/*
- * Copyright (c) 2008, 2009, 2010, 2011, 2012 Nicira, Inc.
+ * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014 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 "byte-order.h"
#include "openflow/nicira-ext.h"
#include "openflow/openflow.h"
+#include "packets.h"
#include "hash.h"
#include "util.h"
struct dpif_flow_stats;
struct ds;
struct flow_wildcards;
+struct minimask;
struct ofpbuf;
+struct pkt_metadata;
/* 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 14
+#define FLOW_WC_SEQ 26
#define FLOW_N_REGS 8
BUILD_ASSERT_DECL(FLOW_N_REGS <= NXM_NX_MAX_REGS);
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)
+
+const char *flow_tun_flag_to_string(uint32_t flags);
+
+/* Maximum number of supported MPLS labels. */
+#define FLOW_MAX_MPLS_LABELS 3
+
+/*
+ * A flow in the network.
+ *
+ * Must be initialized to all zeros to make any compiler-induced padding
+ * zeroed. Helps also in keeping unused fields (such as mutually exclusive
+ * IPv4 and IPv6 addresses) zeroed out.
+ *
+ * The meaning of 'in_port' is context-dependent. In most cases, it is a
+ * 16-bit OpenFlow 1.0 port number. In the software datapath interface (dpif)
+ * layer and its implementations (e.g. dpif-linux, dpif-netdev), it is instead
+ * a 32-bit datapath port number.
+ *
+ * The fields are organized in four segments to facilitate staged lookup, where
+ * lower layer fields are first used to determine if the later fields need to
+ * be looked at. This enables better wildcarding for datapath flows.
+ *
+ * NOTE: Order of the fields is significant, any change in the order must be
+ * reflected in miniflow_extract()!
+ */
struct flow {
- ovs_be64 tun_id; /* Encapsulating tunnel ID. */
+ /* L1 */
+ struct flow_tnl tunnel; /* Encapsulating tunnel parameters. */
ovs_be64 metadata; /* OpenFlow Metadata. */
+ 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. */
+ union flow_in_port in_port; /* Input port.*/
+
+ /* L2, Order the same as in the Ethernet header! */
+ uint8_t dl_dst[6]; /* Ethernet destination address. */
+ uint8_t dl_src[6]; /* 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 */
struct in6_addr ipv6_src; /* IPv6 source address. */
struct in6_addr ipv6_dst; /* IPv6 destination address. */
- struct in6_addr nd_target; /* IPv6 neighbor discovery (ND) target. */
- uint32_t skb_priority; /* Packet priority for QoS. */
- uint32_t regs[FLOW_N_REGS]; /* Registers. */
+ ovs_be32 ipv6_label; /* IPv6 flow label. */
ovs_be32 nw_src; /* IPv4 source address. */
ovs_be32 nw_dst; /* IPv4 destination address. */
- ovs_be32 ipv6_label; /* IPv6 flow label. */
- uint16_t in_port; /* OpenFlow port number of input port. */
- ovs_be16 vlan_tci; /* If 802.1Q, TCI | VLAN_CFI; otherwise 0. */
- ovs_be16 dl_type; /* Ethernet frame type. */
- ovs_be16 tp_src; /* TCP/UDP source port. */
- ovs_be16 tp_dst; /* TCP/UDP destination port. */
- uint8_t dl_src[6]; /* Ethernet source address. */
- uint8_t dl_dst[6]; /* Ethernet destination address. */
- uint8_t nw_proto; /* IP protocol or low 8 bits of ARP opcode. */
+ 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[6]; /* ARP/ND source hardware address. */
uint8_t arp_tha[6]; /* ARP/ND target hardware address. */
- uint8_t nw_ttl; /* IP TTL/Hop Limit. */
- uint8_t nw_frag; /* FLOW_FRAG_* flags. */
- uint8_t reserved[2]; /* Reserved for 64-bit packing. */
+ struct in6_addr nd_target; /* IPv6 neighbor discovery (ND) target. */
+ ovs_be16 tcp_flags; /* TCP flags. With L3 to avoid matching L4. */
+ ovs_be16 pad; /* Padding. */
+
+ /* L4 */
+ ovs_be16 tp_src; /* TCP/UDP/SCTP source port. */
+ ovs_be16 tp_dst; /* TCP/UDP/SCTP destination port.
+ * Keep last for the BUILD_ASSERT_DECL below */
+ uint32_t dp_hash; /* Datapath computed hash value. The exact
+ computation is opaque to the user space.*/
};
+BUILD_ASSERT_DECL(sizeof(struct flow) % 4 == 0);
-/* Represents the metadata fields of struct flow. The masks are used to
- * indicate which metadata fields are relevant in a given context. Typically
- * they will be all 1 or all 0. */
-struct flow_metadata {
- ovs_be64 tun_id; /* Encapsulating tunnel ID. */
- ovs_be64 tun_id_mask; /* 1-bit in each significant tun_id bit.*/
+#define FLOW_U32S (sizeof(struct flow) / 4)
+
+/* 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) + 172
+ && FLOW_WC_SEQ == 26);
+
+/* 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
+ * 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_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( 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));
- ovs_be64 metadata;
- ovs_be64 metadata_mask;
+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 reg_masks[FLOW_N_REGS]; /* 1-bit in each significant regs bit. */
-
- uint16_t in_port; /* OpenFlow port or zero. */
+ uint32_t pkt_mark; /* Packet mark. */
+ ofp_port_t in_port; /* OpenFlow port or zero. */
};
-/* Assert that there are FLOW_SIG_SIZE bytes of significant data in "struct
- * flow", followed by FLOW_PAD_SIZE bytes of padding. */
-#define FLOW_SIG_SIZE (118 + FLOW_N_REGS * 4)
-#define FLOW_PAD_SIZE 2
-BUILD_ASSERT_DECL(offsetof(struct flow, nw_frag) == FLOW_SIG_SIZE - 1);
-BUILD_ASSERT_DECL(sizeof(((struct flow *)0)->nw_frag) == 1);
-BUILD_ASSERT_DECL(sizeof(struct flow) == FLOW_SIG_SIZE + FLOW_PAD_SIZE);
+void flow_extract(struct ofpbuf *, const struct pkt_metadata *md,
+ struct flow *);
-/* Remember to update FLOW_WC_SEQ when changing 'struct flow'. */
-BUILD_ASSERT_DECL(FLOW_SIG_SIZE == 150 && FLOW_WC_SEQ == 14);
-
-void flow_extract(struct ofpbuf *, uint32_t priority, ovs_be64 tun_id,
- uint16_t in_port, 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 *);
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);
+
void flow_format(struct ds *, const struct flow *);
void flow_print(FILE *, const struct flow *);
static inline int flow_compare_3way(const struct flow *, const struct flow *);
void flow_set_vlan_vid(struct flow *, ovs_be16 vid);
void flow_set_vlan_pcp(struct flow *, uint8_t pcp);
+int flow_count_mpls_labels(const struct flow *, struct flow_wildcards *);
+int flow_count_common_mpls_labels(const struct flow *a, int an,
+ const struct flow *b, int bn,
+ struct flow_wildcards *wc);
+void flow_push_mpls(struct flow *, int n, ovs_be16 mpls_eth_type,
+ struct flow_wildcards *);
+bool flow_pop_mpls(struct flow *, int n, ovs_be16 eth_type,
+ struct flow_wildcards *);
+void flow_set_mpls_label(struct flow *, int idx, ovs_be32 label);
+void flow_set_mpls_ttl(struct flow *, int idx, uint8_t ttl);
+void flow_set_mpls_tc(struct flow *, int idx, uint8_t tc);
+void flow_set_mpls_bos(struct flow *, int idx, uint8_t stack);
+void flow_set_mpls_lse(struct flow *, int idx, ovs_be32 lse);
+
void flow_compose(struct ofpbuf *, const struct flow *);
static inline int
flow_compare_3way(const struct flow *a, const struct flow *b)
{
- return memcmp(a, b, FLOW_SIG_SIZE);
+ return memcmp(a, b, sizeof *a);
}
static inline bool
static inline size_t
flow_hash(const struct flow *flow, uint32_t basis)
{
- return hash_bytes(flow, FLOW_SIG_SIZE, basis);
+ return hash_words((const uint32_t *) flow, sizeof *flow / 4, basis);
}
-/* Open vSwitch flow wildcard bits.
- *
- * These are used only internally to Open vSwitch, in the 'wildcards' member of
- * struct flow_wildcards. They never appear in the wire protocol in this
- * form. */
+static inline uint16_t
+ofp_to_u16(ofp_port_t ofp_port)
+{
+ return (OVS_FORCE uint16_t) ofp_port;
+}
+
+static inline uint32_t
+odp_to_u32(odp_port_t odp_port)
+{
+ return (OVS_FORCE uint32_t) odp_port;
+}
+
+static inline uint32_t
+ofp11_to_u32(ofp11_port_t ofp11_port)
+{
+ return (OVS_FORCE uint32_t) ofp11_port;
+}
-typedef unsigned int OVS_BITWISE flow_wildcards_t;
+static inline ofp_port_t
+u16_to_ofp(uint16_t port)
+{
+ return OFP_PORT_C(port);
+}
-/* Same values and meanings as corresponding OFPFW10_* bits. */
-#define FWW_IN_PORT ((OVS_FORCE flow_wildcards_t) (1 << 0))
-#define FWW_DL_TYPE ((OVS_FORCE flow_wildcards_t) (1 << 4))
-#define FWW_NW_PROTO ((OVS_FORCE flow_wildcards_t) (1 << 5))
-/* No corresponding OFPFW10_* bits. */
-#define FWW_NW_DSCP ((OVS_FORCE flow_wildcards_t) (1 << 1))
-#define FWW_NW_ECN ((OVS_FORCE flow_wildcards_t) (1 << 2))
-#define FWW_NW_TTL ((OVS_FORCE flow_wildcards_t) (1 << 3))
-#define FWW_ALL ((OVS_FORCE flow_wildcards_t) (((1 << 6)) - 1))
+static inline odp_port_t
+u32_to_odp(uint32_t port)
+{
+ return ODP_PORT_C(port);
+}
-/* Remember to update FLOW_WC_SEQ when adding or removing FWW_*. */
-BUILD_ASSERT_DECL(FWW_ALL == ((1 << 6) - 1) && FLOW_WC_SEQ == 14);
+static inline ofp11_port_t
+u32_to_ofp11(uint32_t port)
+{
+ return OFP11_PORT_C(port);
+}
+
+static inline uint32_t
+hash_ofp_port(ofp_port_t ofp_port)
+{
+ return hash_int(ofp_to_u16(ofp_port), 0);
+}
-/* Information on wildcards for a flow, as a supplement to "struct flow".
+static inline uint32_t
+hash_odp_port(odp_port_t odp_port)
+{
+ return hash_int(odp_to_u32(odp_port), 0);
+}
+\f
+/* Wildcards for a flow.
*
- * Note that the meaning of 1-bits in 'wildcards' is opposite that of 1-bits in
- * the rest of the members. */
+ * A 1-bit in each bit in 'masks' indicates that the corresponding bit of
+ * the flow is significant (must match). A 0-bit indicates that the
+ * corresponding bit of the flow is wildcarded (need not match). */
struct flow_wildcards {
- ovs_be64 tun_id_mask; /* 1-bit in each significant tun_id bit. */
- ovs_be64 metadata_mask; /* 1-bit in each significant metadata bit. */
- flow_wildcards_t wildcards; /* 1-bit in each FWW_* wildcarded field. */
- uint32_t reg_masks[FLOW_N_REGS]; /* 1-bit in each significant regs bit. */
- ovs_be32 nw_src_mask; /* 1-bit in each significant nw_src bit. */
- ovs_be32 nw_dst_mask; /* 1-bit in each significant nw_dst bit. */
- struct in6_addr ipv6_src_mask; /* 1-bit in each signficant ipv6_src bit. */
- struct in6_addr ipv6_dst_mask; /* 1-bit in each signficant ipv6_dst bit. */
- struct in6_addr nd_target_mask; /* 1-bit in each significant
- nd_target bit. */
- ovs_be32 ipv6_label_mask; /* 1 bit in each significant ipv6_label bit. */
- ovs_be16 vlan_tci_mask; /* 1-bit in each significant vlan_tci bit. */
- ovs_be16 tp_src_mask; /* 1-bit in each significant tp_src bit. */
- ovs_be16 tp_dst_mask; /* 1-bit in each significant tp_dst bit. */
- uint8_t nw_frag_mask; /* 1-bit in each significant nw_frag bit. */
- uint8_t dl_src_mask[6]; /* 1-bit in each significant dl_src bit. */
- uint8_t dl_dst_mask[6]; /* 1-bit in each significant dl_dst bit. */
- uint8_t arp_sha_mask[6]; /* 1-bit in each significant dl_dst bit. */
- uint8_t arp_tha_mask[6]; /* 1-bit in each significant dl_dst bit. */
- uint8_t zeros[1]; /* Padding field set to zero. */
+ struct flow masks;
};
-/* Remember to update FLOW_WC_SEQ when updating struct flow_wildcards. */
-BUILD_ASSERT_DECL(sizeof(struct flow_wildcards) == 144 && FLOW_WC_SEQ == 14);
-
void flow_wildcards_init_catchall(struct flow_wildcards *);
-void flow_wildcards_init_exact(struct flow_wildcards *);
-bool flow_wildcards_is_exact(const struct flow_wildcards *);
+void flow_wildcards_clear_non_packet_fields(struct flow_wildcards *);
+
bool flow_wildcards_is_catchall(const struct flow_wildcards *);
void flow_wildcards_set_reg_mask(struct flow_wildcards *,
int idx, uint32_t mask);
-void flow_wildcards_combine(struct flow_wildcards *dst,
- const struct flow_wildcards *src1,
- const struct flow_wildcards *src2);
+void flow_wildcards_and(struct flow_wildcards *dst,
+ const struct flow_wildcards *src1,
+ const struct flow_wildcards *src2);
+void flow_wildcards_or(struct flow_wildcards *dst,
+ const struct flow_wildcards *src1,
+ const struct flow_wildcards *src2);
bool flow_wildcards_has_extra(const struct flow_wildcards *,
const struct flow_wildcards *);
-
uint32_t flow_wildcards_hash(const struct flow_wildcards *, uint32_t basis);
bool flow_wildcards_equal(const 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);
-const uint8_t *flow_wildcards_to_dl_dst_mask(flow_wildcards_t);
-bool flow_wildcards_is_dl_dst_mask_valid(const uint8_t[6]);
-flow_wildcards_t flow_wildcards_set_dl_dst_mask(flow_wildcards_t,
- const uint8_t mask[6]);
+/* Initialize a flow with random fields that matter for nx_hash_fields. */
+void flow_random_hash_fields(struct flow *);
+void flow_mask_hash_fields(const struct flow *, struct flow_wildcards *,
+ enum nx_hash_fields);
uint32_t flow_hash_fields(const struct flow *, enum nx_hash_fields,
uint16_t basis);
const char *flow_hash_fields_to_str(enum nx_hash_fields);
bool flow_hash_fields_valid(enum nx_hash_fields);
+uint32_t flow_hash_in_wildcards(const struct flow *,
+ const struct flow_wildcards *,
+ uint32_t basis);
+
+bool flow_equal_except(const struct flow *a, const struct flow *b,
+ const struct flow_wildcards *);
+\f
+/* Compressed flow. */
+
+#define MINI_N_INLINE (sizeof(void *) == 4 ? 7 : 8)
+BUILD_ASSERT_DECL(FLOW_U32S <= 64);
+
+/* 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.
+ *
+ * 'values' points to the start of an array that has one element for each 1-bit
+ * in 'map'. The least-numbered 1-bit is in values[0], the next 1-bit is in
+ * values[1], and so on.
+ *
+ * 'values' may point to a few different locations:
+ *
+ * - If 'map' has MINI_N_INLINE or fewer 1-bits, it may point to
+ * 'inline_values'. One hopes that this is the common case.
+ *
+ * - If 'map' has more than MINI_N_INLINE 1-bits, it may point to memory
+ * allocated with malloc().
+ *
+ * - The caller could provide storage on the stack for situations where
+ * that makes sense. So far that's only proved useful for
+ * minimask_combine(), but the principle works elsewhere.
+ *
+ * Elements in 'values' 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;
+ uint32_t *values;
+ uint32_t inline_values[MINI_N_INLINE];
+};
+
+/* 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])
+{
+ mf->map = 0;
+ mf->values = buf;
+}
+
+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);
+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_move(struct miniflow *dst, struct miniflow *);
+void miniflow_destroy(struct miniflow *);
+
+void miniflow_expand(const struct miniflow *, struct flow *);
+
+#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))
+
+static inline uint32_t
+mf_get_next_in_map(uint64_t *fmap, uint64_t rm1bit, const uint32_t **fp,
+ uint32_t *value)
+{
+ *value = 0;
+ if (*fmap & rm1bit) {
+ uint64_t trash = *fmap & (rm1bit - 1);
+
+ if (trash) {
+ *fmap -= trash;
+ *fp += count_1bits(trash);
+ }
+ *value = **fp;
+ }
+ return rm1bit != 0;
+}
+
+/* Iterate through all miniflow u32 values specified by the 'MAP'.
+ * This works as the first statement in a block.*/
+#define MINIFLOW_FOR_EACH_IN_MAP(VALUE, FLOW, MAP) \
+ const uint32_t *fp_ = (FLOW)->values; \
+ uint64_t rm1bit_, fmap_, map_; \
+ for (fmap_ = (FLOW)->map, map_ = (MAP), rm1bit_ = rightmost_1bit(map_); \
+ mf_get_next_in_map(&fmap_, rm1bit_, &fp_, &(VALUE)); \
+ map_ -= rm1bit_, rm1bit_ = rightmost_1bit(map_))
+
+/* 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 *) \
+ ((MF)->values \
+ + 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))
+
+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 *);
+
+bool miniflow_equal(const struct miniflow *a, const struct miniflow *b);
+bool miniflow_equal_in_minimask(const struct miniflow *a,
+ const struct miniflow *b,
+ const struct minimask *);
+bool miniflow_equal_flow_in_minimask(const struct miniflow *a,
+ const struct flow *b,
+ const struct minimask *);
+uint32_t miniflow_hash_5tuple(const struct miniflow *flow, uint32_t basis);
+
+\f
+/* Compressed flow wildcards. */
+
+/* A sparse representation of a "struct flow_wildcards".
+ *
+ * See the large comment on struct miniflow for details.
+ *
+ * Note: While miniflow can have zero data for a 1-bit in the map,
+ * a minimask may not! We rely on this in the implementation. */
+struct minimask {
+ struct miniflow masks;
+};
+
+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);
+void minimask_combine(struct minimask *dst,
+ const struct minimask *a, const struct minimask *b,
+ uint32_t storage[FLOW_U32S]);
+void minimask_destroy(struct minimask *);
+
+void minimask_expand(const struct minimask *, struct flow_wildcards *);
+
+uint32_t minimask_get(const struct minimask *, unsigned int u32_ofs);
+static inline uint16_t minimask_get_vid_mask(const struct minimask *);
+static inline ovs_be64 minimask_get_metadata_mask(const struct minimask *);
+
+bool minimask_equal(const struct minimask *a, const struct minimask *b);
+bool minimask_has_extra(const struct minimask *, const struct minimask *);
+
+/* Returns true if 'mask' matches every packet, false if 'mask' fixes any bits
+ * or fields. */
+static inline bool
+minimask_is_catchall(const struct minimask *mask)
+{
+ /* For every 1-bit in mask's map, the corresponding value is non-zero,
+ * so the only way the mask can not fix any bits or fields is for the
+ * map the be zero. */
+ return mask->masks.map == 0;
+}
+
+\f
+
+/* Returns the VID within the vlan_tci member of the "struct flow" represented
+ * by 'flow'. */
+static inline uint16_t
+miniflow_get_vid(const struct miniflow *flow)
+{
+ ovs_be16 tci = MINIFLOW_GET_BE16(flow, vlan_tci);
+ return vlan_tci_to_vid(tci);
+}
+
+/* Returns the VID mask within the vlan_tci member of the "struct
+ * flow_wildcards" represented by 'mask'. */
+static inline uint16_t
+minimask_get_vid_mask(const struct minimask *mask)
+{
+ return miniflow_get_vid(&mask->masks);
+}
+
+/* Returns the value of the "tcp_flags" field in 'flow'. */
+static inline uint16_t
+miniflow_get_tcp_flags(const struct miniflow *flow)
+{
+ return ntohs(MINIFLOW_GET_BE16(flow, tcp_flags));
+}
+
+/* Returns the value of the OpenFlow 1.1+ "metadata" field in 'flow'. */
+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;
+}
+
+/* Returns the mask for the OpenFlow 1.1+ "metadata" field in 'mask'.
+ *
+ * The return value is all-1-bits if 'mask' matches on the whole value of the
+ * metadata field, all-0-bits if 'mask' entirely wildcards the metadata field,
+ * or some other value if the metadata field is partially matched, partially
+ * wildcarded. */
+static inline ovs_be64
+minimask_get_metadata_mask(const struct minimask *mask)
+{
+ return miniflow_get_metadata(&mask->masks);
+}
+
+/* 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)
+{
+ uint32_t *dst_u32 = (uint32_t *) dst;
+ const uint32_t *p = (uint32_t *)src->values;
+ uint64_t map;
+
+ for (map = src->map; map; map = zero_rightmost_1bit(map)) {
+ dst_u32[raw_ctz(map)] |= *p++;
+ }
+}
+
+static inline struct pkt_metadata
+pkt_metadata_from_flow(const struct flow *flow)
+{
+ struct pkt_metadata md;
+
+ md.recirc_id = flow->recirc_id;
+ md.dp_hash = flow->dp_hash;
+ md.tunnel = flow->tunnel;
+ md.skb_priority = flow->skb_priority;
+ md.pkt_mark = flow->pkt_mark;
+ md.in_port = flow->in_port;
+
+ return md;
+}
+
+static inline bool is_ip_any(const struct flow *flow)
+{
+ return dl_type_is_ip_any(flow->dl_type);
+}
+
+static inline bool is_icmpv4(const struct flow *flow)
+{
+ return (flow->dl_type == htons(ETH_TYPE_IP)
+ && flow->nw_proto == IPPROTO_ICMP);
+}
+
+static inline bool is_icmpv6(const struct flow *flow)
+{
+ return (flow->dl_type == htons(ETH_TYPE_IPV6)
+ && flow->nw_proto == IPPROTO_ICMPV6);
+}
+
+static inline bool is_stp(const struct flow *flow)
+{
+ return (eth_addr_equals(flow->dl_dst, eth_addr_stp)
+ && flow->dl_type == htons(FLOW_DL_TYPE_NONE));
+}
+
#endif /* flow.h */
git://git.onelab.eu / sliver-openvswitch.git / blobdiff