/*
- * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014 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.
#include <stdbool.h>
#include <stdint.h>
#include <string.h>
+#include "bitmap.h"
#include "byte-order.h"
#include "openflow/nicira-ext.h"
#include "openflow/openflow.h"
struct ds;
struct flow_wildcards;
struct minimask;
-struct ofpbuf;
+struct dp_packet;
struct pkt_metadata;
+struct match;
/* This sequence number should be incremented whenever anything involving flows
* or the wildcarding of flows changes. This will cause build assertion
* failures in places which likely need to be updated. */
-#define FLOW_WC_SEQ 27
+#define FLOW_WC_SEQ 35
/* Number of Open vSwitch extension 32-bit registers. */
#define FLOW_N_REGS 8
BUILD_ASSERT_DECL(FLOW_N_REGS <= NXM_NX_MAX_REGS);
+BUILD_ASSERT_DECL(FLOW_N_REGS % 2 == 0); /* Even. */
/* Number of OpenFlow 1.5+ 64-bit registers.
*
BUILD_ASSERT_DECL(FLOW_NW_FRAG_ANY == NX_IP_FRAG_ANY);
BUILD_ASSERT_DECL(FLOW_NW_FRAG_LATER == NX_IP_FRAG_LATER);
-#define FLOW_TNL_F_DONT_FRAGMENT (1 << 0)
-#define FLOW_TNL_F_CSUM (1 << 1)
-#define FLOW_TNL_F_KEY (1 << 2)
-#define FLOW_TNL_F_OAM (1 << 3)
-
-#define FLOW_TNL_F_MASK ((1 << 4) - 1)
+BUILD_ASSERT_DECL(FLOW_TNL_F_OAM == NX_TUN_FLAG_OAM);
const char *flow_tun_flag_to_string(uint32_t flags);
uint32_t regs[FLOW_N_REGS]; /* Registers. */
uint32_t skb_priority; /* Packet priority for QoS. */
uint32_t pkt_mark; /* Packet mark. */
- uint32_t recirc_id; /* Must be exact match. */
+ uint32_t dp_hash; /* Datapath computed hash value. The exact
+ * computation is opaque to the user space. */
union flow_in_port in_port; /* Input port.*/
-
- /* L2, Order the same as in the Ethernet header! */
- uint8_t dl_dst[ETH_ADDR_LEN]; /* Ethernet destination address. */
- uint8_t dl_src[ETH_ADDR_LEN]; /* Ethernet source address. */
+ uint32_t recirc_id; /* Must be exact match. */
+ uint16_t ct_state; /* Connection tracking state. */
+ uint16_t ct_zone; /* Connection tracking zone. */
+ uint32_t ct_mark; /* Connection mark.*/
+ uint8_t pad1[4]; /* Pad to 64 bits. */
+ ovs_u128 ct_label; /* Connection label. */
+ uint32_t conj_id; /* Conjunction ID. */
+ ofp_port_t actset_output; /* Output port in action set. */
+ uint8_t pad2[2]; /* Pad to 64 bits. */
+
+ /* L2, Order the same as in the Ethernet header! (64-bit aligned) */
+ struct eth_addr dl_dst; /* Ethernet destination address. */
+ struct eth_addr dl_src; /* Ethernet source address. */
ovs_be16 dl_type; /* Ethernet frame type. */
ovs_be16 vlan_tci; /* If 802.1Q, TCI | VLAN_CFI; otherwise 0. */
- ovs_be32 mpls_lse[FLOW_MAX_MPLS_LABELS]; /* MPLS label stack entry. */
-
- /* L3 */
+ ovs_be32 mpls_lse[ROUND_UP(FLOW_MAX_MPLS_LABELS, 2)]; /* MPLS label stack
+ (with padding). */
+ /* L3 (64-bit aligned) */
+ ovs_be32 nw_src; /* IPv4 source address. */
+ ovs_be32 nw_dst; /* IPv4 destination address. */
struct in6_addr ipv6_src; /* IPv6 source address. */
struct in6_addr ipv6_dst; /* IPv6 destination address. */
ovs_be32 ipv6_label; /* IPv6 flow label. */
- ovs_be32 nw_src; /* IPv4 source address. */
- ovs_be32 nw_dst; /* IPv4 destination address. */
uint8_t nw_frag; /* FLOW_FRAG_* flags. */
uint8_t nw_tos; /* IP ToS (including DSCP and ECN). */
uint8_t nw_ttl; /* IP TTL/Hop Limit. */
uint8_t nw_proto; /* IP protocol or low 8 bits of ARP opcode. */
- uint8_t arp_sha[ETH_ADDR_LEN]; /* ARP/ND source hardware address. */
- uint8_t arp_tha[ETH_ADDR_LEN]; /* ARP/ND target hardware address. */
struct in6_addr nd_target; /* IPv6 neighbor discovery (ND) target. */
+ struct eth_addr arp_sha; /* ARP/ND source hardware address. */
+ struct eth_addr arp_tha; /* ARP/ND target hardware address. */
ovs_be16 tcp_flags; /* TCP flags. With L3 to avoid matching L4. */
- ovs_be16 pad; /* Padding. */
+ ovs_be16 pad3; /* Pad to 64 bits. */
- /* L4 */
- ovs_be16 tp_src; /* TCP/UDP/SCTP source port. */
- ovs_be16 tp_dst; /* TCP/UDP/SCTP destination port. */
- ovs_be32 igmp_group_ip4; /* IGMP group IPv4 address */
- uint32_t dp_hash; /* Datapath computed hash value. The exact
- * computation is opaque to the user space.
+ /* L4 (64-bit aligned) */
+ ovs_be16 tp_src; /* TCP/UDP/SCTP source port/ICMP type. */
+ ovs_be16 tp_dst; /* TCP/UDP/SCTP destination port/ICMP code. */
+ ovs_be32 igmp_group_ip4; /* IGMP group IPv4 address.
* Keep last for BUILD_ASSERT_DECL below. */
};
-BUILD_ASSERT_DECL(sizeof(struct flow) % 4 == 0);
+BUILD_ASSERT_DECL(sizeof(struct flow) % sizeof(uint64_t) == 0);
+BUILD_ASSERT_DECL(sizeof(struct flow_tnl) % sizeof(uint64_t) == 0);
-#define FLOW_U32S (sizeof(struct flow) / 4)
+#define FLOW_U64S (sizeof(struct flow) / sizeof(uint64_t))
/* Some flow fields are mutually exclusive or only appear within the flow
* pipeline. IPv6 headers are bigger than IPv4 and MPLS, and IPv6 ND packets
* are bigger than TCP,UDP and IGMP packets. */
-#define FLOW_MAX_PACKET_U32S (FLOW_U32S \
- /* Unused in datapath */ - FLOW_U32_SIZE(regs) \
- - FLOW_U32_SIZE(metadata) \
- /* L2.5/3 */ - FLOW_U32_SIZE(nw_src) \
- - FLOW_U32_SIZE(nw_dst) \
- - FLOW_U32_SIZE(mpls_lse) \
- /* L4 */ - FLOW_U32_SIZE(tcp_flags) /* incl. pad. */ \
- - FLOW_U32_SIZE(igmp_group_ip4) \
+#define FLOW_MAX_PACKET_U64S (FLOW_U64S \
+ /* Unused in datapath */ - FLOW_U64_SIZE(regs) \
+ - FLOW_U64_SIZE(metadata) \
+ /* L2.5/3 */ - FLOW_U64_SIZE(nw_src) /* incl. nw_dst */ \
+ - FLOW_U64_SIZE(mpls_lse) \
+ /* L4 */ - FLOW_U64_SIZE(tp_src) \
)
/* Remember to update FLOW_WC_SEQ when changing 'struct flow'. */
-BUILD_ASSERT_DECL(offsetof(struct flow, dp_hash) + sizeof(uint32_t)
- == sizeof(struct flow_tnl) + 176
- && FLOW_WC_SEQ == 27);
+BUILD_ASSERT_DECL(offsetof(struct flow, igmp_group_ip4) + sizeof(uint32_t)
+ == sizeof(struct flow_tnl) + 216
+ && FLOW_WC_SEQ == 35);
/* Incremental points at which flow classification may be performed in
* segments.
* This is located here since this is dependent on the structure of the
* struct flow defined above:
- * Each offset must be on a distinct, successive U32 boundary strictly
+ * Each offset must be on a distinct, successive U64 boundary strictly
* within the struct flow. */
enum {
FLOW_SEGMENT_1_ENDS_AT = offsetof(struct flow, dl_dst),
- FLOW_SEGMENT_2_ENDS_AT = offsetof(struct flow, ipv6_src),
+ FLOW_SEGMENT_2_ENDS_AT = offsetof(struct flow, nw_src),
FLOW_SEGMENT_3_ENDS_AT = offsetof(struct flow, tp_src),
};
-BUILD_ASSERT_DECL(FLOW_SEGMENT_1_ENDS_AT % 4 == 0);
-BUILD_ASSERT_DECL(FLOW_SEGMENT_2_ENDS_AT % 4 == 0);
-BUILD_ASSERT_DECL(FLOW_SEGMENT_3_ENDS_AT % 4 == 0);
+BUILD_ASSERT_DECL(FLOW_SEGMENT_1_ENDS_AT % sizeof(uint64_t) == 0);
+BUILD_ASSERT_DECL(FLOW_SEGMENT_2_ENDS_AT % sizeof(uint64_t) == 0);
+BUILD_ASSERT_DECL(FLOW_SEGMENT_3_ENDS_AT % sizeof(uint64_t) == 0);
BUILD_ASSERT_DECL( 0 < FLOW_SEGMENT_1_ENDS_AT);
BUILD_ASSERT_DECL(FLOW_SEGMENT_1_ENDS_AT < FLOW_SEGMENT_2_ENDS_AT);
BUILD_ASSERT_DECL(FLOW_SEGMENT_2_ENDS_AT < FLOW_SEGMENT_3_ENDS_AT);
BUILD_ASSERT_DECL(FLOW_SEGMENT_3_ENDS_AT < sizeof(struct flow));
-extern const uint8_t flow_segment_u32s[];
-
-/* Represents the metadata fields of struct flow. */
-struct flow_metadata {
- uint32_t dp_hash; /* Datapath computed hash field. */
- uint32_t recirc_id; /* Recirculation ID. */
- ovs_be64 tun_id; /* Encapsulating tunnel ID. */
- ovs_be32 tun_src; /* Tunnel outer IPv4 src addr */
- ovs_be32 tun_dst; /* Tunnel outer IPv4 dst addr */
- ovs_be64 metadata; /* OpenFlow 1.1+ metadata field. */
- uint32_t regs[FLOW_N_REGS]; /* Registers. */
- uint32_t pkt_mark; /* Packet mark. */
- ofp_port_t in_port; /* OpenFlow port or zero. */
-};
+extern const uint8_t flow_segment_u64s[];
+
+#define FLOW_U64_OFFSET(FIELD) \
+ (offsetof(struct flow, FIELD) / sizeof(uint64_t))
+#define FLOW_U64_OFFREM(FIELD) \
+ (offsetof(struct flow, FIELD) % sizeof(uint64_t))
+
+/* Number of 64-bit units spanned by a 'FIELD'. */
+#define FLOW_U64_SIZE(FIELD) \
+ DIV_ROUND_UP(FLOW_U64_OFFREM(FIELD) + MEMBER_SIZEOF(struct flow, FIELD), \
+ sizeof(uint64_t))
-void flow_extract(struct ofpbuf *, const struct pkt_metadata *md,
- struct flow *);
+void flow_extract(struct dp_packet *, struct flow *);
void flow_zero_wildcards(struct flow *, const struct flow_wildcards *);
void flow_unwildcard_tp_ports(const struct flow *, struct flow_wildcards *);
-void flow_get_metadata(const struct flow *, struct flow_metadata *);
+void flow_get_metadata(const struct flow *, struct match *flow_metadata);
+const char *ct_state_to_string(uint32_t state);
char *flow_to_string(const struct flow *);
void format_flags(struct ds *ds, const char *(*bit_to_string)(uint32_t),
uint32_t flags, char del);
void format_flags_masked(struct ds *ds, const char *name,
const char *(*bit_to_string)(uint32_t),
- uint32_t flags, uint32_t mask);
+ uint32_t flags, uint32_t mask, uint32_t max_mask);
+int parse_flags(const char *s, const char *(*bit_to_string)(uint32_t),
+ char end, const char *field_name, char **res_string,
+ uint32_t *res_flags, uint32_t allowed, uint32_t *res_mask);
void flow_format(struct ds *, const struct flow *);
void flow_print(FILE *, const struct flow *);
void flow_set_mpls_bos(struct flow *, int idx, uint8_t stack);
void flow_set_mpls_lse(struct flow *, int idx, ovs_be32 lse);
-void flow_compose(struct ofpbuf *, const struct flow *);
+void flow_compose(struct dp_packet *, const struct flow *);
static inline uint64_t
flow_get_xreg(const struct flow *flow, int idx)
static inline size_t
flow_hash(const struct flow *flow, uint32_t basis)
{
- return hash_words((const uint32_t *) flow, sizeof *flow / 4, basis);
+ return hash_bytes64((const uint64_t *)flow, sizeof *flow, basis);
}
static inline uint16_t
#define WC_MASK_FIELD(WC, FIELD) \
memset(&(WC)->masks.FIELD, 0xff, sizeof (WC)->masks.FIELD)
+#define WC_MASK_FIELD_MASK(WC, FIELD, MASK) \
+ ((WC)->masks.FIELD |= (MASK))
#define WC_UNMASK_FIELD(WC, FIELD) \
memset(&(WC)->masks.FIELD, 0, sizeof (WC)->masks.FIELD)
void flow_wildcards_init_for_packet(struct flow_wildcards *,
const struct flow *);
-uint64_t flow_wc_map(const struct flow *);
void flow_wildcards_clear_non_packet_fields(struct flow_wildcards *);
const struct flow_wildcards *);
uint32_t flow_hash_5tuple(const struct flow *flow, uint32_t basis);
uint32_t flow_hash_symmetric_l4(const struct flow *flow, uint32_t basis);
+uint32_t flow_hash_symmetric_l3l4(const struct flow *flow, uint32_t basis,
+ bool inc_udp_ports );
/* Initialize a flow with random fields that matter for nx_hash_fields. */
void flow_random_hash_fields(struct flow *);
bool flow_equal_except(const struct flow *a, const struct flow *b,
const struct flow_wildcards *);
\f
-/* Compressed flow. */
+/* Bitmap for flow values. For each 1-bit the corresponding flow value is
+ * explicitly specified, other values are zeroes.
+ *
+ * map_t must be wide enough to hold any member of struct flow. */
+typedef unsigned long long map_t;
+#define MAP_T_BITS (sizeof(map_t) * CHAR_BIT)
+#define MAP_1 (map_t)1
+#define MAP_MAX TYPE_MAXIMUM(map_t)
-/* Number of 32-bit words present in struct miniflow. */
-#define MINI_N_INLINE 8
+#define MAP_IS_SET(MAP, IDX) ((MAP) & (MAP_1 << (IDX)))
-/* Maximum number of 32-bit words supported. */
-BUILD_ASSERT_DECL(FLOW_U32S <= 63);
+/* Iterate through the indices of all 1-bits in 'MAP'. */
+#define MAP_FOR_EACH_INDEX(IDX, MAP) \
+ ULLONG_FOR_EACH_1(IDX, MAP)
-/* A sparse representation of a "struct flow".
- *
- * A "struct flow" is fairly large and tends to be mostly zeros. Sparse
- * representation has two advantages. First, it saves memory. Second, it
- * saves time when the goal is to iterate over only the nonzero parts of the
- * struct.
- *
- * The 'map' member holds one bit for each uint32_t in a "struct flow". Each
- * 0-bit indicates that the corresponding uint32_t is zero, each 1-bit that it
- * *may* be nonzero (see below how this applies to minimasks).
- *
- * The 'values_inline' boolean member indicates that the values are at
- * 'inline_values'. If 'values_inline' is zero, then the values are
- * offline at 'offline_values'. In either case, values is an array that has
- * one element for each 1-bit in 'map'. The least-numbered 1-bit is in
- * the first element of the values array, the next 1-bit is in the next array
- * element, and so on.
- *
- * MINI_N_INLINE is the default number of inline words. When a miniflow is
- * dynamically allocated the actual amount of inline storage may be different.
- * In that case 'inline_values' contains storage at least for the number
- * of words indicated by 'map' (one uint32_t for each 1-bit in the map).
- *
- * Elements in values array are allowed to be zero. This is useful for "struct
- * minimatch", for which ensuring that the miniflow and minimask members have
- * same 'map' allows optimization. This allowance applies only to a miniflow
- * that is not a mask. That is, a minimask may NOT have zero elements in
- * its 'values'.
- */
-struct miniflow {
- uint64_t map:63;
- uint64_t values_inline:1;
- union {
- uint32_t *offline_values;
- uint32_t inline_values[MINI_N_INLINE]; /* Minimum inline size. */
- };
+#define FLOWMAP_UNITS DIV_ROUND_UP(FLOW_U64S, MAP_T_BITS)
+
+struct flowmap {
+ map_t bits[FLOWMAP_UNITS];
};
-BUILD_ASSERT_DECL(sizeof(struct miniflow)
- == sizeof(uint64_t) + MINI_N_INLINE * sizeof(uint32_t));
-#define MINIFLOW_VALUES_SIZE(COUNT) ((COUNT) * sizeof(uint32_t))
+#define FLOWMAP_EMPTY_INITIALIZER { { 0 } }
+
+static inline void flowmap_init(struct flowmap *);
+static inline bool flowmap_equal(struct flowmap, struct flowmap);
+static inline bool flowmap_is_set(const struct flowmap *, size_t idx);
+static inline bool flowmap_are_set(const struct flowmap *, size_t idx,
+ unsigned int n_bits);
+static inline void flowmap_set(struct flowmap *, size_t idx,
+ unsigned int n_bits);
+static inline void flowmap_clear(struct flowmap *, size_t idx,
+ unsigned int n_bits);
+static inline struct flowmap flowmap_or(struct flowmap, struct flowmap);
+static inline struct flowmap flowmap_and(struct flowmap, struct flowmap);
+static inline bool flowmap_is_empty(struct flowmap);
+static inline unsigned int flowmap_n_1bits(struct flowmap);
+
+#define FLOWMAP_HAS_FIELD(FM, FIELD) \
+ flowmap_are_set(FM, FLOW_U64_OFFSET(FIELD), FLOW_U64_SIZE(FIELD))
+
+#define FLOWMAP_SET(FM, FIELD) \
+ flowmap_set(FM, FLOW_U64_OFFSET(FIELD), FLOW_U64_SIZE(FIELD))
+
+#define FLOWMAP_SET__(FM, FIELD, SIZE) \
+ flowmap_set(FM, FLOW_U64_OFFSET(FIELD), \
+ DIV_ROUND_UP(SIZE, sizeof(uint64_t)))
+
+/* XXX: Only works for full 64-bit units. */
+#define FLOWMAP_CLEAR(FM, FIELD) \
+ BUILD_ASSERT_DECL(FLOW_U64_OFFREM(FIELD) == 0); \
+ BUILD_ASSERT_DECL(sizeof(((struct flow *)0)->FIELD) % sizeof(uint64_t) == 0); \
+ flowmap_clear(FM, FLOW_U64_OFFSET(FIELD), FLOW_U64_SIZE(FIELD))
+
+/* Iterate through all units in 'FMAP'. */
+#define FLOWMAP_FOR_EACH_UNIT(UNIT) \
+ for ((UNIT) = 0; (UNIT) < FLOWMAP_UNITS; (UNIT)++)
+
+/* Iterate through all map units in 'FMAP'. */
+#define FLOWMAP_FOR_EACH_MAP(MAP, FLOWMAP) \
+ for (size_t unit__ = 0; \
+ unit__ < FLOWMAP_UNITS && ((MAP) = (FLOWMAP).bits[unit__], true); \
+ unit__++)
+
+struct flowmap_aux;
+static inline bool flowmap_next_index(struct flowmap_aux *, size_t *idx);
+
+#define FLOWMAP_AUX_INITIALIZER(FLOWMAP) { .unit = 0, .map = (FLOWMAP) }
+
+/* Iterate through all struct flow u64 indices specified by 'MAP'. This is a
+ * slower but easier version of the FLOWMAP_FOR_EACH_MAP() &
+ * MAP_FOR_EACH_INDEX() combination. */
+#define FLOWMAP_FOR_EACH_INDEX(IDX, MAP) \
+ for (struct flowmap_aux aux__ = FLOWMAP_AUX_INITIALIZER(MAP); \
+ flowmap_next_index(&aux__, &(IDX));)
+
+/* Flowmap inline implementations. */
+static inline void
+flowmap_init(struct flowmap *fm)
+{
+ memset(fm, 0, sizeof *fm);
+}
+
+static inline bool
+flowmap_equal(struct flowmap a, struct flowmap b)
+{
+ return !memcmp(&a, &b, sizeof a);
+}
+
+static inline bool
+flowmap_is_set(const struct flowmap *fm, size_t idx)
+{
+ return (fm->bits[idx / MAP_T_BITS] & (MAP_1 << (idx % MAP_T_BITS))) != 0;
+}
+
+/* Returns 'true' if any of the 'n_bits' bits starting at 'idx' are set in
+ * 'fm'. 'n_bits' can be at most MAP_T_BITS. */
+static inline bool
+flowmap_are_set(const struct flowmap *fm, size_t idx, unsigned int n_bits)
+{
+ map_t n_bits_mask = (MAP_1 << n_bits) - 1;
+ size_t unit = idx / MAP_T_BITS;
+
+ idx %= MAP_T_BITS;
-static inline uint32_t *miniflow_values(struct miniflow *mf)
+ if (fm->bits[unit] & (n_bits_mask << idx)) {
+ return true;
+ }
+ /* The seemingly unnecessary bounds check on 'unit' is a workaround for a
+ * false-positive array out of bounds error by GCC 4.9. */
+ if (unit + 1 < FLOWMAP_UNITS && idx + n_bits > MAP_T_BITS) {
+ /* Check the remaining bits from the next unit. */
+ return fm->bits[unit + 1] & (n_bits_mask >> (MAP_T_BITS - idx));
+ }
+ return false;
+}
+
+/* Set the 'n_bits' consecutive bits in 'fm', starting at bit 'idx'.
+ * 'n_bits' can be at most MAP_T_BITS. */
+static inline void
+flowmap_set(struct flowmap *fm, size_t idx, unsigned int n_bits)
{
- return OVS_LIKELY(mf->values_inline)
- ? mf->inline_values : mf->offline_values;
+ map_t n_bits_mask = (MAP_1 << n_bits) - 1;
+ size_t unit = idx / MAP_T_BITS;
+
+ idx %= MAP_T_BITS;
+
+ fm->bits[unit] |= n_bits_mask << idx;
+ /* The seemingly unnecessary bounds check on 'unit' is a workaround for a
+ * false-positive array out of bounds error by GCC 4.9. */
+ if (unit + 1 < FLOWMAP_UNITS && idx + n_bits > MAP_T_BITS) {
+ /* 'MAP_T_BITS - idx' bits were set on 'unit', set the remaining
+ * bits from the next unit. */
+ fm->bits[unit + 1] |= n_bits_mask >> (MAP_T_BITS - idx);
+ }
}
-static inline const uint32_t *miniflow_get_values(const struct miniflow *mf)
+/* Clears the 'n_bits' consecutive bits in 'fm', starting at bit 'idx'.
+ * 'n_bits' can be at most MAP_T_BITS. */
+static inline void
+flowmap_clear(struct flowmap *fm, size_t idx, unsigned int n_bits)
{
- return OVS_LIKELY(mf->values_inline)
- ? mf->inline_values : mf->offline_values;
+ map_t n_bits_mask = (MAP_1 << n_bits) - 1;
+ size_t unit = idx / MAP_T_BITS;
+
+ idx %= MAP_T_BITS;
+
+ fm->bits[unit] &= ~(n_bits_mask << idx);
+ /* The seemingly unnecessary bounds check on 'unit' is a workaround for a
+ * false-positive array out of bounds error by GCC 4.9. */
+ if (unit + 1 < FLOWMAP_UNITS && idx + n_bits > MAP_T_BITS) {
+ /* 'MAP_T_BITS - idx' bits were cleared on 'unit', clear the
+ * remaining bits from the next unit. */
+ fm->bits[unit + 1] &= ~(n_bits_mask >> (MAP_T_BITS - idx));
+ }
+}
+
+/* OR the bits in the flowmaps. */
+static inline struct flowmap
+flowmap_or(struct flowmap a, struct flowmap b)
+{
+ struct flowmap map;
+ size_t unit;
+
+ FLOWMAP_FOR_EACH_UNIT (unit) {
+ map.bits[unit] = a.bits[unit] | b.bits[unit];
+ }
+ return map;
+}
+
+/* AND the bits in the flowmaps. */
+static inline struct flowmap
+flowmap_and(struct flowmap a, struct flowmap b)
+{
+ struct flowmap map;
+ size_t unit;
+
+ FLOWMAP_FOR_EACH_UNIT (unit) {
+ map.bits[unit] = a.bits[unit] & b.bits[unit];
+ }
+ return map;
}
-static inline const uint32_t *miniflow_get_u32_values(const struct miniflow *mf)
+static inline bool
+flowmap_is_empty(struct flowmap fm)
{
- return miniflow_get_values(mf);
+ map_t map;
+
+ FLOWMAP_FOR_EACH_MAP (map, fm) {
+ if (map) {
+ return false;
+ }
+ }
+ return true;
}
-static inline const ovs_be32 *miniflow_get_be32_values(const struct miniflow *mf)
+static inline unsigned int
+flowmap_n_1bits(struct flowmap fm)
{
- return (OVS_FORCE const ovs_be32 *)miniflow_get_values(mf);
+ unsigned int n_1bits = 0;
+ size_t unit;
+
+ FLOWMAP_FOR_EACH_UNIT (unit) {
+ n_1bits += count_1bits(fm.bits[unit]);
+ }
+ return n_1bits;
}
-/* This is useful for initializing a miniflow for a miniflow_extract() call. */
-static inline void miniflow_initialize(struct miniflow *mf,
- uint32_t buf[FLOW_U32S])
+struct flowmap_aux {
+ size_t unit;
+ struct flowmap map;
+};
+
+static inline bool
+flowmap_next_index(struct flowmap_aux *aux, size_t *idx)
{
- mf->map = 0;
- mf->values_inline = (buf == (uint32_t *)(mf + 1));
- if (!mf->values_inline) {
- mf->offline_values = buf;
+ for (;;) {
+ map_t *map = &aux->map.bits[aux->unit];
+ if (*map) {
+ *idx = aux->unit * MAP_T_BITS + raw_ctz(*map);
+ *map = zero_rightmost_1bit(*map);
+ return true;
+ }
+ if (++aux->unit >= FLOWMAP_UNITS) {
+ return false;
+ }
}
}
+\f
+/* Compressed flow. */
+
+/* A sparse representation of a "struct flow".
+ *
+ * A "struct flow" is fairly large and tends to be mostly zeros. Sparse
+ * representation has two advantages. First, it saves memory and, more
+ * importantly, minimizes the number of accessed cache lines. Second, it saves
+ * time when the goal is to iterate over only the nonzero parts of the struct.
+ *
+ * The map member hold one bit for each uint64_t in a "struct flow". Each
+ * 0-bit indicates that the corresponding uint64_t is zero, each 1-bit that it
+ * *may* be nonzero (see below how this applies to minimasks).
+ *
+ * The values indicated by 'map' always follow the miniflow in memory. The
+ * user of the miniflow is responsible for always having enough storage after
+ * the struct miniflow corresponding to the number of 1-bits in maps.
+ *
+ * Elements in values array are allowed to be zero. This is useful for "struct
+ * minimatch", for which ensuring that the miniflow and minimask members have
+ * same maps allows optimization. This allowance applies only to a miniflow
+ * that is not a mask. That is, a minimask may NOT have zero elements in its
+ * values.
+ *
+ * A miniflow is always dynamically allocated so that the maps are followed by
+ * at least as many elements as there are 1-bits in maps. */
+struct miniflow {
+ struct flowmap map;
+ /* Followed by:
+ * uint64_t values[n];
+ * where 'n' is miniflow_n_values(miniflow). */
+};
+BUILD_ASSERT_DECL(sizeof(struct miniflow) % sizeof(uint64_t) == 0);
+
+#define MINIFLOW_VALUES_SIZE(COUNT) ((COUNT) * sizeof(uint64_t))
+
+static inline uint64_t *miniflow_values(struct miniflow *mf)
+{
+ return (uint64_t *)(mf + 1);
+}
+
+static inline const uint64_t *miniflow_get_values(const struct miniflow *mf)
+{
+ return (const uint64_t *)(mf + 1);
+}
+
struct pkt_metadata;
-/* The 'dst->values' must be initialized with a buffer with space for
- * FLOW_U32S. 'dst->map' is ignored on input and set on output to
- * indicate which fields were extracted. */
-void miniflow_extract(struct ofpbuf *packet, const struct pkt_metadata *,
- struct miniflow *dst);
+/* The 'dst' must follow with buffer space for FLOW_U64S 64-bit units.
+ * 'dst->map' is ignored on input and set on output to indicate which fields
+ * were extracted. */
+void miniflow_extract(struct dp_packet *packet, struct miniflow *dst);
+void miniflow_map_init(struct miniflow *, const struct flow *);
+void flow_wc_map(const struct flow *, struct flowmap *);
+size_t miniflow_alloc(struct miniflow *dsts[], size_t n,
+ const struct miniflow *src);
void miniflow_init(struct miniflow *, const struct flow *);
-void miniflow_init_with_minimask(struct miniflow *, const struct flow *,
- const struct minimask *);
-void miniflow_clone(struct miniflow *, const struct miniflow *);
-void miniflow_clone_inline(struct miniflow *, const struct miniflow *,
- size_t n_values);
-void miniflow_move(struct miniflow *dst, struct miniflow *);
-void miniflow_destroy(struct miniflow *);
+void miniflow_clone(struct miniflow *, const struct miniflow *,
+ size_t n_values);
+struct miniflow * miniflow_create(const struct flow *);
void miniflow_expand(const struct miniflow *, struct flow *);
-static inline uint32_t flow_u32_value(const struct flow *flow, size_t index)
+static inline uint64_t flow_u64_value(const struct flow *flow, size_t index)
+{
+ return ((uint64_t *)flow)[index];
+}
+
+static inline uint64_t *flow_u64_lvalue(struct flow *flow, size_t index)
{
- return ((uint32_t *)(flow))[index];
+ return &((uint64_t *)flow)[index];
}
-static inline uint32_t *flow_u32_lvalue(struct flow *flow, size_t index)
+static inline size_t
+miniflow_n_values(const struct miniflow *flow)
{
- return &((uint32_t *)(flow))[index];
+ return flowmap_n_1bits(flow->map);
}
+struct flow_for_each_in_maps_aux {
+ const struct flow *flow;
+ struct flowmap_aux map_aux;
+};
+
static inline bool
-flow_get_next_in_map(const struct flow *flow, uint64_t map, uint32_t *value)
+flow_values_get_next_in_maps(struct flow_for_each_in_maps_aux *aux,
+ uint64_t *value)
{
- if (map) {
- *value = flow_u32_value(flow, raw_ctz(map));
+ size_t idx;
+
+ if (flowmap_next_index(&aux->map_aux, &idx)) {
+ *value = flow_u64_value(aux->flow, idx);
return true;
}
return false;
}
-/* Iterate through all flow u32 values specified by 'MAP'. */
-#define FLOW_FOR_EACH_IN_MAP(VALUE, FLOW, MAP) \
- for (uint64_t map__ = (MAP); \
- flow_get_next_in_map(FLOW, map__, &(VALUE)); \
- map__ = zero_rightmost_1bit(map__))
-
-/* Iterate through all struct flow u32 indices specified by 'MAP'. */
-#define MAP_FOR_EACH_INDEX(U32IDX, MAP) \
- for (uint64_t map__ = (MAP); \
- ((U32IDX) = ctz64(map__)) < FLOW_U32S; \
- map__ = zero_rightmost_1bit(map__))
-
-#define FLOW_U32_SIZE(FIELD) \
- DIV_ROUND_UP(sizeof(((struct flow *)0)->FIELD), sizeof(uint32_t))
-
-#define MINIFLOW_MAP(FIELD) \
- (((UINT64_C(1) << FLOW_U32_SIZE(FIELD)) - 1) \
- << (offsetof(struct flow, FIELD) / 4))
+/* Iterate through all flow u64 values specified by 'MAPS'. */
+#define FLOW_FOR_EACH_IN_MAPS(VALUE, FLOW, MAPS) \
+ for (struct flow_for_each_in_maps_aux aux__ \
+ = { (FLOW), FLOWMAP_AUX_INITIALIZER(MAPS) }; \
+ flow_values_get_next_in_maps(&aux__, &(VALUE));)
struct mf_for_each_in_map_aux {
- const uint32_t *values;
- uint64_t fmap;
- uint64_t map;
+ size_t unit;
+ struct flowmap fmap;
+ struct flowmap map;
+ const uint64_t *values;
};
static inline bool
-mf_get_next_in_map(struct mf_for_each_in_map_aux *aux, uint32_t *value)
-{
- if (aux->map) {
- uint64_t rm1bit = rightmost_1bit(aux->map);
- aux->map -= rm1bit;
-
- if (aux->fmap & rm1bit) {
- /* Advance 'aux->values' to point to the value for 'rm1bit'. */
- uint64_t trash = aux->fmap & (rm1bit - 1);
- if (trash) {
- aux->fmap -= trash;
- aux->values += count_1bits(trash);
- }
-
- /* Retrieve the value for 'rm1bit' then advance past it. */
- aux->fmap -= rm1bit;
- *value = *aux->values++;
- } else {
- *value = 0;
+mf_get_next_in_map(struct mf_for_each_in_map_aux *aux,
+ uint64_t *value)
+{
+ map_t *map, *fmap;
+ map_t rm1bit;
+
+ while (OVS_UNLIKELY(!*(map = &aux->map.bits[aux->unit]))) {
+ /* Skip remaining data in the previous unit. */
+ aux->values += count_1bits(aux->fmap.bits[aux->unit]);
+ if (++aux->unit == FLOWMAP_UNITS) {
+ return false;
}
- return true;
+ }
+
+ rm1bit = rightmost_1bit(*map);
+ *map -= rm1bit;
+ fmap = &aux->fmap.bits[aux->unit];
+
+ if (OVS_LIKELY(*fmap & rm1bit)) {
+ map_t trash = *fmap & (rm1bit - 1);
+
+ *fmap -= trash;
+ /* count_1bits() is fast for systems where speed matters (e.g.,
+ * DPDK), so we don't try avoid using it.
+ * Advance 'aux->values' to point to the value for 'rm1bit'. */
+ aux->values += count_1bits(trash);
+
+ *value = *aux->values;
} else {
- return false;
+ *value = 0;
}
+ return true;
}
-/* Iterate through all miniflow u32 values specified by 'MAP'. */
-#define MINIFLOW_FOR_EACH_IN_MAP(VALUE, FLOW, MAP) \
- for (struct mf_for_each_in_map_aux aux__ \
- = { miniflow_get_u32_values(FLOW), (FLOW)->map, MAP }; \
- mf_get_next_in_map(&aux__, &(VALUE)); \
- )
-
-/* Get the value of 'FIELD' of an up to 4 byte wide integer type 'TYPE' of
- * a miniflow. */
-#define MINIFLOW_GET_TYPE(MF, TYPE, OFS) \
- (((MF)->map & (UINT64_C(1) << (OFS) / 4)) \
- ? ((OVS_FORCE const TYPE *) \
- (miniflow_get_u32_values(MF) \
- + count_1bits((MF)->map & ((UINT64_C(1) << (OFS) / 4) - 1)))) \
- [(OFS) % 4 / sizeof(TYPE)] \
- : 0) \
-
-#define MINIFLOW_GET_U8(FLOW, FIELD) \
- MINIFLOW_GET_TYPE(FLOW, uint8_t, offsetof(struct flow, FIELD))
-#define MINIFLOW_GET_U16(FLOW, FIELD) \
- MINIFLOW_GET_TYPE(FLOW, uint16_t, offsetof(struct flow, FIELD))
-#define MINIFLOW_GET_BE16(FLOW, FIELD) \
- MINIFLOW_GET_TYPE(FLOW, ovs_be16, offsetof(struct flow, FIELD))
-#define MINIFLOW_GET_U32(FLOW, FIELD) \
- MINIFLOW_GET_TYPE(FLOW, uint32_t, offsetof(struct flow, FIELD))
-#define MINIFLOW_GET_BE32(FLOW, FIELD) \
- MINIFLOW_GET_TYPE(FLOW, ovs_be32, offsetof(struct flow, FIELD))
+/* Iterate through miniflow u64 values specified by 'FLOWMAP'. */
+#define MINIFLOW_FOR_EACH_IN_FLOWMAP(VALUE, FLOW, FLOWMAP) \
+ for (struct mf_for_each_in_map_aux aux__ = \
+ { 0, (FLOW)->map, (FLOWMAP), miniflow_get_values(FLOW) }; \
+ mf_get_next_in_map(&aux__, &(VALUE));)
+/* This can be used when it is known that 'idx' is set in 'map'. */
+static inline const uint64_t *
+miniflow_values_get__(const uint64_t *values, map_t map, size_t idx)
+{
+ return values + count_1bits(map & ((MAP_1 << idx) - 1));
+}
+
+/* This can be used when it is known that 'u64_idx' is set in
+ * the map of 'mf'. */
+static inline const uint64_t *
+miniflow_get__(const struct miniflow *mf, size_t idx)
+{
+ const uint64_t *values = miniflow_get_values(mf);
+ const map_t *map = mf->map.bits;
+
+ while (idx >= MAP_T_BITS) {
+ idx -= MAP_T_BITS;
+ values += count_1bits(*map++);
+ }
+ return miniflow_values_get__(values, *map, idx);
+}
+
+#define MINIFLOW_IN_MAP(MF, IDX) flowmap_is_set(&(MF)->map, IDX)
+
+/* Get the value of the struct flow 'FIELD' as up to 8 byte wide integer type
+ * 'TYPE' from miniflow 'MF'. */
+#define MINIFLOW_GET_TYPE(MF, TYPE, FIELD) \
+ (MINIFLOW_IN_MAP(MF, FLOW_U64_OFFSET(FIELD)) \
+ ? ((OVS_FORCE const TYPE *)miniflow_get__(MF, FLOW_U64_OFFSET(FIELD))) \
+ [FLOW_U64_OFFREM(FIELD) / sizeof(TYPE)] \
+ : 0)
+
+#define MINIFLOW_GET_U128(FLOW, FIELD) \
+ (ovs_u128) { .u64 = { \
+ (MINIFLOW_IN_MAP(FLOW, FLOW_U64_OFFSET(FIELD)) ? \
+ *miniflow_get__(FLOW, FLOW_U64_OFFSET(FIELD)) : 0), \
+ (MINIFLOW_IN_MAP(FLOW, FLOW_U64_OFFSET(FIELD) + 1) ? \
+ *miniflow_get__(FLOW, FLOW_U64_OFFSET(FIELD) + 1) : 0) } }
+
+#define MINIFLOW_GET_U8(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, uint8_t, FIELD)
+#define MINIFLOW_GET_U16(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, uint16_t, FIELD)
+#define MINIFLOW_GET_BE16(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, ovs_be16, FIELD)
+#define MINIFLOW_GET_U32(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, uint32_t, FIELD)
+#define MINIFLOW_GET_BE32(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, ovs_be32, FIELD)
+#define MINIFLOW_GET_U64(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, uint64_t, FIELD)
+#define MINIFLOW_GET_BE64(FLOW, FIELD) \
+ MINIFLOW_GET_TYPE(FLOW, ovs_be64, FIELD)
+
+static inline uint64_t miniflow_get(const struct miniflow *,
+ unsigned int u64_ofs);
+static inline uint32_t miniflow_get_u32(const struct miniflow *,
+ unsigned int u32_ofs);
+static inline ovs_be32 miniflow_get_be32(const struct miniflow *,
+ unsigned int be32_ofs);
static inline uint16_t miniflow_get_vid(const struct miniflow *);
static inline uint16_t miniflow_get_tcp_flags(const struct miniflow *);
static inline ovs_be64 miniflow_get_metadata(const struct miniflow *);
};
void minimask_init(struct minimask *, const struct flow_wildcards *);
-void minimask_clone(struct minimask *, const struct minimask *);
-void minimask_move(struct minimask *dst, struct minimask *src);
+struct minimask * minimask_create(const struct flow_wildcards *);
void minimask_combine(struct minimask *dst,
const struct minimask *a, const struct minimask *b,
- uint32_t storage[FLOW_U32S]);
-void minimask_destroy(struct minimask *);
+ uint64_t storage[FLOW_U64S]);
void minimask_expand(const struct minimask *, struct flow_wildcards *);
-uint32_t minimask_get(const struct minimask *, unsigned int u32_ofs);
+static inline uint32_t minimask_get_u32(const struct minimask *,
+ unsigned int u32_ofs);
+static inline ovs_be32 minimask_get_be32(const struct minimask *,
+ unsigned int be32_ofs);
static inline uint16_t minimask_get_vid_mask(const struct minimask *);
static inline ovs_be64 minimask_get_metadata_mask(const struct minimask *);
/* For every 1-bit in mask's map, the corresponding value is non-zero,
* so the only way the mask can not fix any bits or fields is for the
* map the be zero. */
- return mask->masks.map == 0;
+ return flowmap_is_empty(mask->masks.map);
+}
+
+/* Returns the uint64_t that would be at byte offset '8 * u64_ofs' if 'flow'
+ * were expanded into a "struct flow". */
+static inline uint64_t miniflow_get(const struct miniflow *flow,
+ unsigned int u64_ofs)
+{
+ return MINIFLOW_IN_MAP(flow, u64_ofs) ? *miniflow_get__(flow, u64_ofs) : 0;
+}
+
+static inline uint32_t miniflow_get_u32(const struct miniflow *flow,
+ unsigned int u32_ofs)
+{
+ uint64_t value = miniflow_get(flow, u32_ofs / 2);
+
+#if WORDS_BIGENDIAN
+ return (u32_ofs & 1) ? value : value >> 32;
+#else
+ return (u32_ofs & 1) ? value >> 32 : value;
+#endif
+}
+
+static inline ovs_be32 miniflow_get_be32(const struct miniflow *flow,
+ unsigned int be32_ofs)
+{
+ return (OVS_FORCE ovs_be32)miniflow_get_u32(flow, be32_ofs);
}
/* Returns the VID within the vlan_tci member of the "struct flow" represented
return vlan_tci_to_vid(tci);
}
+/* Returns the uint32_t that would be at byte offset '4 * u32_ofs' if 'mask'
+ * were expanded into a "struct flow_wildcards". */
+static inline uint32_t
+minimask_get_u32(const struct minimask *mask, unsigned int u32_ofs)
+{
+ return miniflow_get_u32(&mask->masks, u32_ofs);
+}
+
+static inline ovs_be32
+minimask_get_be32(const struct minimask *mask, unsigned int be32_ofs)
+{
+ return (OVS_FORCE ovs_be32)minimask_get_u32(mask, be32_ofs);
+}
+
/* Returns the VID mask within the vlan_tci member of the "struct
* flow_wildcards" represented by 'mask'. */
static inline uint16_t
static inline ovs_be64
miniflow_get_metadata(const struct miniflow *flow)
{
- union {
- ovs_be64 be64;
- struct {
- ovs_be32 hi;
- ovs_be32 lo;
- };
- } value;
-
- enum { MD_OFS = offsetof(struct flow, metadata) };
- BUILD_ASSERT_DECL(MD_OFS % sizeof(uint32_t) == 0);
- value.hi = MINIFLOW_GET_TYPE(flow, ovs_be32, MD_OFS);
- value.lo = MINIFLOW_GET_TYPE(flow, ovs_be32, MD_OFS + 4);
-
- return value.be64;
+ return MINIFLOW_GET_BE64(flow, metadata);
}
/* Returns the mask for the OpenFlow 1.1+ "metadata" field in 'mask'.
static inline ovs_be64
minimask_get_metadata_mask(const struct minimask *mask)
{
- return miniflow_get_metadata(&mask->masks);
+ return MINIFLOW_GET_BE64(&mask->masks, metadata);
}
-/* Perform a bitwise OR of miniflow 'src' flow data with the equivalent
- * fields in 'dst', storing the result in 'dst'. */
+/* Perform a bitwise OR of miniflow 'src' flow data specified in 'subset' with
+ * the equivalent fields in 'dst', storing the result in 'dst'. 'subset' must
+ * be a subset of 'src's map. */
static inline void
-flow_union_with_miniflow(struct flow *dst, const struct miniflow *src)
+flow_union_with_miniflow_subset(struct flow *dst, const struct miniflow *src,
+ struct flowmap subset)
{
- uint32_t *dst_u32 = (uint32_t *) dst;
- const uint32_t *p = miniflow_get_u32_values(src);
- uint64_t map;
+ uint64_t *dst_u64 = (uint64_t *) dst;
+ const uint64_t *p = miniflow_get_values(src);
+ map_t map;
- for (map = src->map; map; map = zero_rightmost_1bit(map)) {
- dst_u32[raw_ctz(map)] |= *p++;
+ FLOWMAP_FOR_EACH_MAP (map, subset) {
+ size_t idx;
+
+ MAP_FOR_EACH_INDEX(idx, map) {
+ dst_u64[idx] |= *p++;
+ }
+ dst_u64 += MAP_T_BITS;
}
}
-static inline struct pkt_metadata
-pkt_metadata_from_flow(const struct flow *flow)
+/* Perform a bitwise OR of miniflow 'src' flow data with the equivalent
+ * fields in 'dst', storing the result in 'dst'. */
+static inline void
+flow_union_with_miniflow(struct flow *dst, const struct miniflow *src)
{
- struct pkt_metadata md;
-
- md.recirc_id = flow->recirc_id;
- md.dp_hash = flow->dp_hash;
- md.tunnel = flow->tunnel;
- md.skb_priority = flow->skb_priority;
- md.pkt_mark = flow->pkt_mark;
- md.in_port = flow->in_port;
+ flow_union_with_miniflow_subset(dst, src, src->map);
+}
- return md;
+static inline void
+pkt_metadata_from_flow(struct pkt_metadata *md, const struct flow *flow)
+{
+ md->recirc_id = flow->recirc_id;
+ md->dp_hash = flow->dp_hash;
+ flow_tnl_copy__(&md->tunnel, &flow->tunnel);
+ md->skb_priority = flow->skb_priority;
+ md->pkt_mark = flow->pkt_mark;
+ md->in_port = flow->in_port;
+ md->ct_state = flow->ct_state;
+ md->ct_zone = flow->ct_zone;
+ md->ct_mark = flow->ct_mark;
+ md->ct_label = flow->ct_label;
}
static inline bool is_ip_any(const struct flow *flow)
&& flow->nw_proto == IPPROTO_ICMPV6);
}
+static inline bool is_igmp(const struct flow *flow)
+{
+ return (flow->dl_type == htons(ETH_TYPE_IP)
+ && flow->nw_proto == IPPROTO_IGMP);
+}
+
+static inline bool is_mld(const struct flow *flow)
+{
+ return is_icmpv6(flow)
+ && (flow->tp_src == htons(MLD_QUERY)
+ || flow->tp_src == htons(MLD_REPORT)
+ || flow->tp_src == htons(MLD_DONE)
+ || flow->tp_src == htons(MLD2_REPORT));
+}
+
+static inline bool is_mld_query(const struct flow *flow)
+{
+ return is_icmpv6(flow) && flow->tp_src == htons(MLD_QUERY);
+}
+
+static inline bool is_mld_report(const struct flow *flow)
+{
+ return is_mld(flow) && !is_mld_query(flow);
+}
+
static inline bool is_stp(const struct flow *flow)
{
return (eth_addr_equals(flow->dl_dst, eth_addr_stp)
projects / cascardo / ovs.git / blobdiff