/*
- * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014 Nicira, Inc.
+ * Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015 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 "hash.h"
#include "jhash.h"
#include "match.h"
-#include "ofpbuf.h"
+#include "dp-packet.h"
#include "openflow/openflow.h"
#include "packets.h"
#include "odp-util.h"
COVERAGE_DEFINE(flow_extract);
COVERAGE_DEFINE(miniflow_malloc);
-/* U32 indices for segmented flow classification. */
-const uint8_t flow_segment_u32s[4] = {
- FLOW_SEGMENT_1_ENDS_AT / 4,
- FLOW_SEGMENT_2_ENDS_AT / 4,
- FLOW_SEGMENT_3_ENDS_AT / 4,
- FLOW_U32S
+/* U64 indices for segmented flow classification. */
+const uint8_t flow_segment_u64s[4] = {
+ FLOW_SEGMENT_1_ENDS_AT / sizeof(uint64_t),
+ FLOW_SEGMENT_2_ENDS_AT / sizeof(uint64_t),
+ FLOW_SEGMENT_3_ENDS_AT / sizeof(uint64_t),
+ FLOW_U64S
};
+/* Asserts that field 'f1' follows immediately after 'f0' in struct flow,
+ * without any intervening padding. */
+#define ASSERT_SEQUENTIAL(f0, f1) \
+ BUILD_ASSERT_DECL(offsetof(struct flow, f0) \
+ + MEMBER_SIZEOF(struct flow, f0) \
+ == offsetof(struct flow, f1))
+
+/* Asserts that fields 'f0' and 'f1' are in the same 32-bit aligned word within
+ * struct flow. */
+#define ASSERT_SAME_WORD(f0, f1) \
+ BUILD_ASSERT_DECL(offsetof(struct flow, f0) / 4 \
+ == offsetof(struct flow, f1) / 4)
+
+/* Asserts that 'f0' and 'f1' are both sequential and within the same 32-bit
+ * aligned word in struct flow. */
+#define ASSERT_SEQUENTIAL_SAME_WORD(f0, f1) \
+ ASSERT_SEQUENTIAL(f0, f1); \
+ ASSERT_SAME_WORD(f0, f1)
+
/* miniflow_extract() assumes the following to be true to optimize the
* extraction process. */
-BUILD_ASSERT_DECL(offsetof(struct flow, dl_type) + 2
- == offsetof(struct flow, vlan_tci) &&
- offsetof(struct flow, dl_type) / 4
- == offsetof(struct flow, vlan_tci) / 4 );
-
-BUILD_ASSERT_DECL(offsetof(struct flow, nw_frag) + 3
- == offsetof(struct flow, nw_proto) &&
- offsetof(struct flow, nw_tos) + 2
- == offsetof(struct flow, nw_proto) &&
- offsetof(struct flow, nw_ttl) + 1
- == offsetof(struct flow, nw_proto) &&
- offsetof(struct flow, nw_frag) / 4
- == offsetof(struct flow, nw_tos) / 4 &&
- offsetof(struct flow, nw_ttl) / 4
- == offsetof(struct flow, nw_tos) / 4 &&
- offsetof(struct flow, nw_proto) / 4
- == offsetof(struct flow, nw_tos) / 4);
-
-/* TCP flags in the first half of a BE32, zeroes in the other half. */
-BUILD_ASSERT_DECL(offsetof(struct flow, tcp_flags) + 2
- == offsetof(struct flow, pad) &&
- offsetof(struct flow, tcp_flags) / 4
- == offsetof(struct flow, pad) / 4);
+ASSERT_SEQUENTIAL_SAME_WORD(dl_type, vlan_tci);
+
+ASSERT_SEQUENTIAL_SAME_WORD(nw_frag, nw_tos);
+ASSERT_SEQUENTIAL_SAME_WORD(nw_tos, nw_ttl);
+ASSERT_SEQUENTIAL_SAME_WORD(nw_ttl, nw_proto);
+
+/* TCP flags in the middle of a BE64, zeroes in the other half. */
+BUILD_ASSERT_DECL(offsetof(struct flow, tcp_flags) % 8 == 4);
+
#if WORDS_BIGENDIAN
#define TCP_FLAGS_BE32(tcp_ctl) ((OVS_FORCE ovs_be32)TCP_FLAGS_BE16(tcp_ctl) \
<< 16)
#define TCP_FLAGS_BE32(tcp_ctl) ((OVS_FORCE ovs_be32)TCP_FLAGS_BE16(tcp_ctl))
#endif
-BUILD_ASSERT_DECL(offsetof(struct flow, tp_src) + 2
- == offsetof(struct flow, tp_dst) &&
- offsetof(struct flow, tp_src) / 4
- == offsetof(struct flow, tp_dst) / 4);
+ASSERT_SEQUENTIAL_SAME_WORD(tp_src, tp_dst);
/* Removes 'size' bytes from the head end of '*datap', of size '*sizep', which
* must contain at least 'size' bytes of data. Returns the first byte of data
* removed. */
static inline const void *
-data_pull(void **datap, size_t *sizep, size_t size)
+data_pull(const void **datap, size_t *sizep, size_t size)
{
- char *data = (char *)*datap;
+ const char *data = *datap;
*datap = data + size;
*sizep -= size;
return data;
* the head end of '*datap' and returns the first byte removed. Otherwise,
* returns a null pointer without modifying '*datap'. */
static inline const void *
-data_try_pull(void **datap, size_t *sizep, size_t size)
+data_try_pull(const void **datap, size_t *sizep, size_t size)
{
return OVS_LIKELY(*sizep >= size) ? data_pull(datap, sizep, size) : NULL;
}
/* Context for pushing data to a miniflow. */
struct mf_ctx {
- uint64_t map;
- uint32_t *data;
- uint32_t * const end;
+ struct flowmap map;
+ uint64_t *data;
+ uint64_t * const end;
};
/* miniflow_push_* macros allow filling in a miniflow data values in order.
* Assertions are needed only when the layout of the struct flow is modified.
* 'ofs' is a compile-time constant, which allows most of the code be optimized
- * away. Some GCC versions gave warnigns on ALWAYS_INLINE, so these are
+ * away. Some GCC versions gave warnings on ALWAYS_INLINE, so these are
* defined as macros. */
-#if (FLOW_WC_SEQ != 26)
+#if (FLOW_WC_SEQ != 35)
#define MINIFLOW_ASSERT(X) ovs_assert(X)
+BUILD_MESSAGE("FLOW_WC_SEQ changed: miniflow_extract() will have runtime "
+ "assertions enabled. Consider updating FLOW_WC_SEQ after "
+ "testing")
#else
#define MINIFLOW_ASSERT(X)
#endif
-#define miniflow_push_uint32_(MF, OFS, VALUE) \
-{ \
- MINIFLOW_ASSERT(MF.data < MF.end && (OFS) % 4 == 0 \
- && !(MF.map & (UINT64_MAX << (OFS) / 4))); \
- *MF.data++ = VALUE; \
- MF.map |= UINT64_C(1) << (OFS) / 4; \
+/* True if 'IDX' and higher bits are not set. */
+#define ASSERT_FLOWMAP_NOT_SET(FM, IDX) \
+{ \
+ MINIFLOW_ASSERT(!((FM)->bits[(IDX) / MAP_T_BITS] & \
+ (MAP_MAX << ((IDX) % MAP_T_BITS)))); \
+ for (size_t i = (IDX) / MAP_T_BITS + 1; i < FLOWMAP_UNITS; i++) { \
+ MINIFLOW_ASSERT(!(FM)->bits[i]); \
+ } \
+}
+
+#define miniflow_set_map(MF, OFS) \
+ { \
+ ASSERT_FLOWMAP_NOT_SET(&MF.map, (OFS)); \
+ flowmap_set(&MF.map, (OFS), 1); \
+}
+
+#define miniflow_assert_in_map(MF, OFS) \
+ MINIFLOW_ASSERT(flowmap_is_set(&MF.map, (OFS))); \
+ ASSERT_FLOWMAP_NOT_SET(&MF.map, (OFS) + 1)
+
+#define miniflow_push_uint64_(MF, OFS, VALUE) \
+{ \
+ MINIFLOW_ASSERT(MF.data < MF.end && (OFS) % 8 == 0); \
+ *MF.data++ = VALUE; \
+ miniflow_set_map(MF, OFS / 8); \
+}
+
+#define miniflow_push_be64_(MF, OFS, VALUE) \
+ miniflow_push_uint64_(MF, OFS, (OVS_FORCE uint64_t)(VALUE))
+
+#define miniflow_push_uint32_(MF, OFS, VALUE) \
+ { \
+ MINIFLOW_ASSERT(MF.data < MF.end); \
+ \
+ if ((OFS) % 8 == 0) { \
+ miniflow_set_map(MF, OFS / 8); \
+ *(uint32_t *)MF.data = VALUE; \
+ } else if ((OFS) % 8 == 4) { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint32_t *)MF.data + 1) = VALUE; \
+ MF.data++; \
+ } \
}
-#define miniflow_push_be32_(MF, OFS, VALUE) \
+#define miniflow_push_be32_(MF, OFS, VALUE) \
miniflow_push_uint32_(MF, OFS, (OVS_FORCE uint32_t)(VALUE))
-#define miniflow_push_uint16_(MF, OFS, VALUE) \
+#define miniflow_push_uint16_(MF, OFS, VALUE) \
+{ \
+ MINIFLOW_ASSERT(MF.data < MF.end); \
+ \
+ if ((OFS) % 8 == 0) { \
+ miniflow_set_map(MF, OFS / 8); \
+ *(uint16_t *)MF.data = VALUE; \
+ } else if ((OFS) % 8 == 2) { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint16_t *)MF.data + 1) = VALUE; \
+ } else if ((OFS) % 8 == 4) { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint16_t *)MF.data + 2) = VALUE; \
+ } else if ((OFS) % 8 == 6) { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint16_t *)MF.data + 3) = VALUE; \
+ MF.data++; \
+ } \
+}
+
+#define miniflow_push_uint8_(MF, OFS, VALUE) \
+{ \
+ MINIFLOW_ASSERT(MF.data < MF.end); \
+ \
+ if ((OFS) % 8 == 0) { \
+ miniflow_set_map(MF, OFS / 8); \
+ *(uint8_t *)MF.data = VALUE; \
+ } else if ((OFS) % 8 == 7) { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint8_t *)MF.data + 7) = VALUE; \
+ MF.data++; \
+ } else { \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ *((uint8_t *)MF.data + ((OFS) % 8)) = VALUE; \
+ } \
+}
+
+#define miniflow_pad_to_64_(MF, OFS) \
{ \
- MINIFLOW_ASSERT(MF.data < MF.end && \
- (((OFS) % 4 == 0 && !(MF.map & (UINT64_MAX << (OFS) / 4))) \
- || ((OFS) % 4 == 2 && MF.map & (UINT64_C(1) << (OFS) / 4) \
- && !(MF.map & (UINT64_MAX << ((OFS) / 4 + 1)))))); \
- \
- if ((OFS) % 4 == 0) { \
- *(uint16_t *)MF.data = VALUE; \
- MF.map |= UINT64_C(1) << (OFS) / 4; \
- } else if ((OFS) % 4 == 2) { \
- *((uint16_t *)MF.data + 1) = VALUE; \
- MF.data++; \
- } \
+ MINIFLOW_ASSERT((OFS) % 8 != 0); \
+ miniflow_assert_in_map(MF, OFS / 8); \
+ \
+ memset((uint8_t *)MF.data + (OFS) % 8, 0, 8 - (OFS) % 8); \
+ MF.data++; \
+}
+
+#define miniflow_pad_from_64_(MF, OFS) \
+{ \
+ MINIFLOW_ASSERT(MF.data < MF.end); \
+ \
+ MINIFLOW_ASSERT((OFS) % 8 != 0); \
+ miniflow_set_map(MF, OFS / 8); \
+ \
+ memset((uint8_t *)MF.data, 0, (OFS) % 8); \
}
-#define miniflow_push_be16_(MF, OFS, VALUE) \
+#define miniflow_push_be16_(MF, OFS, VALUE) \
miniflow_push_uint16_(MF, OFS, (OVS_FORCE uint16_t)VALUE);
+#define miniflow_push_be8_(MF, OFS, VALUE) \
+ miniflow_push_uint8_(MF, OFS, (OVS_FORCE uint8_t)VALUE);
+
+#define miniflow_set_maps(MF, OFS, N_WORDS) \
+{ \
+ size_t ofs = (OFS); \
+ size_t n_words = (N_WORDS); \
+ \
+ MINIFLOW_ASSERT(n_words && MF.data + n_words <= MF.end); \
+ ASSERT_FLOWMAP_NOT_SET(&MF.map, ofs); \
+ flowmap_set(&MF.map, ofs, n_words); \
+}
+
/* Data at 'valuep' may be unaligned. */
#define miniflow_push_words_(MF, OFS, VALUEP, N_WORDS) \
{ \
- int ofs32 = (OFS) / 4; \
- \
- MINIFLOW_ASSERT(MF.data + (N_WORDS) <= MF.end && (OFS) % 4 == 0 \
- && !(MF.map & (UINT64_MAX << ofs32))); \
- \
- memcpy(MF.data, (VALUEP), (N_WORDS) * sizeof *MF.data); \
- MF.data += (N_WORDS); \
- MF.map |= ((UINT64_MAX >> (64 - (N_WORDS))) << ofs32); \
+ MINIFLOW_ASSERT((OFS) % 8 == 0); \
+ miniflow_set_maps(MF, (OFS) / 8, (N_WORDS)); \
+ memcpy(MF.data, (VALUEP), (N_WORDS) * sizeof *MF.data); \
+ MF.data += (N_WORDS); \
+}
+
+/* Push 32-bit words padded to 64-bits. */
+#define miniflow_push_words_32_(MF, OFS, VALUEP, N_WORDS) \
+{ \
+ miniflow_set_maps(MF, (OFS) / 8, DIV_ROUND_UP(N_WORDS, 2)); \
+ memcpy(MF.data, (VALUEP), (N_WORDS) * sizeof(uint32_t)); \
+ MF.data += DIV_ROUND_UP(N_WORDS, 2); \
+ if ((N_WORDS) & 1) { \
+ *((uint32_t *)MF.data - 1) = 0; \
+ } \
+}
+
+/* Data at 'valuep' may be unaligned. */
+/* MACs start 64-aligned, and must be followed by other data or padding. */
+#define miniflow_push_macs_(MF, OFS, VALUEP) \
+{ \
+ miniflow_set_maps(MF, (OFS) / 8, 2); \
+ memcpy(MF.data, (VALUEP), 2 * ETH_ADDR_LEN); \
+ MF.data += 1; /* First word only. */ \
}
-#define miniflow_push_uint32(MF, FIELD, VALUE) \
+#define miniflow_push_uint32(MF, FIELD, VALUE) \
miniflow_push_uint32_(MF, offsetof(struct flow, FIELD), VALUE)
-#define miniflow_push_be32(MF, FIELD, VALUE) \
+#define miniflow_push_be32(MF, FIELD, VALUE) \
miniflow_push_be32_(MF, offsetof(struct flow, FIELD), VALUE)
-#define miniflow_push_uint32_check(MF, FIELD, VALUE) \
- { if (OVS_LIKELY(VALUE)) { \
- miniflow_push_uint32_(MF, offsetof(struct flow, FIELD), VALUE); \
- } \
- }
-
-#define miniflow_push_be32_check(MF, FIELD, VALUE) \
- { if (OVS_LIKELY(VALUE)) { \
- miniflow_push_be32_(MF, offsetof(struct flow, FIELD), VALUE); \
- } \
- }
-
-#define miniflow_push_uint16(MF, FIELD, VALUE) \
+#define miniflow_push_uint16(MF, FIELD, VALUE) \
miniflow_push_uint16_(MF, offsetof(struct flow, FIELD), VALUE)
-#define miniflow_push_be16(MF, FIELD, VALUE) \
+#define miniflow_push_be16(MF, FIELD, VALUE) \
miniflow_push_be16_(MF, offsetof(struct flow, FIELD), VALUE)
+#define miniflow_push_uint8(MF, FIELD, VALUE) \
+ miniflow_push_uint8_(MF, offsetof(struct flow, FIELD), VALUE)
+
+#define miniflow_pad_to_64(MF, FIELD) \
+ miniflow_pad_to_64_(MF, OFFSETOFEND(struct flow, FIELD))
+
+#define miniflow_pad_from_64(MF, FIELD) \
+ miniflow_pad_from_64_(MF, offsetof(struct flow, FIELD))
+
#define miniflow_push_words(MF, FIELD, VALUEP, N_WORDS) \
miniflow_push_words_(MF, offsetof(struct flow, FIELD), VALUEP, N_WORDS)
+#define miniflow_push_words_32(MF, FIELD, VALUEP, N_WORDS) \
+ miniflow_push_words_32_(MF, offsetof(struct flow, FIELD), VALUEP, N_WORDS)
+
+#define miniflow_push_macs(MF, FIELD, VALUEP) \
+ miniflow_push_macs_(MF, offsetof(struct flow, FIELD), VALUEP)
+
/* Pulls the MPLS headers at '*datap' and returns the count of them. */
static inline int
-parse_mpls(void **datap, size_t *sizep)
+parse_mpls(const void **datap, size_t *sizep)
{
const struct mpls_hdr *mh;
int count = 0;
break;
}
}
- return MAX(count, FLOW_MAX_MPLS_LABELS);
+ return MIN(count, FLOW_MAX_MPLS_LABELS);
}
static inline ovs_be16
-parse_vlan(void **datap, size_t *sizep)
+parse_vlan(const void **datap, size_t *sizep)
{
const struct eth_header *eth = *datap;
}
static inline ovs_be16
-parse_ethertype(void **datap, size_t *sizep)
+parse_ethertype(const void **datap, size_t *sizep)
{
const struct llc_snap_header *llc;
ovs_be16 proto;
return htons(FLOW_DL_TYPE_NONE);
}
-static inline bool
-parse_icmpv6(void **datap, size_t *sizep, const struct icmp6_hdr *icmp,
+static inline void
+parse_icmpv6(const void **datap, size_t *sizep, const struct icmp6_hdr *icmp,
const struct in6_addr **nd_target,
- uint8_t arp_buf[2][ETH_ADDR_LEN])
+ struct eth_addr arp_buf[2])
{
if (icmp->icmp6_code == 0 &&
(icmp->icmp6_type == ND_NEIGHBOR_SOLICIT ||
icmp->icmp6_type == ND_NEIGHBOR_ADVERT)) {
- *nd_target = data_try_pull(datap, sizep, sizeof *nd_target);
+ *nd_target = data_try_pull(datap, sizep, sizeof **nd_target);
if (OVS_UNLIKELY(!*nd_target)) {
- return false;
+ return;
}
while (*sizep >= 8) {
/* The minimum size of an option is 8 bytes, which also is
* the size of Ethernet link-layer options. */
- const struct nd_opt_hdr *nd_opt = *datap;
- int opt_len = nd_opt->nd_opt_len * 8;
+ const struct ovs_nd_opt *nd_opt = *datap;
+ int opt_len = nd_opt->nd_opt_len * ND_OPT_LEN;
if (!opt_len || opt_len > *sizep) {
- goto invalid;
+ return;
}
/* Store the link layer address if the appropriate option is
* provided. It is considered an error if the same link
* layer option is specified twice. */
if (nd_opt->nd_opt_type == ND_OPT_SOURCE_LINKADDR
- && opt_len == 8) {
+ && opt_len == 8) {
if (OVS_LIKELY(eth_addr_is_zero(arp_buf[0]))) {
- memcpy(arp_buf[0], nd_opt + 1, ETH_ADDR_LEN);
+ arp_buf[0] = nd_opt->nd_opt_mac;
} else {
goto invalid;
}
} else if (nd_opt->nd_opt_type == ND_OPT_TARGET_LINKADDR
- && opt_len == 8) {
+ && opt_len == 8) {
if (OVS_LIKELY(eth_addr_is_zero(arp_buf[1]))) {
- memcpy(arp_buf[1], nd_opt + 1, ETH_ADDR_LEN);
+ arp_buf[1] = nd_opt->nd_opt_mac;
} else {
goto invalid;
}
}
if (OVS_UNLIKELY(!data_try_pull(datap, sizep, opt_len))) {
- goto invalid;
+ return;
}
}
}
- return true;
+ return;
invalid:
- return false;
+ *nd_target = NULL;
+ arp_buf[0] = eth_addr_zero;
+ arp_buf[1] = eth_addr_zero;
}
/* Initializes 'flow' members from 'packet' and 'md'
* otherwise UINT16_MAX.
*/
void
-flow_extract(struct ofpbuf *packet, const struct pkt_metadata *md,
- struct flow *flow)
+flow_extract(struct dp_packet *packet, struct flow *flow)
{
- uint32_t buf[FLOW_U32S];
- struct miniflow mf;
+ struct {
+ struct miniflow mf;
+ uint64_t buf[FLOW_U64S];
+ } m;
COVERAGE_INC(flow_extract);
- miniflow_initialize(&mf, buf);
- miniflow_extract(packet, md, &mf);
- miniflow_expand(&mf, flow);
+ miniflow_extract(packet, &m.mf);
+ miniflow_expand(&m.mf, flow);
}
-/* Caller is responsible for initializing 'dst->values' with enough storage
- * (FLOW_U64S * 8 bytes is enough). */
+/* Caller is responsible for initializing 'dst' with enough storage for
+ * FLOW_U64S * 8 bytes. */
void
-miniflow_extract(struct ofpbuf *packet, const struct pkt_metadata *md,
- struct miniflow *dst)
-{
- void *data = ofpbuf_data(packet);
- size_t size = ofpbuf_size(packet);
- char *l2;
- struct mf_ctx mf = { 0, dst->values, dst->values + FLOW_U32S };
+miniflow_extract(struct dp_packet *packet, struct miniflow *dst)
+{
+ const struct pkt_metadata *md = &packet->md;
+ const void *data = dp_packet_data(packet);
+ size_t size = dp_packet_size(packet);
+ uint64_t *values = miniflow_values(dst);
+ struct mf_ctx mf = { FLOWMAP_EMPTY_INITIALIZER, values,
+ values + FLOW_U64S };
+ const char *l2;
ovs_be16 dl_type;
uint8_t nw_frag, nw_tos, nw_ttl, nw_proto;
/* Metadata. */
- if (md) {
- if (md->tunnel.ip_dst) {
- miniflow_push_words(mf, tunnel, &md->tunnel,
- sizeof md->tunnel / 4);
+ if (flow_tnl_dst_is_set(&md->tunnel)) {
+ miniflow_push_words(mf, tunnel, &md->tunnel,
+ offsetof(struct flow_tnl, metadata) /
+ sizeof(uint64_t));
+
+ if (!(md->tunnel.flags & FLOW_TNL_F_UDPIF)) {
+ if (md->tunnel.metadata.present.map) {
+ miniflow_push_words(mf, tunnel.metadata, &md->tunnel.metadata,
+ sizeof md->tunnel.metadata /
+ sizeof(uint64_t));
+ }
+ } else {
+ if (md->tunnel.metadata.present.len) {
+ miniflow_push_words(mf, tunnel.metadata.present,
+ &md->tunnel.metadata.present, 1);
+ miniflow_push_words(mf, tunnel.metadata.opts.gnv,
+ md->tunnel.metadata.opts.gnv,
+ DIV_ROUND_UP(md->tunnel.metadata.present.len,
+ sizeof(uint64_t)));
+ }
+ }
+ }
+ if (md->skb_priority || md->pkt_mark) {
+ miniflow_push_uint32(mf, skb_priority, md->skb_priority);
+ miniflow_push_uint32(mf, pkt_mark, md->pkt_mark);
+ }
+ miniflow_push_uint32(mf, dp_hash, md->dp_hash);
+ miniflow_push_uint32(mf, in_port, odp_to_u32(md->in_port.odp_port));
+ if (md->recirc_id || md->ct_state) {
+ miniflow_push_uint32(mf, recirc_id, md->recirc_id);
+ miniflow_push_uint16(mf, ct_state, md->ct_state);
+ miniflow_push_uint16(mf, ct_zone, md->ct_zone);
+ }
+
+ if (md->ct_state) {
+ miniflow_push_uint32(mf, ct_mark, md->ct_mark);
+ miniflow_pad_to_64(mf, ct_mark);
+
+ if (!ovs_u128_is_zero(&md->ct_label)) {
+ miniflow_push_words(mf, ct_label, &md->ct_label,
+ sizeof md->ct_label / sizeof(uint64_t));
}
- miniflow_push_uint32_check(mf, skb_priority, md->skb_priority);
- miniflow_push_uint32_check(mf, pkt_mark, md->pkt_mark);
- miniflow_push_uint32_check(mf, recirc_id, md->recirc_id);
- miniflow_push_uint32(mf, in_port, odp_to_u32(md->in_port.odp_port));
}
/* Initialize packet's layer pointer and offsets. */
l2 = data;
- ofpbuf_set_frame(packet, data);
+ dp_packet_reset_offsets(packet);
/* Must have full Ethernet header to proceed. */
if (OVS_UNLIKELY(size < sizeof(struct eth_header))) {
ovs_be16 vlan_tci;
/* Link layer. */
- BUILD_ASSERT(offsetof(struct flow, dl_dst) + 6
- == offsetof(struct flow, dl_src));
- miniflow_push_words(mf, dl_dst, data, ETH_ADDR_LEN * 2 / 4);
+ ASSERT_SEQUENTIAL(dl_dst, dl_src);
+ miniflow_push_macs(mf, dl_dst, data);
/* dl_type, vlan_tci. */
vlan_tci = parse_vlan(&data, &size);
dl_type = parse_ethertype(&data, &size);
packet->l2_5_ofs = (char *)data - l2;
count = parse_mpls(&data, &size);
- miniflow_push_words(mf, mpls_lse, mpls, count);
+ miniflow_push_words_32(mf, mpls_lse, mpls, count);
}
/* Network layer. */
if (OVS_LIKELY(dl_type == htons(ETH_TYPE_IP))) {
const struct ip_header *nh = data;
int ip_len;
+ uint16_t tot_len;
if (OVS_UNLIKELY(size < IP_HEADER_LEN)) {
goto out;
if (OVS_UNLIKELY(ip_len < IP_HEADER_LEN)) {
goto out;
}
+ if (OVS_UNLIKELY(size < ip_len)) {
+ goto out;
+ }
+ tot_len = ntohs(nh->ip_tot_len);
+ if (OVS_UNLIKELY(tot_len > size)) {
+ goto out;
+ }
+ if (OVS_UNLIKELY(size - tot_len > UINT8_MAX)) {
+ goto out;
+ }
+ dp_packet_set_l2_pad_size(packet, size - tot_len);
+ size = tot_len; /* Never pull padding. */
/* Push both source and destination address at once. */
- miniflow_push_words(mf, nw_src, &nh->ip_src, 2);
+ miniflow_push_words(mf, nw_src, &nh->ip_src, 1);
+
+ miniflow_push_be32(mf, ipv6_label, 0); /* Padding for IPv4. */
nw_tos = nh->ip_tos;
nw_ttl = nh->ip_ttl;
nw_frag |= FLOW_NW_FRAG_LATER;
}
}
- if (OVS_UNLIKELY(size < ip_len)) {
- goto out;
- }
data_pull(&data, &size, ip_len);
-
} else if (dl_type == htons(ETH_TYPE_IPV6)) {
const struct ovs_16aligned_ip6_hdr *nh;
ovs_be32 tc_flow;
+ uint16_t plen;
if (OVS_UNLIKELY(size < sizeof *nh)) {
goto out;
}
nh = data_pull(&data, &size, sizeof *nh);
+ plen = ntohs(nh->ip6_plen);
+ if (OVS_UNLIKELY(plen > size)) {
+ goto out;
+ }
+ /* Jumbo Payload option not supported yet. */
+ if (OVS_UNLIKELY(size - plen > UINT8_MAX)) {
+ goto out;
+ }
+ dp_packet_set_l2_pad_size(packet, size - plen);
+ size = plen; /* Never pull padding. */
+
miniflow_push_words(mf, ipv6_src, &nh->ip6_src,
- sizeof nh->ip6_src / 4);
+ sizeof nh->ip6_src / 8);
miniflow_push_words(mf, ipv6_dst, &nh->ip6_dst,
- sizeof nh->ip6_dst / 4);
+ sizeof nh->ip6_dst / 8);
tc_flow = get_16aligned_be32(&nh->ip6_flow);
{
ovs_be32 label = tc_flow & htonl(IPV6_LABEL_MASK);
- miniflow_push_be32_check(mf, ipv6_label, label);
+ miniflow_push_be32(mf, ipv6_label, label);
}
nw_tos = ntohl(tc_flow) >> 20;
} else {
if (dl_type == htons(ETH_TYPE_ARP) ||
dl_type == htons(ETH_TYPE_RARP)) {
- uint8_t arp_buf[2][ETH_ADDR_LEN];
+ struct eth_addr arp_buf[2];
const struct arp_eth_header *arp = (const struct arp_eth_header *)
data_try_pull(&data, &size, ARP_ETH_HEADER_LEN);
&& OVS_LIKELY(arp->ar_pro == htons(ETH_TYPE_IP))
&& OVS_LIKELY(arp->ar_hln == ETH_ADDR_LEN)
&& OVS_LIKELY(arp->ar_pln == 4)) {
- miniflow_push_words(mf, nw_src, &arp->ar_spa, 1);
- miniflow_push_words(mf, nw_dst, &arp->ar_tpa, 1);
+ miniflow_push_be32(mf, nw_src,
+ get_16aligned_be32(&arp->ar_spa));
+ miniflow_push_be32(mf, nw_dst,
+ get_16aligned_be32(&arp->ar_tpa));
/* We only match on the lower 8 bits of the opcode. */
if (OVS_LIKELY(ntohs(arp->ar_op) <= 0xff)) {
+ miniflow_push_be32(mf, ipv6_label, 0); /* Pad with ARP. */
miniflow_push_be32(mf, nw_frag, htonl(ntohs(arp->ar_op)));
}
/* Must be adjacent. */
- BUILD_ASSERT(offsetof(struct flow, arp_sha) + 6
- == offsetof(struct flow, arp_tha));
+ ASSERT_SEQUENTIAL(arp_sha, arp_tha);
- memcpy(arp_buf[0], arp->ar_sha, ETH_ADDR_LEN);
- memcpy(arp_buf[1], arp->ar_tha, ETH_ADDR_LEN);
- miniflow_push_words(mf, arp_sha, arp_buf,
- ETH_ADDR_LEN * 2 / 4);
+ arp_buf[0] = arp->ar_sha;
+ arp_buf[1] = arp->ar_tha;
+ miniflow_push_macs(mf, arp_sha, arp_buf);
+ miniflow_pad_to_64(mf, arp_tha);
}
}
goto out;
if (OVS_LIKELY(size >= TCP_HEADER_LEN)) {
const struct tcp_header *tcp = data;
+ miniflow_push_be32(mf, arp_tha.ea[2], 0);
miniflow_push_be32(mf, tcp_flags,
TCP_FLAGS_BE32(tcp->tcp_ctl));
- miniflow_push_words(mf, tp_src, &tcp->tcp_src, 1);
+ miniflow_push_be16(mf, tp_src, tcp->tcp_src);
+ miniflow_push_be16(mf, tp_dst, tcp->tcp_dst);
+ miniflow_pad_to_64(mf, tp_dst);
}
} else if (OVS_LIKELY(nw_proto == IPPROTO_UDP)) {
if (OVS_LIKELY(size >= UDP_HEADER_LEN)) {
const struct udp_header *udp = data;
- miniflow_push_words(mf, tp_src, &udp->udp_src, 1);
+ miniflow_push_be16(mf, tp_src, udp->udp_src);
+ miniflow_push_be16(mf, tp_dst, udp->udp_dst);
+ miniflow_pad_to_64(mf, tp_dst);
}
} else if (OVS_LIKELY(nw_proto == IPPROTO_SCTP)) {
if (OVS_LIKELY(size >= SCTP_HEADER_LEN)) {
const struct sctp_header *sctp = data;
- miniflow_push_words(mf, tp_src, &sctp->sctp_src, 1);
+ miniflow_push_be16(mf, tp_src, sctp->sctp_src);
+ miniflow_push_be16(mf, tp_dst, sctp->sctp_dst);
+ miniflow_pad_to_64(mf, tp_dst);
}
} else if (OVS_LIKELY(nw_proto == IPPROTO_ICMP)) {
if (OVS_LIKELY(size >= ICMP_HEADER_LEN)) {
miniflow_push_be16(mf, tp_src, htons(icmp->icmp_type));
miniflow_push_be16(mf, tp_dst, htons(icmp->icmp_code));
+ miniflow_pad_to_64(mf, tp_dst);
+ }
+ } else if (OVS_LIKELY(nw_proto == IPPROTO_IGMP)) {
+ if (OVS_LIKELY(size >= IGMP_HEADER_LEN)) {
+ const struct igmp_header *igmp = data;
+
+ miniflow_push_be16(mf, tp_src, htons(igmp->igmp_type));
+ miniflow_push_be16(mf, tp_dst, htons(igmp->igmp_code));
+ miniflow_push_be32(mf, igmp_group_ip4,
+ get_16aligned_be32(&igmp->group));
}
} else if (OVS_LIKELY(nw_proto == IPPROTO_ICMPV6)) {
if (OVS_LIKELY(size >= sizeof(struct icmp6_hdr))) {
const struct in6_addr *nd_target = NULL;
- uint8_t arp_buf[2][ETH_ADDR_LEN];
+ struct eth_addr arp_buf[2] = { { { { 0 } } } };
const struct icmp6_hdr *icmp = data_pull(&data, &size,
sizeof *icmp);
- memset(arp_buf, 0, sizeof arp_buf);
- if (OVS_LIKELY(parse_icmpv6(&data, &size, icmp, &nd_target,
- arp_buf))) {
- if (nd_target) {
- miniflow_push_words(mf, nd_target, nd_target,
- sizeof *nd_target / 4);
- }
- miniflow_push_words(mf, arp_sha, arp_buf,
- ETH_ADDR_LEN * 2 / 4);
- miniflow_push_be16(mf, tp_src, htons(icmp->icmp6_type));
- miniflow_push_be16(mf, tp_dst, htons(icmp->icmp6_code));
+ parse_icmpv6(&data, &size, icmp, &nd_target, arp_buf);
+ if (nd_target) {
+ miniflow_push_words(mf, nd_target, nd_target,
+ sizeof *nd_target / sizeof(uint64_t));
}
+ miniflow_push_macs(mf, arp_sha, arp_buf);
+ miniflow_pad_to_64(mf, arp_tha);
+ miniflow_push_be16(mf, tp_src, htons(icmp->icmp6_type));
+ miniflow_push_be16(mf, tp_dst, htons(icmp->icmp6_code));
+ miniflow_pad_to_64(mf, tp_dst);
}
}
}
- if (md) {
- miniflow_push_uint32_check(mf, dp_hash, md->dp_hash);
- }
out:
dst->map = mf.map;
}
void
flow_zero_wildcards(struct flow *flow, const struct flow_wildcards *wildcards)
{
- uint32_t *flow_u32 = (uint32_t *) flow;
- const uint32_t *wc_u32 = (const uint32_t *) &wildcards->masks;
+ uint64_t *flow_u64 = (uint64_t *) flow;
+ const uint64_t *wc_u64 = (const uint64_t *) &wildcards->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- flow_u32[i] &= wc_u32[i];
+ for (i = 0; i < FLOW_U64S; i++) {
+ flow_u64[i] &= wc_u64[i];
}
}
}
}
-/* Initializes 'fmd' with the metadata found in 'flow'. */
+/* Initializes 'flow_metadata' with the metadata found in 'flow'. */
void
-flow_get_metadata(const struct flow *flow, struct flow_metadata *fmd)
+flow_get_metadata(const struct flow *flow, struct match *flow_metadata)
{
- BUILD_ASSERT_DECL(FLOW_WC_SEQ == 26);
+ int i;
+
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 35);
- fmd->dp_hash = flow->dp_hash;
- fmd->recirc_id = flow->recirc_id;
- fmd->tun_id = flow->tunnel.tun_id;
- fmd->tun_src = flow->tunnel.ip_src;
- fmd->tun_dst = flow->tunnel.ip_dst;
- fmd->metadata = flow->metadata;
- memcpy(fmd->regs, flow->regs, sizeof fmd->regs);
- fmd->pkt_mark = flow->pkt_mark;
- fmd->in_port = flow->in_port.ofp_port;
+ match_init_catchall(flow_metadata);
+ if (flow->tunnel.tun_id != htonll(0)) {
+ match_set_tun_id(flow_metadata, flow->tunnel.tun_id);
+ }
+ if (flow->tunnel.flags & FLOW_TNL_PUB_F_MASK) {
+ match_set_tun_flags(flow_metadata,
+ flow->tunnel.flags & FLOW_TNL_PUB_F_MASK);
+ }
+ if (flow->tunnel.ip_src) {
+ match_set_tun_src(flow_metadata, flow->tunnel.ip_src);
+ }
+ if (flow->tunnel.ip_dst) {
+ match_set_tun_dst(flow_metadata, flow->tunnel.ip_dst);
+ }
+ if (ipv6_addr_is_set(&flow->tunnel.ipv6_src)) {
+ match_set_tun_ipv6_src(flow_metadata, &flow->tunnel.ipv6_src);
+ }
+ if (ipv6_addr_is_set(&flow->tunnel.ipv6_dst)) {
+ match_set_tun_ipv6_dst(flow_metadata, &flow->tunnel.ipv6_dst);
+ }
+ if (flow->tunnel.gbp_id != htons(0)) {
+ match_set_tun_gbp_id(flow_metadata, flow->tunnel.gbp_id);
+ }
+ if (flow->tunnel.gbp_flags) {
+ match_set_tun_gbp_flags(flow_metadata, flow->tunnel.gbp_flags);
+ }
+ tun_metadata_get_fmd(&flow->tunnel, flow_metadata);
+ if (flow->metadata != htonll(0)) {
+ match_set_metadata(flow_metadata, flow->metadata);
+ }
+
+ for (i = 0; i < FLOW_N_REGS; i++) {
+ if (flow->regs[i]) {
+ match_set_reg(flow_metadata, i, flow->regs[i]);
+ }
+ }
+
+ if (flow->pkt_mark != 0) {
+ match_set_pkt_mark(flow_metadata, flow->pkt_mark);
+ }
+
+ match_set_in_port(flow_metadata, flow->in_port.ofp_port);
+ if (flow->ct_state != 0) {
+ match_set_ct_state(flow_metadata, flow->ct_state);
+ }
+ if (flow->ct_zone != 0) {
+ match_set_ct_zone(flow_metadata, flow->ct_zone);
+ }
+ if (flow->ct_mark != 0) {
+ match_set_ct_mark(flow_metadata, flow->ct_mark);
+ }
+ if (!ovs_u128_is_zero(&flow->ct_label)) {
+ match_set_ct_label(flow_metadata, flow->ct_label);
+ }
+}
+
+const char *ct_state_to_string(uint32_t state)
+{
+ switch (state) {
+ case CS_REPLY_DIR:
+ return "rpl";
+ case CS_TRACKED:
+ return "trk";
+ case CS_NEW:
+ return "new";
+ case CS_ESTABLISHED:
+ return "est";
+ case CS_RELATED:
+ return "rel";
+ case CS_INVALID:
+ return "inv";
+ case CS_SRC_NAT:
+ return "snat";
+ case CS_DST_NAT:
+ return "dnat";
+ default:
+ return NULL;
+ }
}
char *
return "csum";
case FLOW_TNL_F_KEY:
return "key";
+ case FLOW_TNL_F_OAM:
+ return "oam";
default:
return NULL;
}
uint32_t bad = 0;
if (!flags) {
+ ds_put_char(ds, '0');
return;
}
while (flags) {
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 mask, uint32_t max_mask)
{
if (name) {
ds_put_format(ds, "%s=", name);
}
+
+ if (mask == max_mask) {
+ format_flags(ds, bit_to_string, flags, '|');
+ return;
+ }
+
+ if (!mask) {
+ ds_put_cstr(ds, "0/0");
+ return;
+ }
+
while (mask) {
uint32_t bit = rightmost_1bit(mask);
const char *s = bit_to_string(bit);
}
}
+/* Scans a string 's' of flags to determine their numerical value and
+ * returns the number of characters parsed using 'bit_to_string' to
+ * lookup flag names. Scanning continues until the character 'end' is
+ * reached.
+ *
+ * In the event of a failure, a negative error code will be returned. In
+ * addition, if 'res_string' is non-NULL then a descriptive string will
+ * be returned incorporating the identifying string 'field_name'. This
+ * error string must be freed by the caller.
+ *
+ * Upon success, the flag values will be stored in 'res_flags' and
+ * optionally 'res_mask', if it is non-NULL (if it is NULL then any masks
+ * present in the original string will be considered an error). The
+ * caller may restrict the acceptable set of values through the mask
+ * 'allowed'. */
+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)
+{
+ uint32_t result = 0;
+ int n;
+
+ /* Parse masked flags in numeric format? */
+ if (res_mask && ovs_scan(s, "%"SCNi32"/%"SCNi32"%n",
+ res_flags, res_mask, &n) && n > 0) {
+ if (*res_flags & ~allowed || *res_mask & ~allowed) {
+ goto unknown;
+ }
+ return n;
+ }
+
+ n = 0;
+
+ if (res_mask && (*s == '+' || *s == '-')) {
+ uint32_t flags = 0, mask = 0;
+
+ /* Parse masked flags. */
+ while (s[0] != end) {
+ bool set;
+ uint32_t bit;
+ size_t len;
+
+ if (s[0] == '+') {
+ set = true;
+ } else if (s[0] == '-') {
+ set = false;
+ } else {
+ if (res_string) {
+ *res_string = xasprintf("%s: %s must be preceded by '+' "
+ "(for SET) or '-' (NOT SET)", s,
+ field_name);
+ }
+ return -EINVAL;
+ }
+ s++;
+ n++;
+
+ for (bit = 1; bit; bit <<= 1) {
+ const char *fname = bit_to_string(bit);
+
+ if (!fname) {
+ continue;
+ }
+
+ len = strlen(fname);
+ if (strncmp(s, fname, len) ||
+ (s[len] != '+' && s[len] != '-' && s[len] != end)) {
+ continue;
+ }
+
+ if (mask & bit) {
+ /* bit already set. */
+ if (res_string) {
+ *res_string = xasprintf("%s: Each %s flag can be "
+ "specified only once", s,
+ field_name);
+ }
+ return -EINVAL;
+ }
+ if (!(bit & allowed)) {
+ goto unknown;
+ }
+ if (set) {
+ flags |= bit;
+ }
+ mask |= bit;
+ break;
+ }
+
+ if (!bit) {
+ goto unknown;
+ }
+ s += len;
+ n += len;
+ }
+
+ *res_flags = flags;
+ *res_mask = mask;
+ return n;
+ }
+
+ /* Parse unmasked flags. If a flag is present, it is set, otherwise
+ * it is not set. */
+ while (s[n] != end) {
+ unsigned long long int flags;
+ uint32_t bit;
+ int n0;
+
+ if (ovs_scan(&s[n], "%lli%n", &flags, &n0)) {
+ if (flags & ~allowed) {
+ goto unknown;
+ }
+ n += n0 + (s[n + n0] == '|');
+ result |= flags;
+ continue;
+ }
+
+ for (bit = 1; bit; bit <<= 1) {
+ const char *name = bit_to_string(bit);
+ size_t len;
+
+ if (!name) {
+ continue;
+ }
+
+ len = strlen(name);
+ if (!strncmp(s + n, name, len) &&
+ (s[n + len] == '|' || s[n + len] == end)) {
+ if (!(bit & allowed)) {
+ goto unknown;
+ }
+ result |= bit;
+ n += len + (s[n + len] == '|');
+ break;
+ }
+ }
+
+ if (!bit) {
+ goto unknown;
+ }
+ }
+
+ *res_flags = result;
+ if (res_mask) {
+ *res_mask = UINT32_MAX;
+ }
+ if (res_string) {
+ *res_string = NULL;
+ }
+ return n;
+
+unknown:
+ if (res_string) {
+ *res_string = xasprintf("%s: unknown %s flag(s)", s, field_name);
+ }
+ return -EINVAL;
+}
+
void
flow_format(struct ds *ds, const struct flow *flow)
{
struct match match;
+ struct flow_wildcards *wc = &match.wc;
match_wc_init(&match, flow);
+
+ /* As this function is most often used for formatting a packet in a
+ * packet-in message, skip formatting the packet context fields that are
+ * all-zeroes to make the print-out easier on the eyes. This means that a
+ * missing context field implies a zero value for that field. This is
+ * similar to OpenFlow encoding of these fields, as the specification
+ * states that all-zeroes context fields should not be encoded in the
+ * packet-in messages. */
+ if (!flow->in_port.ofp_port) {
+ WC_UNMASK_FIELD(wc, in_port);
+ }
+ if (!flow->skb_priority) {
+ WC_UNMASK_FIELD(wc, skb_priority);
+ }
+ if (!flow->pkt_mark) {
+ WC_UNMASK_FIELD(wc, pkt_mark);
+ }
+ if (!flow->recirc_id) {
+ WC_UNMASK_FIELD(wc, recirc_id);
+ }
+ if (!flow->dp_hash) {
+ WC_UNMASK_FIELD(wc, dp_hash);
+ }
+ if (!flow->ct_state) {
+ WC_UNMASK_FIELD(wc, ct_state);
+ }
+ if (!flow->ct_zone) {
+ WC_UNMASK_FIELD(wc, ct_zone);
+ }
+ if (!flow->ct_mark) {
+ WC_UNMASK_FIELD(wc, ct_mark);
+ }
+ if (ovs_u128_is_zero(&flow->ct_label)) {
+ WC_UNMASK_FIELD(wc, ct_label);
+ }
+ for (int i = 0; i < FLOW_N_REGS; i++) {
+ if (!flow->regs[i]) {
+ WC_UNMASK_FIELD(wc, regs[i]);
+ }
+ }
+ if (!flow->metadata) {
+ WC_UNMASK_FIELD(wc, metadata);
+ }
+
match_format(&match, ds, OFP_DEFAULT_PRIORITY);
}
memset(&wc->masks, 0, sizeof wc->masks);
}
+/* Converts a flow into flow wildcards. It sets the wildcard masks based on
+ * the packet headers extracted to 'flow'. It will not set the mask for fields
+ * that do not make sense for the packet type. OpenFlow-only metadata is
+ * wildcarded, but other metadata is unconditionally exact-matched. */
+void flow_wildcards_init_for_packet(struct flow_wildcards *wc,
+ const struct flow *flow)
+{
+ memset(&wc->masks, 0x0, sizeof wc->masks);
+
+ /* Update this function whenever struct flow changes. */
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 35);
+
+ if (flow_tnl_dst_is_set(&flow->tunnel)) {
+ if (flow->tunnel.flags & FLOW_TNL_F_KEY) {
+ WC_MASK_FIELD(wc, tunnel.tun_id);
+ }
+ WC_MASK_FIELD(wc, tunnel.ip_src);
+ WC_MASK_FIELD(wc, tunnel.ip_dst);
+ WC_MASK_FIELD(wc, tunnel.ipv6_src);
+ WC_MASK_FIELD(wc, tunnel.ipv6_dst);
+ WC_MASK_FIELD(wc, tunnel.flags);
+ WC_MASK_FIELD(wc, tunnel.ip_tos);
+ WC_MASK_FIELD(wc, tunnel.ip_ttl);
+ WC_MASK_FIELD(wc, tunnel.tp_src);
+ WC_MASK_FIELD(wc, tunnel.tp_dst);
+ WC_MASK_FIELD(wc, tunnel.gbp_id);
+ WC_MASK_FIELD(wc, tunnel.gbp_flags);
+
+ if (!(flow->tunnel.flags & FLOW_TNL_F_UDPIF)) {
+ if (flow->tunnel.metadata.present.map) {
+ wc->masks.tunnel.metadata.present.map =
+ flow->tunnel.metadata.present.map;
+ WC_MASK_FIELD(wc, tunnel.metadata.opts.u8);
+ }
+ } else {
+ WC_MASK_FIELD(wc, tunnel.metadata.present.len);
+ memset(wc->masks.tunnel.metadata.opts.gnv, 0xff,
+ flow->tunnel.metadata.present.len);
+ }
+ } else if (flow->tunnel.tun_id) {
+ WC_MASK_FIELD(wc, tunnel.tun_id);
+ }
+
+ /* metadata, regs, and conj_id wildcarded. */
+
+ WC_MASK_FIELD(wc, skb_priority);
+ WC_MASK_FIELD(wc, pkt_mark);
+ WC_MASK_FIELD(wc, ct_state);
+ WC_MASK_FIELD(wc, ct_zone);
+ WC_MASK_FIELD(wc, ct_mark);
+ WC_MASK_FIELD(wc, ct_label);
+ WC_MASK_FIELD(wc, recirc_id);
+ WC_MASK_FIELD(wc, dp_hash);
+ WC_MASK_FIELD(wc, in_port);
+
+ /* actset_output wildcarded. */
+
+ WC_MASK_FIELD(wc, dl_dst);
+ WC_MASK_FIELD(wc, dl_src);
+ WC_MASK_FIELD(wc, dl_type);
+ WC_MASK_FIELD(wc, vlan_tci);
+
+ if (flow->dl_type == htons(ETH_TYPE_IP)) {
+ WC_MASK_FIELD(wc, nw_src);
+ WC_MASK_FIELD(wc, nw_dst);
+ } else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
+ WC_MASK_FIELD(wc, ipv6_src);
+ WC_MASK_FIELD(wc, ipv6_dst);
+ WC_MASK_FIELD(wc, ipv6_label);
+ } else if (flow->dl_type == htons(ETH_TYPE_ARP) ||
+ flow->dl_type == htons(ETH_TYPE_RARP)) {
+ WC_MASK_FIELD(wc, nw_src);
+ WC_MASK_FIELD(wc, nw_dst);
+ WC_MASK_FIELD(wc, nw_proto);
+ WC_MASK_FIELD(wc, arp_sha);
+ WC_MASK_FIELD(wc, arp_tha);
+ return;
+ } else if (eth_type_mpls(flow->dl_type)) {
+ for (int i = 0; i < FLOW_MAX_MPLS_LABELS; i++) {
+ WC_MASK_FIELD(wc, mpls_lse[i]);
+ if (flow->mpls_lse[i] & htonl(MPLS_BOS_MASK)) {
+ break;
+ }
+ }
+ return;
+ } else {
+ return; /* Unknown ethertype. */
+ }
+
+ /* IPv4 or IPv6. */
+ WC_MASK_FIELD(wc, nw_frag);
+ WC_MASK_FIELD(wc, nw_tos);
+ WC_MASK_FIELD(wc, nw_ttl);
+ WC_MASK_FIELD(wc, nw_proto);
+
+ /* No transport layer header in later fragments. */
+ if (!(flow->nw_frag & FLOW_NW_FRAG_LATER) &&
+ (flow->nw_proto == IPPROTO_ICMP ||
+ flow->nw_proto == IPPROTO_ICMPV6 ||
+ flow->nw_proto == IPPROTO_TCP ||
+ flow->nw_proto == IPPROTO_UDP ||
+ flow->nw_proto == IPPROTO_SCTP ||
+ flow->nw_proto == IPPROTO_IGMP)) {
+ WC_MASK_FIELD(wc, tp_src);
+ WC_MASK_FIELD(wc, tp_dst);
+
+ if (flow->nw_proto == IPPROTO_TCP) {
+ WC_MASK_FIELD(wc, tcp_flags);
+ } else if (flow->nw_proto == IPPROTO_ICMPV6) {
+ WC_MASK_FIELD(wc, arp_sha);
+ WC_MASK_FIELD(wc, arp_tha);
+ WC_MASK_FIELD(wc, nd_target);
+ } else if (flow->nw_proto == IPPROTO_IGMP) {
+ WC_MASK_FIELD(wc, igmp_group_ip4);
+ }
+ }
+}
+
+/* Return a map of possible fields for a packet of the same type as 'flow'.
+ * Including extra bits in the returned mask is not wrong, it is just less
+ * optimal.
+ *
+ * This is a less precise version of flow_wildcards_init_for_packet() above. */
+void
+flow_wc_map(const struct flow *flow, struct flowmap *map)
+{
+ /* Update this function whenever struct flow changes. */
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 35);
+
+ flowmap_init(map);
+
+ if (flow_tnl_dst_is_set(&flow->tunnel)) {
+ FLOWMAP_SET__(map, tunnel, offsetof(struct flow_tnl, metadata));
+ if (!(flow->tunnel.flags & FLOW_TNL_F_UDPIF)) {
+ if (flow->tunnel.metadata.present.map) {
+ FLOWMAP_SET(map, tunnel.metadata);
+ }
+ } else {
+ FLOWMAP_SET(map, tunnel.metadata.present.len);
+ FLOWMAP_SET__(map, tunnel.metadata.opts.gnv,
+ flow->tunnel.metadata.present.len);
+ }
+ }
+
+ /* Metadata fields that can appear on packet input. */
+ FLOWMAP_SET(map, skb_priority);
+ FLOWMAP_SET(map, pkt_mark);
+ FLOWMAP_SET(map, recirc_id);
+ FLOWMAP_SET(map, dp_hash);
+ FLOWMAP_SET(map, in_port);
+ FLOWMAP_SET(map, dl_dst);
+ FLOWMAP_SET(map, dl_src);
+ FLOWMAP_SET(map, dl_type);
+ FLOWMAP_SET(map, vlan_tci);
+ FLOWMAP_SET(map, ct_state);
+ FLOWMAP_SET(map, ct_zone);
+ FLOWMAP_SET(map, ct_mark);
+ FLOWMAP_SET(map, ct_label);
+
+ /* Ethertype-dependent fields. */
+ if (OVS_LIKELY(flow->dl_type == htons(ETH_TYPE_IP))) {
+ FLOWMAP_SET(map, nw_src);
+ FLOWMAP_SET(map, nw_dst);
+ FLOWMAP_SET(map, nw_proto);
+ FLOWMAP_SET(map, nw_frag);
+ FLOWMAP_SET(map, nw_tos);
+ FLOWMAP_SET(map, nw_ttl);
+
+ if (OVS_UNLIKELY(flow->nw_proto == IPPROTO_IGMP)) {
+ FLOWMAP_SET(map, igmp_group_ip4);
+ } else {
+ FLOWMAP_SET(map, tcp_flags);
+ FLOWMAP_SET(map, tp_src);
+ FLOWMAP_SET(map, tp_dst);
+ }
+ } else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
+ FLOWMAP_SET(map, ipv6_src);
+ FLOWMAP_SET(map, ipv6_dst);
+ FLOWMAP_SET(map, ipv6_label);
+ FLOWMAP_SET(map, nw_proto);
+ FLOWMAP_SET(map, nw_frag);
+ FLOWMAP_SET(map, nw_tos);
+ FLOWMAP_SET(map, nw_ttl);
+
+ if (OVS_UNLIKELY(flow->nw_proto == IPPROTO_ICMPV6)) {
+ FLOWMAP_SET(map, nd_target);
+ FLOWMAP_SET(map, arp_sha);
+ FLOWMAP_SET(map, arp_tha);
+ } else {
+ FLOWMAP_SET(map, tcp_flags);
+ FLOWMAP_SET(map, tp_src);
+ FLOWMAP_SET(map, tp_dst);
+ }
+ } else if (eth_type_mpls(flow->dl_type)) {
+ FLOWMAP_SET(map, mpls_lse);
+ } else if (flow->dl_type == htons(ETH_TYPE_ARP) ||
+ flow->dl_type == htons(ETH_TYPE_RARP)) {
+ FLOWMAP_SET(map, nw_src);
+ FLOWMAP_SET(map, nw_dst);
+ FLOWMAP_SET(map, nw_proto);
+ FLOWMAP_SET(map, arp_sha);
+ FLOWMAP_SET(map, arp_tha);
+ }
+}
+
/* Clear the metadata and register wildcard masks. They are not packet
* header fields. */
void
flow_wildcards_clear_non_packet_fields(struct flow_wildcards *wc)
{
+ /* Update this function whenever struct flow changes. */
+ BUILD_ASSERT_DECL(FLOW_WC_SEQ == 35);
+
memset(&wc->masks.metadata, 0, sizeof wc->masks.metadata);
memset(&wc->masks.regs, 0, sizeof wc->masks.regs);
+ wc->masks.actset_output = 0;
+ wc->masks.conj_id = 0;
}
/* Returns true if 'wc' matches every packet, false if 'wc' fixes any bits or
bool
flow_wildcards_is_catchall(const struct flow_wildcards *wc)
{
- const uint32_t *wc_u32 = (const uint32_t *) &wc->masks;
+ const uint64_t *wc_u64 = (const uint64_t *) &wc->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- if (wc_u32[i]) {
+ for (i = 0; i < FLOW_U64S; i++) {
+ if (wc_u64[i]) {
return false;
}
}
const struct flow_wildcards *src1,
const struct flow_wildcards *src2)
{
- uint32_t *dst_u32 = (uint32_t *) &dst->masks;
- const uint32_t *src1_u32 = (const uint32_t *) &src1->masks;
- const uint32_t *src2_u32 = (const uint32_t *) &src2->masks;
+ uint64_t *dst_u64 = (uint64_t *) &dst->masks;
+ const uint64_t *src1_u64 = (const uint64_t *) &src1->masks;
+ const uint64_t *src2_u64 = (const uint64_t *) &src2->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- dst_u32[i] = src1_u32[i] & src2_u32[i];
+ for (i = 0; i < FLOW_U64S; i++) {
+ dst_u64[i] = src1_u64[i] & src2_u64[i];
}
}
const struct flow_wildcards *src1,
const struct flow_wildcards *src2)
{
- uint32_t *dst_u32 = (uint32_t *) &dst->masks;
- const uint32_t *src1_u32 = (const uint32_t *) &src1->masks;
- const uint32_t *src2_u32 = (const uint32_t *) &src2->masks;
+ uint64_t *dst_u64 = (uint64_t *) &dst->masks;
+ const uint64_t *src1_u64 = (const uint64_t *) &src1->masks;
+ const uint64_t *src2_u64 = (const uint64_t *) &src2->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- dst_u32[i] = src1_u32[i] | src2_u32[i];
- }
-}
-
-/* Perform a bitwise OR of miniflow 'src' flow data with the equivalent
- * fields in 'dst', storing the result in 'dst'. */
-static void
-flow_union_with_miniflow(struct flow *dst, const struct miniflow *src)
-{
- uint32_t *dst_u32 = (uint32_t *) dst;
- const uint32_t *p = src->values;
- uint64_t map;
-
- for (map = src->map; map; map = zero_rightmost_1bit(map)) {
- dst_u32[raw_ctz(map)] |= *p++;
- }
-}
-
-/* Fold minimask 'mask''s wildcard mask into 'wc's wildcard mask. */
-void
-flow_wildcards_fold_minimask(struct flow_wildcards *wc,
- const struct minimask *mask)
-{
- flow_union_with_miniflow(&wc->masks, &mask->masks);
-}
-
-uint64_t
-miniflow_get_map_in_range(const struct miniflow *miniflow,
- uint8_t start, uint8_t end, unsigned int *offset)
-{
- uint64_t map = miniflow->map;
- *offset = 0;
-
- if (start > 0) {
- uint64_t msk = (UINT64_C(1) << start) - 1; /* 'start' LSBs set */
- *offset = count_1bits(map & msk);
- map &= ~msk;
- }
- if (end < FLOW_U32S) {
- uint64_t msk = (UINT64_C(1) << end) - 1; /* 'end' LSBs set */
- map &= msk;
- }
- return map;
-}
-
-/* Fold minimask 'mask''s wildcard mask into 'wc's wildcard mask
- * in range [start, end). */
-void
-flow_wildcards_fold_minimask_range(struct flow_wildcards *wc,
- const struct minimask *mask,
- uint8_t start, uint8_t end)
-{
- uint32_t *dst_u32 = (uint32_t *)&wc->masks;
- unsigned int offset;
- uint64_t map = miniflow_get_map_in_range(&mask->masks, start, end,
- &offset);
- const uint32_t *p = mask->masks.values + offset;
-
- for (; map; map = zero_rightmost_1bit(map)) {
- dst_u32[raw_ctz(map)] |= *p++;
+ for (i = 0; i < FLOW_U64S; i++) {
+ dst_u64[i] = src1_u64[i] | src2_u64[i];
}
}
flow_wildcards_has_extra(const struct flow_wildcards *a,
const struct flow_wildcards *b)
{
- const uint32_t *a_u32 = (const uint32_t *) &a->masks;
- const uint32_t *b_u32 = (const uint32_t *) &b->masks;
+ const uint64_t *a_u64 = (const uint64_t *) &a->masks;
+ const uint64_t *b_u64 = (const uint64_t *) &b->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- if ((a_u32[i] & b_u32[i]) != b_u32[i]) {
+ for (i = 0; i < FLOW_U64S; i++) {
+ if ((a_u64[i] & b_u64[i]) != b_u64[i]) {
return true;
}
}
flow_equal_except(const struct flow *a, const struct flow *b,
const struct flow_wildcards *wc)
{
- const uint32_t *a_u32 = (const uint32_t *) a;
- const uint32_t *b_u32 = (const uint32_t *) b;
- const uint32_t *wc_u32 = (const uint32_t *) &wc->masks;
+ const uint64_t *a_u64 = (const uint64_t *) a;
+ const uint64_t *b_u64 = (const uint64_t *) b;
+ const uint64_t *wc_u64 = (const uint64_t *) &wc->masks;
size_t i;
- for (i = 0; i < FLOW_U32S; i++) {
- if ((a_u32[i] ^ b_u32[i]) & wc_u32[i]) {
+ for (i = 0; i < FLOW_U64S; i++) {
+ if ((a_u64[i] ^ b_u64[i]) & wc_u64[i]) {
return false;
}
}
wc->masks.regs[idx] = mask;
}
-/* Calculates the 5-tuple hash from the given flow. */
+/* Sets the wildcard mask for register 'idx' in 'wc' to 'mask'.
+ * (A 0-bit indicates a wildcard bit.) */
+void
+flow_wildcards_set_xreg_mask(struct flow_wildcards *wc, int idx, uint64_t mask)
+{
+ flow_set_xreg(&wc->masks, idx, mask);
+}
+
+/* Calculates the 5-tuple hash from the given miniflow.
+ * This returns the same value as flow_hash_5tuple for the corresponding
+ * flow. */
uint32_t
miniflow_hash_5tuple(const struct miniflow *flow, uint32_t basis)
{
- uint32_t hash = 0;
+ uint32_t hash = basis;
- if (!flow) {
- return 0;
- }
+ if (flow) {
+ ovs_be16 dl_type = MINIFLOW_GET_BE16(flow, dl_type);
- hash = mhash_add(basis,
- miniflow_get_u32(flow, offsetof(struct flow, nw_src)));
- hash = mhash_add(hash,
- miniflow_get_u32(flow, offsetof(struct flow, nw_dst)));
- hash = mhash_add(hash,
- miniflow_get_u32(flow, offsetof(struct flow, tp_src)));
- hash = mhash_add(hash,
- miniflow_get_u8(flow, offsetof(struct flow, nw_proto)));
+ hash = hash_add(hash, MINIFLOW_GET_U8(flow, nw_proto));
- return mhash_finish(hash, 13);
+ /* Separate loops for better optimization. */
+ if (dl_type == htons(ETH_TYPE_IPV6)) {
+ struct flowmap map = FLOWMAP_EMPTY_INITIALIZER;
+ uint64_t value;
+
+ FLOWMAP_SET(&map, ipv6_src);
+ FLOWMAP_SET(&map, ipv6_dst);
+
+ MINIFLOW_FOR_EACH_IN_FLOWMAP(value, flow, map) {
+ hash = hash_add64(hash, value);
+ }
+ } else {
+ hash = hash_add(hash, MINIFLOW_GET_U32(flow, nw_src));
+ hash = hash_add(hash, MINIFLOW_GET_U32(flow, nw_dst));
+ }
+ /* Add both ports at once. */
+ hash = hash_add(hash, MINIFLOW_GET_U32(flow, tp_src));
+ hash = hash_finish(hash, 42); /* Arbitrary number. */
+ }
+ return hash;
}
-BUILD_ASSERT_DECL(offsetof(struct flow, tp_src) + 2
- == offsetof(struct flow, tp_dst) &&
- offsetof(struct flow, tp_src) / 4
- == offsetof(struct flow, tp_dst) / 4);
+ASSERT_SEQUENTIAL_SAME_WORD(tp_src, tp_dst);
+ASSERT_SEQUENTIAL(ipv6_src, ipv6_dst);
/* Calculates the 5-tuple hash from the given flow. */
uint32_t
flow_hash_5tuple(const struct flow *flow, uint32_t basis)
{
- const uint32_t *flow_u32 = (const uint32_t *)flow;
- uint32_t hash = 0;
+ uint32_t hash = basis;
- if (!flow) {
- return 0;
- }
+ if (flow) {
+ hash = hash_add(hash, flow->nw_proto);
- hash = mhash_add(basis, (OVS_FORCE uint32_t) flow->nw_src);
- hash = mhash_add(hash, (OVS_FORCE uint32_t) flow->nw_dst);
- hash = mhash_add(hash, flow_u32[offsetof(struct flow, tp_src) / 4]);
- hash = mhash_add(hash, flow->nw_proto);
+ if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
+ const uint64_t *flow_u64 = (const uint64_t *)flow;
+ int ofs = offsetof(struct flow, ipv6_src) / 8;
+ int end = ofs + 2 * sizeof flow->ipv6_src / 8;
- return mhash_finish(hash, 13);
+ for (;ofs < end; ofs++) {
+ hash = hash_add64(hash, flow_u64[ofs]);
+ }
+ } else {
+ hash = hash_add(hash, (OVS_FORCE uint32_t) flow->nw_src);
+ hash = hash_add(hash, (OVS_FORCE uint32_t) flow->nw_dst);
+ }
+ /* Add both ports at once. */
+ hash = hash_add(hash,
+ ((const uint32_t *)flow)[offsetof(struct flow, tp_src)
+ / sizeof(uint32_t)]);
+ hash = hash_finish(hash, 42); /* Arbitrary number. */
+ }
+ return hash;
}
/* Hashes 'flow' based on its L2 through L4 protocol information. */
ovs_be16 eth_type;
ovs_be16 vlan_tci;
ovs_be16 tp_port;
- uint8_t eth_addr[ETH_ADDR_LEN];
+ struct eth_addr eth_addr;
uint8_t ip_proto;
} fields;
int i;
memset(&fields, 0, sizeof fields);
- for (i = 0; i < ETH_ADDR_LEN; i++) {
- fields.eth_addr[i] = flow->dl_src[i] ^ flow->dl_dst[i];
+ for (i = 0; i < ARRAY_SIZE(fields.eth_addr.be16); i++) {
+ fields.eth_addr.be16[i] = flow->dl_src.be16[i] ^ flow->dl_dst.be16[i];
}
fields.vlan_tci = flow->vlan_tci & htons(VLAN_VID_MASK);
fields.eth_type = flow->dl_type;
return jhash_bytes(&fields, sizeof fields, basis);
}
+/* Hashes 'flow' based on its L3 through L4 protocol information */
+uint32_t
+flow_hash_symmetric_l3l4(const struct flow *flow, uint32_t basis,
+ bool inc_udp_ports)
+{
+ uint32_t hash = basis;
+
+ /* UDP source and destination port are also taken into account. */
+ if (flow->dl_type == htons(ETH_TYPE_IP)) {
+ hash = hash_add(hash,
+ (OVS_FORCE uint32_t) (flow->nw_src ^ flow->nw_dst));
+ } else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
+ /* IPv6 addresses are 64-bit aligned inside struct flow. */
+ const uint64_t *a = ALIGNED_CAST(uint64_t *, flow->ipv6_src.s6_addr);
+ const uint64_t *b = ALIGNED_CAST(uint64_t *, flow->ipv6_dst.s6_addr);
+
+ for (int i = 0; i < 4; i++) {
+ hash = hash_add64(hash, a[i] ^ b[i]);
+ }
+ } else {
+ /* Cannot hash non-IP flows */
+ return 0;
+ }
+
+ hash = hash_add(hash, flow->nw_proto);
+ if (flow->nw_proto == IPPROTO_TCP || flow->nw_proto == IPPROTO_SCTP ||
+ (inc_udp_ports && flow->nw_proto == IPPROTO_UDP)) {
+ hash = hash_add(hash,
+ (OVS_FORCE uint16_t) (flow->tp_src ^ flow->tp_dst));
+ }
+
+ return hash_finish(hash, basis);
+}
+
/* Initialize a flow with random fields that matter for nx_hash_fields. */
void
flow_random_hash_fields(struct flow *flow)
/* Initialize to all zeros. */
memset(flow, 0, sizeof *flow);
- eth_addr_random(flow->dl_src);
- eth_addr_random(flow->dl_dst);
+ eth_addr_random(&flow->dl_src);
+ eth_addr_random(&flow->dl_dst);
flow->vlan_tci = (OVS_FORCE ovs_be16) (random_uint16() & VLAN_VID_MASK);
wc->masks.vlan_tci |= htons(VLAN_VID_MASK | VLAN_CFI);
break;
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4_UDP:
+ if (is_ip_any(flow) && flow->nw_proto == IPPROTO_UDP) {
+ memset(&wc->masks.tp_src, 0xff, sizeof wc->masks.tp_src);
+ memset(&wc->masks.tp_dst, 0xff, sizeof wc->masks.tp_dst);
+ }
+ /* no break */
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4:
+ if (flow->dl_type == htons(ETH_TYPE_IP)) {
+ memset(&wc->masks.nw_src, 0xff, sizeof wc->masks.nw_src);
+ memset(&wc->masks.nw_dst, 0xff, sizeof wc->masks.nw_dst);
+ } else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
+ memset(&wc->masks.ipv6_src, 0xff, sizeof wc->masks.ipv6_src);
+ memset(&wc->masks.ipv6_dst, 0xff, sizeof wc->masks.ipv6_dst);
+ } else {
+ break; /* non-IP flow */
+ }
+
+ memset(&wc->masks.nw_proto, 0xff, sizeof wc->masks.nw_proto);
+ if (flow->nw_proto == IPPROTO_TCP || flow->nw_proto == IPPROTO_SCTP) {
+ memset(&wc->masks.tp_src, 0xff, sizeof wc->masks.tp_src);
+ memset(&wc->masks.tp_dst, 0xff, sizeof wc->masks.tp_dst);
+ }
+ break;
+
default:
OVS_NOT_REACHED();
}
switch (fields) {
case NX_HASH_FIELDS_ETH_SRC:
- return jhash_bytes(flow->dl_src, sizeof flow->dl_src, basis);
+ return jhash_bytes(&flow->dl_src, sizeof flow->dl_src, basis);
case NX_HASH_FIELDS_SYMMETRIC_L4:
return flow_hash_symmetric_l4(flow, basis);
+
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4:
+ return flow_hash_symmetric_l3l4(flow, basis, false);
+
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4_UDP:
+ return flow_hash_symmetric_l3l4(flow, basis, true);
+
}
OVS_NOT_REACHED();
switch (fields) {
case NX_HASH_FIELDS_ETH_SRC: return "eth_src";
case NX_HASH_FIELDS_SYMMETRIC_L4: return "symmetric_l4";
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4: return "symmetric_l3l4";
+ case NX_HASH_FIELDS_SYMMETRIC_L3L4_UDP: return "symmetric_l3l4+udp";
default: return "<unknown>";
}
}
flow_hash_fields_valid(enum nx_hash_fields fields)
{
return fields == NX_HASH_FIELDS_ETH_SRC
- || fields == NX_HASH_FIELDS_SYMMETRIC_L4;
+ || fields == NX_HASH_FIELDS_SYMMETRIC_L4
+ || fields == NX_HASH_FIELDS_SYMMETRIC_L3L4
+ || fields == NX_HASH_FIELDS_SYMMETRIC_L3L4_UDP;
}
/* Returns a hash value for the bits of 'flow' that are active based on
flow_hash_in_wildcards(const struct flow *flow,
const struct flow_wildcards *wc, uint32_t basis)
{
- const uint32_t *wc_u32 = (const uint32_t *) &wc->masks;
- const uint32_t *flow_u32 = (const uint32_t *) flow;
+ const uint64_t *wc_u64 = (const uint64_t *) &wc->masks;
+ const uint64_t *flow_u64 = (const uint64_t *) flow;
uint32_t hash;
size_t i;
hash = basis;
- for (i = 0; i < FLOW_U32S; i++) {
- hash = mhash_add(hash, flow_u32[i] & wc_u32[i]);
+ for (i = 0; i < FLOW_U64S; i++) {
+ hash = hash_add64(hash, flow_u64[i] & wc_u64[i]);
}
- return mhash_finish(hash, 4 * FLOW_U32S);
+ return hash_finish(hash, 8 * FLOW_U64S);
}
/* Sets the VLAN VID that 'flow' matches to 'vid', which is interpreted as an
int
flow_count_mpls_labels(const struct flow *flow, struct flow_wildcards *wc)
{
- if (wc) {
- wc->masks.dl_type = OVS_BE16_MAX;
- }
+ /* dl_type is always masked. */
if (eth_type_mpls(flow->dl_type)) {
int i;
- int len = FLOW_MAX_MPLS_LABELS;
+ int cnt;
- for (i = 0; i < len; i++) {
+ cnt = 0;
+ for (i = 0; i < FLOW_MAX_MPLS_LABELS; i++) {
if (wc) {
wc->masks.mpls_lse[i] |= htonl(MPLS_BOS_MASK);
}
if (flow->mpls_lse[i] & htonl(MPLS_BOS_MASK)) {
return i + 1;
}
+ if (flow->mpls_lse[i]) {
+ cnt++;
+ }
}
-
- return len;
+ return cnt;
} else {
return 0;
}
*
* - BoS: 1.
*
- * If the new label is the second or label MPLS label in 'flow', it is
+ * If the new label is the second or later label MPLS label in 'flow', it is
* generated as;
*
* - label: Copied from outer label.
ovs_assert(eth_type_mpls(mpls_eth_type));
ovs_assert(n < FLOW_MAX_MPLS_LABELS);
- memset(wc->masks.mpls_lse, 0xff, sizeof wc->masks.mpls_lse);
if (n) {
int i;
+ if (wc) {
+ memset(&wc->masks.mpls_lse, 0xff, sizeof *wc->masks.mpls_lse * n);
+ }
for (i = n; i >= 1; i--) {
flow->mpls_lse[i] = flow->mpls_lse[i - 1];
}
- flow->mpls_lse[0] = (flow->mpls_lse[1]
- & htonl(~MPLS_BOS_MASK));
+ flow->mpls_lse[0] = (flow->mpls_lse[1] & htonl(~MPLS_BOS_MASK));
} else {
int label = 0; /* IPv4 Explicit Null. */
int tc = 0;
if (is_ip_any(flow)) {
tc = (flow->nw_tos & IP_DSCP_MASK) >> 2;
- wc->masks.nw_tos |= IP_DSCP_MASK;
+ if (wc) {
+ wc->masks.nw_tos |= IP_DSCP_MASK;
+ wc->masks.nw_ttl = 0xff;
+ }
if (flow->nw_ttl) {
ttl = flow->nw_ttl;
}
- wc->masks.nw_ttl = 0xff;
}
flow->mpls_lse[0] = set_mpls_lse_values(ttl, tc, 1, htonl(label));
- /* Clear all L3 and L4 fields. */
- BUILD_ASSERT(FLOW_WC_SEQ == 26);
+ /* Clear all L3 and L4 fields and dp_hash. */
+ BUILD_ASSERT(FLOW_WC_SEQ == 35);
memset((char *) flow + FLOW_SEGMENT_2_ENDS_AT, 0,
sizeof(struct flow) - FLOW_SEGMENT_2_ENDS_AT);
+ flow->dp_hash = 0;
}
flow->dl_type = mpls_eth_type;
}
if (n == 0) {
/* Nothing to pop. */
return false;
- } else if (n == FLOW_MAX_MPLS_LABELS
- && !(flow->mpls_lse[n - 1] & htonl(MPLS_BOS_MASK))) {
- /* Can't pop because we don't know what to fill in mpls_lse[n - 1]. */
- return false;
+ } else if (n == FLOW_MAX_MPLS_LABELS) {
+ if (wc) {
+ wc->masks.mpls_lse[n - 1] |= htonl(MPLS_BOS_MASK);
+ }
+ if (!(flow->mpls_lse[n - 1] & htonl(MPLS_BOS_MASK))) {
+ /* Can't pop because don't know what to fill in mpls_lse[n - 1]. */
+ return false;
+ }
}
- memset(wc->masks.mpls_lse, 0xff, sizeof wc->masks.mpls_lse);
+ if (wc) {
+ memset(&wc->masks.mpls_lse[1], 0xff,
+ sizeof *wc->masks.mpls_lse * (n - 1));
+ }
for (i = 1; i < n; i++) {
flow->mpls_lse[i - 1] = flow->mpls_lse[i];
}
}
static size_t
-flow_compose_l4(struct ofpbuf *b, const struct flow *flow)
+flow_compose_l4(struct dp_packet *p, const struct flow *flow)
{
size_t l4_len = 0;
struct tcp_header *tcp;
l4_len = sizeof *tcp;
- tcp = ofpbuf_put_zeros(b, l4_len);
+ tcp = dp_packet_put_zeros(p, l4_len);
tcp->tcp_src = flow->tp_src;
tcp->tcp_dst = flow->tp_dst;
tcp->tcp_ctl = TCP_CTL(ntohs(flow->tcp_flags), 5);
struct udp_header *udp;
l4_len = sizeof *udp;
- udp = ofpbuf_put_zeros(b, l4_len);
+ udp = dp_packet_put_zeros(p, l4_len);
udp->udp_src = flow->tp_src;
udp->udp_dst = flow->tp_dst;
} else if (flow->nw_proto == IPPROTO_SCTP) {
struct sctp_header *sctp;
l4_len = sizeof *sctp;
- sctp = ofpbuf_put_zeros(b, l4_len);
+ sctp = dp_packet_put_zeros(p, l4_len);
sctp->sctp_src = flow->tp_src;
sctp->sctp_dst = flow->tp_dst;
} else if (flow->nw_proto == IPPROTO_ICMP) {
struct icmp_header *icmp;
l4_len = sizeof *icmp;
- icmp = ofpbuf_put_zeros(b, l4_len);
+ icmp = dp_packet_put_zeros(p, l4_len);
icmp->icmp_type = ntohs(flow->tp_src);
icmp->icmp_code = ntohs(flow->tp_dst);
icmp->icmp_csum = csum(icmp, ICMP_HEADER_LEN);
+ } else if (flow->nw_proto == IPPROTO_IGMP) {
+ struct igmp_header *igmp;
+
+ l4_len = sizeof *igmp;
+ igmp = dp_packet_put_zeros(p, l4_len);
+ igmp->igmp_type = ntohs(flow->tp_src);
+ igmp->igmp_code = ntohs(flow->tp_dst);
+ put_16aligned_be32(&igmp->group, flow->igmp_group_ip4);
+ igmp->igmp_csum = csum(igmp, IGMP_HEADER_LEN);
} else if (flow->nw_proto == IPPROTO_ICMPV6) {
struct icmp6_hdr *icmp;
l4_len = sizeof *icmp;
- icmp = ofpbuf_put_zeros(b, l4_len);
+ icmp = dp_packet_put_zeros(p, l4_len);
icmp->icmp6_type = ntohs(flow->tp_src);
icmp->icmp6_code = ntohs(flow->tp_dst);
(icmp->icmp6_type == ND_NEIGHBOR_SOLICIT ||
icmp->icmp6_type == ND_NEIGHBOR_ADVERT)) {
struct in6_addr *nd_target;
- struct nd_opt_hdr *nd_opt;
+ struct ovs_nd_opt *nd_opt;
l4_len += sizeof *nd_target;
- nd_target = ofpbuf_put_zeros(b, sizeof *nd_target);
+ nd_target = dp_packet_put_zeros(p, sizeof *nd_target);
*nd_target = flow->nd_target;
if (!eth_addr_is_zero(flow->arp_sha)) {
l4_len += 8;
- nd_opt = ofpbuf_put_zeros(b, 8);
+ nd_opt = dp_packet_put_zeros(p, 8);
nd_opt->nd_opt_len = 1;
nd_opt->nd_opt_type = ND_OPT_SOURCE_LINKADDR;
- memcpy(nd_opt + 1, flow->arp_sha, ETH_ADDR_LEN);
+ nd_opt->nd_opt_mac = flow->arp_sha;
}
if (!eth_addr_is_zero(flow->arp_tha)) {
l4_len += 8;
- nd_opt = ofpbuf_put_zeros(b, 8);
+ nd_opt = dp_packet_put_zeros(p, 8);
nd_opt->nd_opt_len = 1;
nd_opt->nd_opt_type = ND_OPT_TARGET_LINKADDR;
- memcpy(nd_opt + 1, flow->arp_tha, ETH_ADDR_LEN);
+ nd_opt->nd_opt_mac = flow->arp_tha;
}
}
icmp->icmp6_cksum = (OVS_FORCE uint16_t)
- csum(icmp, (char *)ofpbuf_tail(b) - (char *)icmp);
+ csum(icmp, (char *)dp_packet_tail(p) - (char *)icmp);
}
}
return l4_len;
* valid. It hasn't got some checksums filled in, for one, and lots of fields
* are just zeroed.) */
void
-flow_compose(struct ofpbuf *b, const struct flow *flow)
+flow_compose(struct dp_packet *p, const struct flow *flow)
{
size_t l4_len;
/* eth_compose() sets l3 pointer and makes sure it is 32-bit aligned. */
- eth_compose(b, flow->dl_dst, flow->dl_src, ntohs(flow->dl_type), 0);
+ eth_compose(p, flow->dl_dst, flow->dl_src, ntohs(flow->dl_type), 0);
if (flow->dl_type == htons(FLOW_DL_TYPE_NONE)) {
- struct eth_header *eth = ofpbuf_l2(b);
- eth->eth_type = htons(ofpbuf_size(b));
+ struct eth_header *eth = dp_packet_l2(p);
+ eth->eth_type = htons(dp_packet_size(p));
return;
}
if (flow->vlan_tci & htons(VLAN_CFI)) {
- eth_push_vlan(b, htons(ETH_TYPE_VLAN), flow->vlan_tci);
+ eth_push_vlan(p, htons(ETH_TYPE_VLAN), flow->vlan_tci);
}
if (flow->dl_type == htons(ETH_TYPE_IP)) {
struct ip_header *ip;
- ip = ofpbuf_put_zeros(b, sizeof *ip);
+ ip = dp_packet_put_zeros(p, sizeof *ip);
ip->ip_ihl_ver = IP_IHL_VER(5, 4);
ip->ip_tos = flow->nw_tos;
ip->ip_ttl = flow->nw_ttl;
}
}
- ofpbuf_set_l4(b, ofpbuf_tail(b));
+ dp_packet_set_l4(p, dp_packet_tail(p));
- l4_len = flow_compose_l4(b, flow);
+ l4_len = flow_compose_l4(p, flow);
- ip->ip_tot_len = htons(b->l4_ofs - b->l3_ofs + l4_len);
+ ip = dp_packet_l3(p);
+ ip->ip_tot_len = htons(p->l4_ofs - p->l3_ofs + l4_len);
ip->ip_csum = csum(ip, sizeof *ip);
} else if (flow->dl_type == htons(ETH_TYPE_IPV6)) {
struct ovs_16aligned_ip6_hdr *nh;
- nh = ofpbuf_put_zeros(b, sizeof *nh);
+ nh = dp_packet_put_zeros(p, sizeof *nh);
put_16aligned_be32(&nh->ip6_flow, htonl(6 << 28) |
htonl(flow->nw_tos << 20) | flow->ipv6_label);
nh->ip6_hlim = flow->nw_ttl;
memcpy(&nh->ip6_src, &flow->ipv6_src, sizeof(nh->ip6_src));
memcpy(&nh->ip6_dst, &flow->ipv6_dst, sizeof(nh->ip6_dst));
- ofpbuf_set_l4(b, ofpbuf_tail(b));
+ dp_packet_set_l4(p, dp_packet_tail(p));
- l4_len = flow_compose_l4(b, flow);
+ l4_len = flow_compose_l4(p, flow);
+ nh = dp_packet_l3(p);
nh->ip6_plen = htons(l4_len);
} else if (flow->dl_type == htons(ETH_TYPE_ARP) ||
flow->dl_type == htons(ETH_TYPE_RARP)) {
struct arp_eth_header *arp;
- arp = ofpbuf_put_zeros(b, sizeof *arp);
- ofpbuf_set_l3(b, arp);
+ arp = dp_packet_put_zeros(p, sizeof *arp);
+ dp_packet_set_l3(p, arp);
arp->ar_hrd = htons(1);
arp->ar_pro = htons(ETH_TYPE_IP);
arp->ar_hln = ETH_ADDR_LEN;
flow->nw_proto == ARP_OP_REPLY) {
put_16aligned_be32(&arp->ar_spa, flow->nw_src);
put_16aligned_be32(&arp->ar_tpa, flow->nw_dst);
- memcpy(arp->ar_sha, flow->arp_sha, ETH_ADDR_LEN);
- memcpy(arp->ar_tha, flow->arp_tha, ETH_ADDR_LEN);
+ arp->ar_sha = flow->arp_sha;
+ arp->ar_tha = flow->arp_tha;
}
}
if (eth_type_mpls(flow->dl_type)) {
int n;
- b->l2_5_ofs = b->l3_ofs;
+ p->l2_5_ofs = p->l3_ofs;
for (n = 1; n < FLOW_MAX_MPLS_LABELS; n++) {
if (flow->mpls_lse[n - 1] & htonl(MPLS_BOS_MASK)) {
break;
}
}
while (n > 0) {
- push_mpls(b, flow->dl_type, flow->mpls_lse[--n]);
+ push_mpls(p, flow->dl_type, flow->mpls_lse[--n]);
}
}
}
\f
/* Compressed flow. */
-static int
-miniflow_n_values(const struct miniflow *flow)
-{
- return count_1bits(flow->map);
-}
-
-static uint32_t *
-miniflow_alloc_values(struct miniflow *flow, int n)
-{
- if (n <= MINI_N_INLINE) {
- return flow->inline_values;
- } else {
- COVERAGE_INC(miniflow_malloc);
- return xmalloc(n * sizeof *flow->values);
- }
-}
-
/* Completes an initialization of 'dst' as a miniflow copy of 'src' begun by
- * the caller. The caller must have already initialized 'dst->map' properly
- * to indicate the significant uint32_t elements of 'src'. 'n' must be the
- * number of 1-bits in 'dst->map'.
+ * the caller. The caller must have already computed 'dst->map' properly to
+ * indicate the significant uint64_t elements of 'src'.
*
* Normally the significant elements are the ones that are non-zero. However,
* when a miniflow is initialized from a (mini)mask, the values can be zeroes,
- * so that the flow and mask always have the same maps.
- *
- * This function initializes 'dst->values' (either inline if possible or with
- * malloc() otherwise) and copies the uint32_t elements of 'src' indicated by
- * 'dst->map' into it. */
-static void
-miniflow_init__(struct miniflow *dst, const struct flow *src, int n)
+ * so that the flow and mask always have the same maps. */
+void
+miniflow_init(struct miniflow *dst, const struct flow *src)
{
- const uint32_t *src_u32 = (const uint32_t *) src;
- unsigned int ofs;
- uint64_t map;
+ uint64_t *dst_u64 = miniflow_values(dst);
+ size_t idx;
- dst->values = miniflow_alloc_values(dst, n);
- ofs = 0;
- for (map = dst->map; map; map = zero_rightmost_1bit(map)) {
- dst->values[ofs++] = src_u32[raw_ctz(map)];
+ FLOWMAP_FOR_EACH_INDEX(idx, dst->map) {
+ *dst_u64++ = flow_u64_value(src, idx);
}
}
-/* Initializes 'dst' as a copy of 'src'. The caller must eventually free 'dst'
- * with miniflow_destroy(). */
+/* Initialize the maps of 'flow' from 'src'. */
void
-miniflow_init(struct miniflow *dst, const struct flow *src)
+miniflow_map_init(struct miniflow *flow, const struct flow *src)
{
- const uint32_t *src_u32 = (const uint32_t *) src;
- unsigned int i;
- int n;
-
- /* Initialize dst->map, counting the number of nonzero elements. */
- n = 0;
- dst->map = 0;
-
- for (i = 0; i < FLOW_U32S; i++) {
- if (src_u32[i]) {
- dst->map |= UINT64_C(1) << i;
- n++;
+ /* Initialize map, counting the number of nonzero elements. */
+ flowmap_init(&flow->map);
+ for (size_t i = 0; i < FLOW_U64S; i++) {
+ if (flow_u64_value(src, i)) {
+ flowmap_set(&flow->map, i, 1);
}
}
-
- miniflow_init__(dst, src, n);
}
-/* Initializes 'dst' as a copy of 'src', using 'mask->map' as 'dst''s map. The
- * caller must eventually free 'dst' with miniflow_destroy(). */
-void
-miniflow_init_with_minimask(struct miniflow *dst, const struct flow *src,
- const struct minimask *mask)
+/* Allocates 'n' count of miniflows, consecutive in memory, initializing the
+ * map of each from 'src'.
+ * Returns the size of the miniflow data. */
+size_t
+miniflow_alloc(struct miniflow *dsts[], size_t n, const struct miniflow *src)
{
- dst->map = mask->masks.map;
- miniflow_init__(dst, src, miniflow_n_values(dst));
-}
+ size_t n_values = miniflow_n_values(src);
+ size_t data_size = MINIFLOW_VALUES_SIZE(n_values);
+ struct miniflow *dst = xmalloc(n * (sizeof *src + data_size));
+ size_t i;
-/* Initializes 'dst' as a copy of 'src'. The caller must eventually free 'dst'
- * with miniflow_destroy(). */
-void
-miniflow_clone(struct miniflow *dst, const struct miniflow *src)
-{
- int n = miniflow_n_values(src);
- dst->map = src->map;
- dst->values = miniflow_alloc_values(dst, n);
- memcpy(dst->values, src->values, n * sizeof *dst->values);
+ COVERAGE_INC(miniflow_malloc);
+
+ for (i = 0; i < n; i++) {
+ *dst = *src; /* Copy maps. */
+ dsts[i] = dst;
+ dst += 1; /* Just past the maps. */
+ dst = (struct miniflow *)((uint64_t *)dst + n_values); /* Skip data. */
+ }
+ return data_size;
}
-/* Initializes 'dst' with the data in 'src', destroying 'src'.
- * The caller must eventually free 'dst' with miniflow_destroy(). */
-void
-miniflow_move(struct miniflow *dst, struct miniflow *src)
+/* Returns a miniflow copy of 'src'. The caller must eventually free() the
+ * returned miniflow. */
+struct miniflow *
+miniflow_create(const struct flow *src)
{
- if (src->values == src->inline_values) {
- dst->values = dst->inline_values;
- memcpy(dst->values, src->values,
- miniflow_n_values(src) * sizeof *dst->values);
- } else {
- dst->values = src->values;
- }
- dst->map = src->map;
+ struct miniflow tmp;
+ struct miniflow *dst;
+
+ miniflow_map_init(&tmp, src);
+
+ miniflow_alloc(&dst, 1, &tmp);
+ miniflow_init(dst, src);
+ return dst;
}
-/* Frees any memory owned by 'flow'. Does not free the storage in which 'flow'
- * itself resides; the caller is responsible for that. */
+/* Initializes 'dst' as a copy of 'src'. The caller must have allocated
+ * 'dst' to have inline space for 'n_values' data in 'src'. */
void
-miniflow_destroy(struct miniflow *flow)
+miniflow_clone(struct miniflow *dst, const struct miniflow *src,
+ size_t n_values)
{
- if (flow->values != flow->inline_values) {
- free(flow->values);
- }
+ *dst = *src; /* Copy maps. */
+ memcpy(miniflow_values(dst), miniflow_get_values(src),
+ MINIFLOW_VALUES_SIZE(n_values));
}
/* Initializes 'dst' as a copy of 'src'. */
flow_union_with_miniflow(dst, src);
}
-/* Returns the uint32_t that would be at byte offset '4 * u32_ofs' if 'flow'
- * were expanded into a "struct flow". */
-static uint32_t
-miniflow_get(const struct miniflow *flow, unsigned int u32_ofs)
-{
- return (flow->map & UINT64_C(1) << u32_ofs)
- ? *(flow->values +
- count_1bits(flow->map & ((UINT64_C(1) << u32_ofs) - 1)))
- : 0;
-}
-
-/* Returns true if 'a' and 'b' are the same flow, false otherwise. */
+/* Returns true if 'a' and 'b' are equal miniflows, false otherwise. */
bool
miniflow_equal(const struct miniflow *a, const struct miniflow *b)
{
- const uint32_t *ap = a->values;
- const uint32_t *bp = b->values;
- const uint64_t a_map = a->map;
- const uint64_t b_map = b->map;
- uint64_t map;
+ const uint64_t *ap = miniflow_get_values(a);
+ const uint64_t *bp = miniflow_get_values(b);
- if (a_map == b_map) {
- for (map = a_map; map; map = zero_rightmost_1bit(map)) {
- if (*ap++ != *bp++) {
- return false;
- }
- }
+ /* This is mostly called after a matching hash, so it is highly likely that
+ * the maps are equal as well. */
+ if (OVS_LIKELY(flowmap_equal(a->map, b->map))) {
+ return !memcmp(ap, bp, miniflow_n_values(a) * sizeof *ap);
} else {
- for (map = a_map | b_map; map; map = zero_rightmost_1bit(map)) {
- uint64_t bit = rightmost_1bit(map);
- uint64_t a_value = a_map & bit ? *ap++ : 0;
- uint64_t b_value = b_map & bit ? *bp++ : 0;
+ size_t idx;
- if (a_value != b_value) {
+ FLOWMAP_FOR_EACH_INDEX (idx, flowmap_or(a->map, b->map)) {
+ if ((flowmap_is_set(&a->map, idx) ? *ap++ : 0)
+ != (flowmap_is_set(&b->map, idx) ? *bp++ : 0)) {
return false;
}
}
return true;
}
-/* Returns true if 'a' and 'b' are equal at the places where there are 1-bits
- * in 'mask', false if they differ. */
+/* Returns false if 'a' and 'b' differ at the places where there are 1-bits
+ * in 'mask', true otherwise. */
bool
miniflow_equal_in_minimask(const struct miniflow *a, const struct miniflow *b,
const struct minimask *mask)
{
- const uint32_t *p;
- uint64_t map;
-
- p = mask->masks.values;
+ const uint64_t *p = miniflow_get_values(&mask->masks);
+ size_t idx;
- for (map = mask->masks.map; map; map = zero_rightmost_1bit(map)) {
- int ofs = raw_ctz(map);
-
- if ((miniflow_get(a, ofs) ^ miniflow_get(b, ofs)) & *p) {
+ FLOWMAP_FOR_EACH_INDEX(idx, mask->masks.map) {
+ if ((miniflow_get(a, idx) ^ miniflow_get(b, idx)) & *p++) {
return false;
}
- p++;
}
return true;
miniflow_equal_flow_in_minimask(const struct miniflow *a, const struct flow *b,
const struct minimask *mask)
{
- const uint32_t *b_u32 = (const uint32_t *) b;
- const uint32_t *p;
- uint64_t map;
-
- p = mask->masks.values;
-
- for (map = mask->masks.map; map; map = zero_rightmost_1bit(map)) {
- int ofs = raw_ctz(map);
+ const uint64_t *p = miniflow_get_values(&mask->masks);
+ size_t idx;
- if ((miniflow_get(a, ofs) ^ b_u32[ofs]) & *p) {
+ FLOWMAP_FOR_EACH_INDEX(idx, mask->masks.map) {
+ if ((miniflow_get(a, idx) ^ flow_u64_value(b, idx)) & *p++) {
return false;
}
- p++;
}
return true;
}
-/* Returns a hash value for 'flow', given 'basis'. */
-uint32_t
-miniflow_hash(const struct miniflow *flow, uint32_t basis)
-{
- const uint32_t *p = flow->values;
- uint32_t hash = basis;
- uint64_t hash_map = 0;
- uint64_t map;
-
- for (map = flow->map; map; map = zero_rightmost_1bit(map)) {
- if (*p) {
- hash = mhash_add(hash, *p);
- hash_map |= rightmost_1bit(map);
- }
- p++;
- }
- hash = mhash_add(hash, hash_map);
- hash = mhash_add(hash, hash_map >> 32);
-
- return mhash_finish(hash, p - flow->values);
-}
-
-/* Returns a hash value for the bits of 'flow' where there are 1-bits in
- * 'mask', given 'basis'.
- *
- * The hash values returned by this function are the same as those returned by
- * flow_hash_in_minimask(), only the form of the arguments differ. */
-uint32_t
-miniflow_hash_in_minimask(const struct miniflow *flow,
- const struct minimask *mask, uint32_t basis)
-{
- const uint32_t *p = mask->masks.values;
- uint32_t hash = basis;
- uint32_t flow_u32;
-
- MINIFLOW_FOR_EACH_IN_MAP(flow_u32, flow, mask->masks.map) {
- hash = mhash_add(hash, flow_u32 & *p++);
- }
-
- return mhash_finish(hash, (p - mask->masks.values) * 4);
-}
-
-/* Returns a hash value for the bits of 'flow' where there are 1-bits in
- * 'mask', given 'basis'.
- *
- * The hash values returned by this function are the same as those returned by
- * miniflow_hash_in_minimask(), only the form of the arguments differ. */
-uint32_t
-flow_hash_in_minimask(const struct flow *flow, const struct minimask *mask,
- uint32_t basis)
-{
- const uint32_t *flow_u32 = (const uint32_t *)flow;
- const uint32_t *p = mask->masks.values;
- uint32_t hash;
- uint64_t map;
-
- hash = basis;
- for (map = mask->masks.map; map; map = zero_rightmost_1bit(map)) {
- hash = mhash_add(hash, flow_u32[raw_ctz(map)] & *p++);
- }
-
- return mhash_finish(hash, (p - mask->masks.values) * 4);
-}
-
-/* Returns a hash value for the bits of range [start, end) in 'flow',
- * where there are 1-bits in 'mask', given 'hash'.
- *
- * The hash values returned by this function are the same as those returned by
- * minimatch_hash_range(), only the form of the arguments differ. */
-uint32_t
-flow_hash_in_minimask_range(const struct flow *flow,
- const struct minimask *mask,
- uint8_t start, uint8_t end, uint32_t *basis)
-{
- const uint32_t *flow_u32 = (const uint32_t *)flow;
- unsigned int offset;
- uint64_t map = miniflow_get_map_in_range(&mask->masks, start, end,
- &offset);
- const uint32_t *p = mask->masks.values + offset;
- uint32_t hash = *basis;
-
- for (; map; map = zero_rightmost_1bit(map)) {
- hash = mhash_add(hash, flow_u32[raw_ctz(map)] & *p++);
- }
-
- *basis = hash; /* Allow continuation from the unfinished value. */
- return mhash_finish(hash, (p - mask->masks.values) * 4);
-}
-
\f
-/* Initializes 'dst' as a copy of 'src'. The caller must eventually free 'dst'
- * with minimask_destroy(). */
void
minimask_init(struct minimask *mask, const struct flow_wildcards *wc)
{
miniflow_init(&mask->masks, &wc->masks);
}
-/* Initializes 'dst' as a copy of 'src'. The caller must eventually free 'dst'
- * with minimask_destroy(). */
-void
-minimask_clone(struct minimask *dst, const struct minimask *src)
+/* Returns a minimask copy of 'wc'. The caller must eventually free the
+ * returned minimask with free(). */
+struct minimask *
+minimask_create(const struct flow_wildcards *wc)
{
- miniflow_clone(&dst->masks, &src->masks);
-}
-
-/* Initializes 'dst' with the data in 'src', destroying 'src'.
- * The caller must eventually free 'dst' with minimask_destroy(). */
-void
-minimask_move(struct minimask *dst, struct minimask *src)
-{
- miniflow_move(&dst->masks, &src->masks);
+ return (struct minimask *)miniflow_create(&wc->masks);
}
/* Initializes 'dst_' as the bit-wise "and" of 'a_' and 'b_'.
*
- * The caller must provide room for FLOW_U32S "uint32_t"s in 'storage', for use
- * by 'dst_'. The caller must *not* free 'dst_' with minimask_destroy(). */
+ * The caller must provide room for FLOW_U64S "uint64_t"s in 'storage', which
+ * must follow '*dst_' in memory, for use by 'dst_'. The caller must *not*
+ * free 'dst_' free(). */
void
minimask_combine(struct minimask *dst_,
const struct minimask *a_, const struct minimask *b_,
- uint32_t storage[FLOW_U32S])
+ uint64_t storage[FLOW_U64S])
{
struct miniflow *dst = &dst_->masks;
+ uint64_t *dst_values = storage;
const struct miniflow *a = &a_->masks;
const struct miniflow *b = &b_->masks;
- uint64_t map;
- int n = 0;
+ size_t idx;
- dst->values = storage;
+ flowmap_init(&dst->map);
- dst->map = 0;
- for (map = a->map & b->map; map; map = zero_rightmost_1bit(map)) {
- int ofs = raw_ctz(map);
- uint32_t mask = miniflow_get(a, ofs) & miniflow_get(b, ofs);
+ FLOWMAP_FOR_EACH_INDEX(idx, flowmap_and(a->map, b->map)) {
+ /* Both 'a' and 'b' have non-zero data at 'idx'. */
+ uint64_t mask = *miniflow_get__(a, idx) & *miniflow_get__(b, idx);
if (mask) {
- dst->map |= rightmost_1bit(map);
- dst->values[n++] = mask;
+ flowmap_set(&dst->map, idx, 1);
+ *dst_values++ = mask;
}
}
}
-/* Frees any memory owned by 'mask'. Does not free the storage in which 'mask'
- * itself resides; the caller is responsible for that. */
-void
-minimask_destroy(struct minimask *mask)
-{
- miniflow_destroy(&mask->masks);
-}
-
-/* Initializes 'dst' as a copy of 'src'. */
+/* Initializes 'wc' as a copy of 'mask'. */
void
minimask_expand(const struct minimask *mask, struct flow_wildcards *wc)
{
miniflow_expand(&mask->masks, &wc->masks);
}
-/* Returns the uint32_t that would be at byte offset '4 * u32_ofs' if 'mask'
- * were expanded into a "struct flow_wildcards". */
-uint32_t
-minimask_get(const struct minimask *mask, unsigned int u32_ofs)
-{
- return miniflow_get(&mask->masks, u32_ofs);
-}
-
-/* Returns true if 'a' and 'b' are the same flow mask, false otherwise. */
+/* Returns true if 'a' and 'b' are the same flow mask, false otherwise.
+ * Minimasks may not have zero data values, so for the minimasks to be the
+ * same, they need to have the same map and the same data values. */
bool
minimask_equal(const struct minimask *a, const struct minimask *b)
{
- return miniflow_equal(&a->masks, &b->masks);
-}
-
-/* Returns a hash value for 'mask', given 'basis'. */
-uint32_t
-minimask_hash(const struct minimask *mask, uint32_t basis)
-{
- return miniflow_hash(&mask->masks, basis);
+ return !memcmp(a, b, sizeof *a
+ + MINIFLOW_VALUES_SIZE(miniflow_n_values(&a->masks)));
}
-/* Returns true if at least one bit is wildcarded in 'a_' but not in 'b_',
+/* Returns true if at least one bit matched by 'b' is wildcarded by 'a',
* false otherwise. */
bool
-minimask_has_extra(const struct minimask *a_, const struct minimask *b_)
+minimask_has_extra(const struct minimask *a, const struct minimask *b)
{
- const struct miniflow *a = &a_->masks;
- const struct miniflow *b = &b_->masks;
- uint64_t map;
+ const uint64_t *bp = miniflow_get_values(&b->masks);
+ size_t idx;
- for (map = a->map | b->map; map; map = zero_rightmost_1bit(map)) {
- int ofs = raw_ctz(map);
- uint32_t a_u32 = miniflow_get(a, ofs);
- uint32_t b_u32 = miniflow_get(b, ofs);
+ FLOWMAP_FOR_EACH_INDEX(idx, b->masks.map) {
+ uint64_t b_u64 = *bp++;
- if ((a_u32 & b_u32) != b_u32) {
- return true;
+ /* 'b_u64' is non-zero, check if the data in 'a' is either zero
+ * or misses some of the bits in 'b_u64'. */
+ if (!MINIFLOW_IN_MAP(&a->masks, idx)
+ || ((*miniflow_get__(&a->masks, idx) & b_u64) != b_u64)) {
+ return true; /* 'a' wildcards some bits 'b' doesn't. */
}
}
return false;
}
-
-/* Returns true if 'mask' matches every packet, false if 'mask' fixes any bits
- * or fields. */
-bool
-minimask_is_catchall(const struct minimask *mask_)
-{
- const struct miniflow *mask = &mask_->masks;
- const uint32_t *p = mask->values;
- uint64_t map;
-
- for (map = mask->map; map; map = zero_rightmost_1bit(map)) {
- if (*p++) {
- return false;
- }
- }
- return true;
-}
projects / cascardo / ovs.git / blobdiff