofputil_protocol -ofputil_protocol_from_ofp_version(enum ofp_version version); -enum ofp_version ofputil_protocol_to_ofp_version(enum ofputil_protocol); +enum ofputil_protocol ofputil_protocol_from_ofp_version(enum ofp_version); +enum ofputil_protocol ofputil_protocols_from_ofp_version(enum ofp_version); +enum ofp_version ofputil_protocol_to_ofp_version(enum ofputil_protocol); bool ofputil_protocol_is_valid(enum ofputil_protocol); enum ofputil_protocol ofputil_protocol_set_tid(enum ofputil_protocol, @@ -96,7 +150,40 @@ enum ofputil_protocol ofputil_protocol_set_base( const char *ofputil_protocol_to_string(enum ofputil_protocol); char *ofputil_protocols_to_string(enum ofputil_protocol); enum ofputil_protocol ofputil_protocols_from_string(const char *); -enum ofputil_protocol ofputil_usable_protocols(const struct match *); + +void ofputil_format_version(struct ds *, enum ofp_version); +void ofputil_format_version_name(struct ds *, enum ofp_version); + +/* A bitmap of version numbers + * + * Bit offsets correspond to ofp_version numbers which in turn correspond to + * wire-protocol numbers for Open Flow versions.. E.g. (1u << OFP11_VERSION) + * is the mask for Open Flow 1.1. If the bit for a version is set then it is + * allowed, otherwise it is disallowed. */ + +void ofputil_format_version_bitmap(struct ds *msg, uint32_t bitmap); +void ofputil_format_version_bitmap_names(struct ds *msg, uint32_t bitmap); + +uint32_t ofputil_protocols_to_version_bitmap(enum ofputil_protocol); +enum ofputil_protocol ofputil_protocols_from_version_bitmap(uint32_t bitmap); + +/* Bitmap of OpenFlow versions that Open vSwitch supports. */ +#define OFPUTIL_SUPPORTED_VERSIONS \ + ((1u << OFP10_VERSION) | (1u << OFP12_VERSION) | (1u << OFP13_VERSION)) + +/* Bitmap of OpenFlow versions to enable by default (a subset of + * OFPUTIL_SUPPORTED_VERSIONS). */ +#define OFPUTIL_DEFAULT_VERSIONS (1u << OFP10_VERSION) + +enum ofputil_protocol ofputil_protocols_from_string(const char *s); + +const char *ofputil_version_to_string(enum ofp_version ofp_version); +uint32_t ofputil_versions_from_string(const char *s); +uint32_t ofputil_versions_from_strings(char ** const s, size_t count); + +bool ofputil_decode_hello(const struct ofp_header *, + uint32_t *allowed_versions); +struct ofpbuf *ofputil_encode_hello(uint32_t version_bitmap); struct ofpbuf *ofputil_encode_set_protocol(enum ofputil_protocol current, enum ofputil_protocol want, @@ -121,7 +208,10 @@ enum ofperr ofputil_pull_ofp11_match(struct ofpbuf *, struct match *, uint16_t *padded_match_len); enum ofperr ofputil_match_from_ofp11_match(const struct ofp11_match *, struct match *); +int ofputil_put_ofp11_match(struct ofpbuf *, const struct match *, + enum ofputil_protocol); void ofputil_match_to_ofp11_match(const struct match *, struct ofp11_match *); +int ofputil_match_typical_len(enum ofputil_protocol); /* dl_type translation between OpenFlow and 'struct flow' format. */ ovs_be16 ofputil_dl_type_to_openflow(ovs_be16 flow_dl_type); @@ -137,37 +227,67 @@ struct ofpbuf *ofputil_make_set_packet_in_format(enum ofp_version, /* NXT_FLOW_MOD_TABLE_ID extension. */ struct ofpbuf *ofputil_make_flow_mod_table_id(bool flow_mod_table_id); +/* Protocol-independent flow_mod flags. */ +enum ofputil_flow_mod_flags { + /* Flags that are maintained with a flow as part of its state. + * + * (OFPUTIL_FF_EMERG would be here too, if OVS supported it.) */ + OFPUTIL_FF_SEND_FLOW_REM = 1 << 0, /* All versions. */ + OFPUTIL_FF_NO_PKT_COUNTS = 1 << 1, /* OpenFlow 1.3+. */ + OFPUTIL_FF_NO_BYT_COUNTS = 1 << 2, /* OpenFlow 1.3+. */ +#define OFPUTIL_FF_STATE (OFPUTIL_FF_SEND_FLOW_REM \ + | OFPUTIL_FF_NO_PKT_COUNTS \ + | OFPUTIL_FF_NO_BYT_COUNTS) + + /* Flags that affect flow_mod behavior but are not part of flow state. */ + OFPUTIL_FF_CHECK_OVERLAP = 1 << 3, /* All versions. */ + OFPUTIL_FF_EMERG = 1 << 4, /* OpenFlow 1.0 only. */ + OFPUTIL_FF_RESET_COUNTS = 1 << 5, /* OpenFlow 1.2+. */ +}; + /* Protocol-independent flow_mod. * * The handling of cookies across multiple versions of OpenFlow is a bit - * confusing. A full description of Open vSwitch's cookie handling is - * in the DESIGN file. The following table shows the expected values of - * the cookie-related fields for the different flow_mod commands in - * OpenFlow 1.0 ("OF10") and NXM. "" and "-" indicate a value - * that may be populated and an ignored field, respectively. - * - * cookie cookie_mask new_cookie - * ====== =========== ========== - * OF10 Add - 0 - * OF10 Modify - 0 - * OF10 Delete - 0 - - * NXM Add - 0 - * NXM Modify - * NXM Delete - - */ + * confusing. See DESIGN for the details. */ struct ofputil_flow_mod { + struct list list_node; /* For queuing flow_mods. */ + struct match match; unsigned int priority; + + /* Cookie matching. The flow_mod affects only flows that have cookies that + * bitwise match 'cookie' bits in positions where 'cookie_mask has 1-bits. + * + * 'cookie_mask' should be zero for OFPFC_ADD flow_mods. */ ovs_be64 cookie; /* Cookie bits to match. */ ovs_be64 cookie_mask; /* 1-bit in each 'cookie' bit to match. */ - ovs_be64 new_cookie; /* New cookie to install or -1. */ + + /* Cookie changes. + * + * OFPFC_ADD uses 'new_cookie' as the new flow's cookie. 'new_cookie' + * should not be UINT64_MAX. + * + * OFPFC_MODIFY and OFPFC_MODIFY_STRICT have two cases: + * + * - If one or more matching flows exist and 'modify_cookie' is true, + * then the flow_mod changes the existing flows' cookies to + * 'new_cookie'. 'new_cookie' should not be UINT64_MAX. + * + * - If no matching flow exists, 'new_cookie' is not UINT64_MAX, and + * 'cookie_mask' is 0, then the flow_mod adds a new flow with + * 'new_cookie' as its cookie. + */ + ovs_be64 new_cookie; /* New cookie to install or UINT64_MAX. */ + bool modify_cookie; /* Set cookie of existing flow to 'new_cookie'? */ + uint8_t table_id; uint16_t command; uint16_t idle_timeout; uint16_t hard_timeout; uint32_t buffer_id; - uint16_t out_port; - uint16_t flags; + ofp_port_t out_port; + uint32_t out_group; + enum ofputil_flow_mod_flags flags; struct ofpact *ofpacts; /* Series of "struct ofpact"s. */ size_t ofpacts_len; /* Length of ofpacts, in bytes. */ }; @@ -179,16 +299,14 @@ enum ofperr ofputil_decode_flow_mod(struct ofputil_flow_mod *, struct ofpbuf *ofputil_encode_flow_mod(const struct ofputil_flow_mod *, enum ofputil_protocol); -enum ofputil_protocol ofputil_flow_mod_usable_protocols( - const struct ofputil_flow_mod *fms, size_t n_fms); - /* Flow stats or aggregate stats request, independent of protocol. */ struct ofputil_flow_stats_request { bool aggregate; /* Aggregate results? */ struct match match; ovs_be64 cookie; ovs_be64 cookie_mask; - uint16_t out_port; + ofp_port_t out_port; + uint32_t out_group; uint8_t table_id; }; @@ -196,8 +314,6 @@ enum ofperr ofputil_decode_flow_stats_request( struct ofputil_flow_stats_request *, const struct ofp_header *); struct ofpbuf *ofputil_encode_flow_stats_request( const struct ofputil_flow_stats_request *, enum ofputil_protocol); -enum ofputil_protocol ofputil_flow_stats_request_usable_protocols( - const struct ofputil_flow_stats_request *); /* Flow stats reply, independent of protocol. */ struct ofputil_flow_stats { @@ -215,6 +331,7 @@ struct ofputil_flow_stats { uint64_t byte_count; /* Byte count, UINT64_MAX if unknown. */ struct ofpact *ofpacts; size_t ofpacts_len; + enum ofputil_flow_mod_flags flags; }; int ofputil_decode_flow_stats_reply(struct ofputil_flow_stats *, @@ -260,19 +377,33 @@ struct ofpbuf *ofputil_encode_flow_removed(const struct ofputil_flow_removed *, /* Abstract packet-in message. */ struct ofputil_packet_in { + /* Packet data and metadata. + * + * To save bandwidth, in some cases a switch may send only the first + * several bytes of a packet, indicated by 'packet_len < total_len'. When + * the full packet is included, 'packet_len == total_len'. */ const void *packet; - size_t packet_len; + size_t packet_len; /* Number of bytes in 'packet'. */ + size_t total_len; /* Size of packet, pre-truncation. */ + struct flow_metadata fmd; + + /* Identifies a buffer in the switch that contains the full packet, to + * allow the controller to reference it later without having to send the + * entire packet back to the switch. + * + * UINT32_MAX indicates that the packet is not buffered in the switch. A + * switch should only use UINT32_MAX when it sends the entire packet. */ + uint32_t buffer_id; + /* Reason that the packet-in is being sent. */ enum ofp_packet_in_reason reason; /* One of OFPR_*. */ - uint16_t controller_id; /* Controller ID to send to. */ - uint8_t table_id; - ovs_be64 cookie; - uint32_t buffer_id; - int send_len; - uint16_t total_len; /* Full length of frame. */ - - struct flow_metadata fmd; /* Metadata at creation time. */ + /* Information about the OpenFlow flow that triggered the packet-in. + * + * A packet-in triggered by a flow table miss has no associated flow. In + * that case, 'cookie' is UINT64_MAX. */ + uint8_t table_id; /* OpenFlow table ID. */ + ovs_be64 cookie; /* Flow's cookie. */ }; enum ofperr ofputil_decode_packet_in(struct ofputil_packet_in *, @@ -281,7 +412,10 @@ struct ofpbuf *ofputil_encode_packet_in(const struct ofputil_packet_in *, enum ofputil_protocol protocol, enum nx_packet_in_format); -const char *ofputil_packet_in_reason_to_string(enum ofp_packet_in_reason); +enum { OFPUTIL_PACKET_IN_REASON_BUFSIZE = INT_STRLEN(int) + 1 }; +const char *ofputil_packet_in_reason_to_string(enum ofp_packet_in_reason, + char *reasonbuf, + size_t bufsize); bool ofputil_packet_in_reason_from_string(const char *, enum ofp_packet_in_reason *); @@ -293,7 +427,7 @@ struct ofputil_packet_out { const void *packet; /* Packet data, if buffer_id == UINT32_MAX. */ size_t packet_len; /* Length of packet data in bytes. */ uint32_t buffer_id; /* Buffer id or UINT32_MAX if no buffer. */ - uint16_t in_port; /* Packet's input port. */ + ofp_port_t in_port; /* Packet's input port. */ struct ofpact *ofpacts; /* Actions. */ size_t ofpacts_len; /* Size of ofpacts in bytes. */ }; @@ -333,7 +467,7 @@ enum ofputil_port_state { /* Abstract ofp10_phy_port or ofp11_port. */ struct ofputil_phy_port { - uint16_t port_no; + ofp_port_t port_no; uint8_t hw_addr[OFP_ETH_ALEN]; char name[OFP_MAX_PORT_NAME_LEN]; enum ofputil_port_config config; @@ -351,7 +485,7 @@ struct ofputil_phy_port { }; enum ofputil_capabilities { - /* OpenFlow 1.0, 1.1 and 1.2 share these values for these capabilities. */ + /* OpenFlow 1.0, 1.1, 1.2, and 1.3 share these capability values. */ OFPUTIL_C_FLOW_STATS = 1 << 0, /* Flow statistics. */ OFPUTIL_C_TABLE_STATS = 1 << 1, /* Table statistics. */ OFPUTIL_C_PORT_STATS = 1 << 2, /* Port statistics. */ @@ -364,10 +498,10 @@ enum ofputil_capabilities { /* OpenFlow 1.0 only. */ OFPUTIL_C_STP = 1 << 3, /* 802.1d spanning tree. */ - /* OpenFlow 1.1 and 1.2 share this capability. */ + /* OpenFlow 1.1, 1.2, and 1.3 share this capability. */ OFPUTIL_C_GROUP_STATS = 1 << 4, /* Group statistics. */ - /* OpenFlow 1.2 only */ + /* OpenFlow 1.2 and 1.3 share this capability */ OFPUTIL_C_PORT_BLOCKED = 1 << 8, /* Switch will block looping ports */ }; @@ -407,6 +541,7 @@ struct ofputil_switch_features { uint64_t datapath_id; /* Datapath unique ID. */ uint32_t n_buffers; /* Max packets buffered at once. */ uint8_t n_tables; /* Number of tables supported by datapath. */ + uint8_t auxiliary_id; /* Identify auxiliary connections */ enum ofputil_capabilities capabilities; enum ofputil_action_bitmap actions; }; @@ -440,7 +575,7 @@ struct ofpbuf *ofputil_encode_port_status(const struct ofputil_port_status *, /* Abstract ofp_port_mod. */ struct ofputil_port_mod { - uint16_t port_no; + ofp_port_t port_no; uint8_t hw_addr[OFP_ETH_ALEN]; enum ofputil_port_config config; enum ofputil_port_config mask; @@ -452,6 +587,116 @@ enum ofperr ofputil_decode_port_mod(const struct ofp_header *, struct ofpbuf *ofputil_encode_port_mod(const struct ofputil_port_mod *, enum ofputil_protocol); +/* Abstract ofp_table_mod. */ +struct ofputil_table_mod { + uint8_t table_id; /* ID of the table, 0xff indicates all tables. */ + uint32_t config; +}; + +enum ofperr ofputil_decode_table_mod(const struct ofp_header *, + struct ofputil_table_mod *); +struct ofpbuf *ofputil_encode_table_mod(const struct ofputil_table_mod *, + enum ofputil_protocol); + +/* Meter band configuration for all supported band types. */ +struct ofputil_meter_band { + uint16_t type; + uint8_t prec_level; /* Non-zero if type == OFPMBT_DSCP_REMARK. */ + uint32_t rate; + uint32_t burst_size; +}; + +struct ofputil_meter_band_stats { + uint64_t packet_count; + uint64_t byte_count; +}; + +struct ofputil_meter_config { + uint32_t meter_id; + uint16_t flags; + uint16_t n_bands; + struct ofputil_meter_band *bands; +}; + +/* Abstract ofp_meter_mod. */ +struct ofputil_meter_mod { + uint16_t command; + struct ofputil_meter_config meter; +}; + +struct ofputil_meter_stats { + uint32_t meter_id; + uint32_t flow_count; + uint64_t packet_in_count; + uint64_t byte_in_count; + uint32_t duration_sec; + uint32_t duration_nsec; + uint16_t n_bands; + struct ofputil_meter_band_stats *bands; +}; + +struct ofputil_meter_features { + uint32_t max_meters; /* Maximum number of meters. */ + uint32_t band_types; /* Can support max 32 band types. */ + uint32_t capabilities; /* Supported flags. */ + uint8_t max_bands; + uint8_t max_color; +}; + +enum ofperr ofputil_decode_meter_mod(const struct ofp_header *, + struct ofputil_meter_mod *, + struct ofpbuf *bands); +struct ofpbuf *ofputil_encode_meter_mod(enum ofp_version, + const struct ofputil_meter_mod *); + +void ofputil_decode_meter_features(const struct ofp_header *, + struct ofputil_meter_features *); +struct ofpbuf *ofputil_encode_meter_features_reply(const struct + ofputil_meter_features *, + const struct ofp_header * + request); +void ofputil_decode_meter_request(const struct ofp_header *, + uint32_t *meter_id); + +void ofputil_append_meter_config(struct list *replies, + const struct ofputil_meter_config *); + +void ofputil_append_meter_stats(struct list *replies, + const struct ofputil_meter_stats *); + +enum ofputil_meter_request_type { + OFPUTIL_METER_FEATURES, + OFPUTIL_METER_CONFIG, + OFPUTIL_METER_STATS +}; + +struct ofpbuf *ofputil_encode_meter_request(enum ofp_version, + enum ofputil_meter_request_type, + uint32_t meter_id); + +int ofputil_decode_meter_stats(struct ofpbuf *, + struct ofputil_meter_stats *, + struct ofpbuf *bands); + +int ofputil_decode_meter_config(struct ofpbuf *, + struct ofputil_meter_config *, + struct ofpbuf *bands); + +/* Type for meter_id in ofproto provider interface, UINT32_MAX if invalid. */ +typedef struct { uint32_t uint32; } ofproto_meter_id; + +/* Abstract ofp_role_request and reply. */ +struct ofputil_role_request { + enum ofp12_controller_role role; + bool have_generation_id; + uint64_t generation_id; +}; + +enum ofperr ofputil_decode_role_message(const struct ofp_header *, + struct ofputil_role_request *); +struct ofpbuf *ofputil_encode_role_reply(const struct ofp_header *, + const struct ofputil_role_request *); + /* Abstract table stats. * * For now we use ofp12_table_stats as a superset of the other protocol @@ -465,7 +710,7 @@ struct ofpbuf *ofputil_encode_table_stats_reply( struct ofputil_flow_monitor_request { uint32_t id; enum nx_flow_monitor_flags flags; - uint16_t out_port; + ofp_port_t out_port; uint8_t table_id; struct match match; }; @@ -540,6 +785,7 @@ bool ofputil_frag_handling_from_string(const char *, enum ofp_config_flags *); * OFPUTIL_OFPAT10_ENQUEUE * OFPUTIL_NXAST_RESUBMIT * OFPUTIL_NXAST_SET_TUNNEL + * OFPUTIL_NXAST_SET_METADATA * OFPUTIL_NXAST_SET_QUEUE * OFPUTIL_NXAST_POP_QUEUE * OFPUTIL_NXAST_REG_MOVE @@ -547,7 +793,6 @@ bool ofputil_frag_handling_from_string(const char *, enum ofp_config_flags *); * OFPUTIL_NXAST_NOTE * OFPUTIL_NXAST_SET_TUNNEL64 * OFPUTIL_NXAST_MULTIPATH - * OFPUTIL_NXAST_AUTOPATH * OFPUTIL_NXAST_BUNDLE * OFPUTIL_NXAST_BUNDLE_LOAD * OFPUTIL_NXAST_RESUBMIT_TABLE @@ -606,8 +851,6 @@ void *ofputil_put_action(enum ofputil_action_code, struct ofpbuf *buf); #define OFP_ACTION_ALIGN 8 /* Alignment of ofp_actions. */ -enum ofperr validate_actions(const union ofp_action *, size_t n_actions, - const struct flow *, int max_ports); bool action_outputs_to_port(const union ofp_action *, ovs_be16 port); enum ofperr ofputil_pull_actions(struct ofpbuf *, unsigned int actions_len, @@ -620,24 +863,24 @@ union ofp_action *ofputil_actions_clone(const union ofp_action *, size_t n); /* Handy utility for parsing flows and actions. */ bool ofputil_parse_key_value(char **stringp, char **keyp, char **valuep); -struct ofpbuf *ofputlil_dump_ports(enum ofp_version ofp_version, int16_t port); - struct ofputil_port_stats { - uint16_t port_no; + ofp_port_t port_no; struct netdev_stats stats; + uint32_t duration_sec; /* UINT32_MAX if unknown. */ + uint32_t duration_nsec; }; struct ofpbuf *ofputil_encode_dump_ports_request(enum ofp_version ofp_version, - int16_t port); + ofp_port_t port); void ofputil_append_port_stat(struct list *replies, const struct ofputil_port_stats *ops); size_t ofputil_count_port_stats(const struct ofp_header *); int ofputil_decode_port_stats(struct ofputil_port_stats *, struct ofpbuf *msg); enum ofperr ofputil_decode_port_stats_request(const struct ofp_header *request, - uint16_t *ofp10_port); + ofp_port_t *ofp10_port); struct ofputil_queue_stats_request { - uint16_t port_no; + ofp_port_t port_no; /* OFPP_ANY means "all ports". */ uint32_t queue_id; }; @@ -649,9 +892,17 @@ ofputil_encode_queue_stats_request(enum ofp_version ofp_version, const struct ofputil_queue_stats_request *oqsr); struct ofputil_queue_stats { - uint16_t port_no; + ofp_port_t port_no; uint32_t queue_id; - struct netdev_queue_stats stats; + + /* Values of unsupported statistics are set to all-1-bits (UINT64_MAX). */ + uint64_t tx_bytes; + uint64_t tx_packets; + uint64_t tx_errors; + + /* UINT32_MAX if unknown. */ + uint32_t duration_sec; + uint32_t duration_nsec; }; size_t ofputil_count_queue_stats(const struct ofp_header *); @@ -659,4 +910,88 @@ int ofputil_decode_queue_stats(struct ofputil_queue_stats *qs, struct ofpbuf *ms void ofputil_append_queue_stat(struct list *replies, const struct ofputil_queue_stats *oqs); +/* Bucket for use in groups. */ +struct ofputil_bucket { + struct list list_node; + uint16_t weight; /* Relative weight, for "select" groups. */ + ofp_port_t watch_port; /* Port whose state affects whether this bucket + * is live. Only required for fast failover + * groups. */ + uint32_t watch_group; /* Group whose state affects whether this + * bucket is live. Only required for fast + * failover groups. */ + struct ofpact *ofpacts; /* Series of "struct ofpact"s. */ + size_t ofpacts_len; /* Length of ofpacts, in bytes. */ +}; + +/* Protocol-independent group_mod. */ +struct ofputil_group_mod { + uint16_t command; /* One of OFPGC11_*. */ + uint8_t type; /* One of OFPGT11_*. */ + uint32_t group_id; /* Group identifier. */ + struct list buckets; /* Contains "struct ofputil_bucket"s. */ +}; + +struct bucket_counter { + uint64_t packet_count; /* Number of packets processed by bucket. */ + uint64_t byte_count; /* Number of bytes processed by bucket. */ +}; + +/* Group stats reply, independent of protocol. */ +struct ofputil_group_stats { + uint32_t group_id; /* Group identifier. */ + uint32_t ref_count; + uint64_t packet_count; /* Packet count, UINT64_MAX if unknown. */ + uint64_t byte_count; /* Byte count, UINT64_MAX if unknown. */ + uint32_t duration_sec; /* UINT32_MAX if unknown. */ + uint32_t duration_nsec; + uint32_t n_buckets; + struct bucket_counter *bucket_stats; +}; + +/* Group features reply, independent of protocol. */ +struct ofputil_group_features { + uint32_t types; /* Bitmap of OFPGT_* values supported. */ + uint32_t capabilities; /* Bitmap of OFPGFC12_* capability supported. */ + uint32_t max_groups[4]; /* Maximum number of groups for each type. */ + uint32_t actions[4]; /* Bitmaps of OFPAT_* that are supported. */ +}; + +/* Group desc reply, independent of protocol. */ +struct ofputil_group_desc { + uint8_t type; /* One of OFPGT_*. */ + uint32_t group_id; /* Group identifier. */ + struct list buckets; /* Contains "struct ofputil_bucket"s. */ +}; + +void ofputil_bucket_list_destroy(struct list *buckets); + +struct ofpbuf *ofputil_encode_group_stats_request(enum ofp_version, + uint32_t group_id); +enum ofperr ofputil_decode_group_stats_request( + const struct ofp_header *request, uint32_t *group_id); +void ofputil_append_group_stats(struct list *replies, + const struct ofputil_group_stats *); +struct ofpbuf *ofputil_encode_group_features_request(enum ofp_version); +struct ofpbuf *ofputil_encode_group_features_reply( + const struct ofputil_group_features *, const struct ofp_header *request); +void ofputil_decode_group_features_reply(const struct ofp_header *, + struct ofputil_group_features *); +struct ofpbuf *ofputil_encode_group_mod(enum ofp_version ofp_version, + const struct ofputil_group_mod *gm); + +enum ofperr ofputil_decode_group_mod(const struct ofp_header *, + struct ofputil_group_mod *); + +int ofputil_decode_group_stats_reply(struct ofpbuf *, + struct ofputil_group_stats *); + +int ofputil_decode_group_desc_reply(struct ofputil_group_desc *, + struct ofpbuf *, enum ofp_version); + +void ofputil_append_group_desc_reply(const struct ofputil_group_desc *, + struct list *buckets, + struct list *replies); +struct ofpbuf *ofputil_encode_group_desc_request(enum ofp_version); + #endif /* ofp-util.h */