From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga06.intel.com (mga06.intel.com [134.134.136.31]) by dpdk.org (Postfix) with ESMTP id D1CBC7CAA for ; Sat, 2 Sep 2017 03:25:11 +0200 (CEST) Received: from fmsmga006.fm.intel.com ([10.253.24.20]) by orsmga104.jf.intel.com with ESMTP; 01 Sep 2017 18:25:11 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.41,459,1498546800"; d="scan'208";a="147303382" Received: from bdw-yipeng.jf.intel.com ([10.54.81.30]) by fmsmga006.fm.intel.com with ESMTP; 01 Sep 2017 18:25:11 -0700 From: Yipeng Wang To: dev@dpdk.org Cc: stephen@networkplumber.org, luca.boccassi@gmail.com, charlie.tai@intel.com, sameh.gobriel@intel.com, ren.wang@intel.com, pablo.de.lara.guarch@intel.com, yipeng1.wang@intel.com Date: Fri, 1 Sep 2017 18:24:36 -0700 Message-Id: <1504315481-12854-3-git-send-email-yipeng1.wang@intel.com> X-Mailer: git-send-email 2.7.4 In-Reply-To: <1504315481-12854-1-git-send-email-yipeng1.wang@intel.com> References: <1503361193-36699-1-git-send-email-yipeng1.wang@intel.com> <1504315481-12854-1-git-send-email-yipeng1.wang@intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Subject: [dpdk-dev] [PATCH v2 2/7] member: implement HT mode X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 02 Sep 2017 01:25:12 -0000 One of the set-summray structure is hash-table based set-summary (HTSS). One example is cuckoo filter [1]. Comparing to a traditional hash table, HTSS has a much more compact structure. For each element, only one signature and its corresponding set ID is stored. No key comparison is required during lookup. For the table structure, there are multiple entries in each bucket, and the table is composed of many buckets. Two modes are supported for HTSS, "cache" and "none-cache" modes. The non-cache mode is more similar to traditional cuckoo filter. When a bucket is full, one entry will be evicted to its alternative bucket to make space for the new key. The table could be full and then no more keys could be inserted. This mode has false-positive rate but no false-negative. Multiple entries with same signature could stay in the same bucket. The "cache" mode does not evict key to its alternative bucket when a bucket is full, an existing key will be evicted out of the table like a cache. So the table will never reject keys when it is full. Another property is in each bucket, there cannot be multiple entries with same signature. The mode could have both false-positive and false-negative probability. This patch add the implementation of HTSS. [1] B Fan, D G Andersen and M Kaminsky, “Cuckoo Filter: Practically Better Than Bloom,” in Conference on emerging Networking Experiments and Technologies, 2014. Signed-off-by: Yipeng Wang --- lib/librte_member/Makefile | 2 +- lib/librte_member/rte_member_ht.c | 490 ++++++++++++++++++++++++++++++++++++++ lib/librte_member/rte_member_ht.h | 115 +++++++++ 3 files changed, 606 insertions(+), 1 deletion(-) create mode 100644 lib/librte_member/rte_member_ht.c create mode 100644 lib/librte_member/rte_member_ht.h diff --git a/lib/librte_member/Makefile b/lib/librte_member/Makefile index 1a79eaa..ad26548 100644 --- a/lib/librte_member/Makefile +++ b/lib/librte_member/Makefile @@ -42,7 +42,7 @@ EXPORT_MAP := rte_member_version.map LIBABIVER := 1 # all source are stored in SRCS-y -SRCS-$(CONFIG_RTE_LIBRTE_MEMBER) += rte_member.c +SRCS-$(CONFIG_RTE_LIBRTE_MEMBER) += rte_member.c rte_member_ht.c # install includes SYMLINK-$(CONFIG_RTE_LIBRTE_MEMBER)-include := rte_member.h diff --git a/lib/librte_member/rte_member_ht.c b/lib/librte_member/rte_member_ht.c new file mode 100644 index 0000000..b2ae6d0 --- /dev/null +++ b/lib/librte_member/rte_member_ht.c @@ -0,0 +1,490 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2017 Intel Corporation. All rights reserved. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in + * the documentation and/or other materials provided with the + * distribution. + * * Neither the name of Intel Corporation nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#include +#include +#include +#include +#include + +#include "rte_member.h" +#include "rte_member_ht.h" + + +static inline int +insert_overwrite_search(uint32_t bucket, SIG_TYPE tmp_sig, + struct member_ht_bucket *buckets, + MEMBER_SET_TYPE set_id) +{ + int i; + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + if (buckets[bucket].sigs[i] == tmp_sig) { + buckets[bucket].sets[i] = set_id; + return 1; + } + } + return 0; +} + + + + +static inline int +search_bucket_single(uint32_t bucket, SIG_TYPE tmp_sig, + struct member_ht_bucket *buckets, + MEMBER_SET_TYPE *set_id) +{ + int iter; + for (iter = 0; iter < RTE_MEMBER_BUCKET_ENTRIES; iter++) { + if (tmp_sig == buckets[bucket].sigs[iter] && + buckets[bucket].sets[iter] != + RTE_MEMBER_NO_MATCH) { + *set_id = buckets[bucket].sets[iter]; + return 1; + } + } + return 0; +} + + +static inline void +search_bucket_multi(uint32_t bucket, SIG_TYPE tmp_sig, + struct member_ht_bucket *buckets, + uint32_t *counter, + uint32_t match_per_key, + MEMBER_SET_TYPE *set_id) +{ + int iter; + for (iter = 0; iter < RTE_MEMBER_BUCKET_ENTRIES; iter++) { + if (tmp_sig == buckets[bucket].sigs[iter] && + buckets[bucket].sets[iter] != + RTE_MEMBER_NO_MATCH) { + set_id[*counter] = buckets[bucket].sets[iter]; + (*counter)++; + if (*counter >= match_per_key) + return; + } + } +} + + +int +rte_member_create_ht(struct rte_member_setsum *ss, + const struct rte_member_parameters *params) +{ + uint32_t i, j; + uint32_t size_bucket_t; + uint32_t num_entries = rte_align32pow2(params->num_keys); + + if ((num_entries > RTE_MEMBER_ENTRIES_MAX) || + !rte_is_power_of_2(RTE_MEMBER_BUCKET_ENTRIES) || + num_entries < RTE_MEMBER_BUCKET_ENTRIES) { + rte_errno = EINVAL; + RTE_LOG(ERR, MEMBER, + "Membership HT create with invalid parameters\n"); + return -EINVAL; + } + + uint32_t num_buckets = num_entries / RTE_MEMBER_BUCKET_ENTRIES; + + size_bucket_t = sizeof(struct member_ht_bucket); + + struct member_ht_bucket *buckets = rte_zmalloc_socket(NULL, + num_buckets * size_bucket_t, + RTE_CACHE_LINE_SIZE, ss->socket_id); + + if (buckets == NULL) + return -ENOMEM; + + ss->table = buckets; + ss->bucket_cnt = num_buckets; + ss->bucket_mask = num_buckets - 1; + ss->cache = params->iscache; + + for (i = 0; i < num_buckets; i++) { + for (j = 0; j < RTE_MEMBER_BUCKET_ENTRIES; j++) + buckets[i].sets[j] = RTE_MEMBER_NO_MATCH; + } + + + RTE_LOG(DEBUG, MEMBER, "Hash table based filter created, " + "the table has %u entries, %u buckets\n", + num_buckets, + num_buckets / RTE_MEMBER_BUCKET_ENTRIES); + return 0; +} + +static inline +void get_buckets_index(const struct rte_member_setsum *ss, const void *key, + uint32_t *prim_bkt, uint32_t *sec_bkt, SIG_TYPE *sig) +{ + uint32_t first_hash = MEMBER_PRIM_HASH(key, ss->key_len, + ss->prim_hash_seed); + uint32_t sec_hash = MEMBER_SEC_HASH(&first_hash, 4, ss->sec_hash_seed); + *sig = first_hash & SIG_BITMASK; + if (ss->cache) { + *prim_bkt = sec_hash & ss->bucket_mask; + *sec_bkt = (sec_hash >> 16) & ss->bucket_mask; + } else { + *prim_bkt = sec_hash & ss->bucket_mask; + *sec_bkt = (*prim_bkt ^ *sig) & ss->bucket_mask; + } +} + + +int +rte_member_lookup_ht(const struct rte_member_setsum *ss, + const void *key, MEMBER_SET_TYPE *set_id) +{ + uint32_t prim_bucket, sec_bucket; + SIG_TYPE tmp_sig; + struct member_ht_bucket *buckets = ss->table; + + + *set_id = RTE_MEMBER_NO_MATCH; + get_buckets_index(ss, key, &prim_bucket, &sec_bucket, &tmp_sig); + + if (search_bucket_single(prim_bucket, tmp_sig, buckets, + set_id) || + search_bucket_single(sec_bucket, tmp_sig, + buckets, set_id)) + return 1; + + return 0; +} + + +uint32_t +rte_member_lookup_bulk_ht(const struct rte_member_setsum *ss, + const void **keys, uint32_t num_keys, MEMBER_SET_TYPE *set_id) +{ + uint32_t i; + uint32_t ret = 0; + struct member_ht_bucket *buckets = ss->table; + SIG_TYPE tmp_sig[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + uint32_t prim_buckets[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + uint32_t sec_buckets[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + + for (i = 0; i < num_keys; i++) { + get_buckets_index(ss, keys[i], &prim_buckets[i], + &sec_buckets[i], &tmp_sig[i]); + rte_prefetch0(&buckets[prim_buckets[i]]); + rte_prefetch0(&buckets[sec_buckets[i]]); + } + + for (i = 0; i < num_keys; i++) { + if (search_bucket_single(prim_buckets[i], tmp_sig[i], + buckets, &set_id[i]) || + search_bucket_single(sec_buckets[i], + tmp_sig[i], buckets, &set_id[i])) + ret++; + else + set_id[i] = RTE_MEMBER_NO_MATCH; + } + return ret; +} + + +uint32_t +rte_member_lookup_multi_ht(const struct rte_member_setsum *ss, + const void *key, uint32_t match_per_key, + MEMBER_SET_TYPE *set_id) +{ + uint32_t ret = 0; + uint32_t prim_bucket, sec_bucket; + SIG_TYPE tmp_sig; + struct member_ht_bucket *buckets = ss->table; + + get_buckets_index(ss, key, &prim_bucket, &sec_bucket, &tmp_sig); + + search_bucket_multi(prim_bucket, tmp_sig, buckets, &ret, + match_per_key, set_id); + if (ret < match_per_key) + search_bucket_multi(sec_bucket, tmp_sig, + buckets, &ret, match_per_key, set_id); + return ret; +} + + +uint32_t +rte_member_lookup_multi_bulk_ht(const struct rte_member_setsum *ss, + const void **keys, uint32_t num_keys, uint32_t match_per_key, + uint32_t *match_count, + MEMBER_SET_TYPE *set_ids) +{ + uint32_t i; + uint32_t ret = 0; + struct member_ht_bucket *buckets = ss->table; + uint32_t match_cnt_t; + SIG_TYPE tmp_sig[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + uint32_t prim_buckets[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + uint32_t sec_buckets[RTE_MEMBER_LOOKUP_BULK_MAX] = {0}; + + for (i = 0; i < num_keys; i++) { + get_buckets_index(ss, keys[i], &prim_buckets[i], + &sec_buckets[i], &tmp_sig[i]); + rte_prefetch0(&buckets[prim_buckets[i]]); + rte_prefetch0(&buckets[sec_buckets[i]]); + } + for (i = 0; i < num_keys; i++) { + match_cnt_t = 0; + + search_bucket_multi(prim_buckets[i], tmp_sig[i], + buckets, &match_cnt_t, match_per_key, + &set_ids[i*match_per_key]); + if (match_cnt_t < match_per_key) + search_bucket_multi(sec_buckets[i], tmp_sig[i], + buckets, &match_cnt_t, match_per_key, + &set_ids[i*match_per_key]); + match_count[i] = match_cnt_t; + if (match_cnt_t != 0) + ret++; + } + return ret; +} + +static inline int +try_insert(struct member_ht_bucket *buckets, uint32_t prim, uint32_t sec, + SIG_TYPE sig, MEMBER_SET_TYPE set_id) +{ + int i; + /* If not full then insert into one slot*/ + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + if (buckets[prim].sets[i] == RTE_MEMBER_NO_MATCH) { + buckets[prim].sigs[i] = sig; + buckets[prim].sets[i] = set_id; + return 0; + } + } + /* if prim failed, we need to access second cache line */ + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + if (buckets[sec].sets[i] == RTE_MEMBER_NO_MATCH) { + buckets[sec].sigs[i] = sig; + buckets[sec].sets[i] = set_id; + return 0; + } + } + return -1; +} + + +static inline int +try_overwrite(struct member_ht_bucket *buckets, uint32_t prim, uint32_t sec, + SIG_TYPE sig, MEMBER_SET_TYPE set_id) +{ + if (insert_overwrite_search(prim, sig, buckets, set_id) || + insert_overwrite_search(sec, sig, buckets, + set_id)) + return 0; + return -1; +} + +static inline int +evict_from_bucket(void) +{ + /* for now, we randomly pick one entry to evict */ + return rte_rand() & (RTE_MEMBER_BUCKET_ENTRIES - 1); +} + + + +/* + * This function is similar to the cuckoo hash make_space function in hash + * library + */ +static inline int +make_space_bucket(const struct rte_member_setsum *ss, uint32_t bkt_num) +{ + + static unsigned int nr_pushes; + + unsigned int i, j; + int ret; + struct member_ht_bucket *buckets = ss->table; + uint32_t next_bucket_idx; + struct member_ht_bucket *next_bkt[RTE_MEMBER_BUCKET_ENTRIES]; + struct member_ht_bucket *bkt = &buckets[bkt_num]; + MEMBER_SET_TYPE flag_mask = 1U << (sizeof(MEMBER_SET_TYPE) * 8 - 1); + /* + * Push existing item (search for bucket with space in + * alternative locations) to its alternative location + */ + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + /* Search for space in alternative locations */ + next_bucket_idx = (bkt->sigs[i] ^ bkt_num) & ss->bucket_mask; + next_bkt[i] = &buckets[next_bucket_idx]; + for (j = 0; j < RTE_MEMBER_BUCKET_ENTRIES; j++) { + if (next_bkt[i]->sets[j] == RTE_MEMBER_NO_MATCH) + break; + } + + if (j != RTE_MEMBER_BUCKET_ENTRIES) + break; + } + + /* Alternative location has spare room (end of recursive function) */ + if (i != RTE_MEMBER_BUCKET_ENTRIES) { + next_bkt[i]->sigs[j] = bkt->sigs[i]; + next_bkt[i]->sets[j] = bkt->sets[i]; + return i; + } + + /* Pick entry that has not been pushed yet */ + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) + if ((bkt->sets[i] & flag_mask) == 0) + break; + + /* All entries have been pushed, so entry cannot be added */ + if (i == RTE_MEMBER_BUCKET_ENTRIES || + nr_pushes > RTE_MEMBER_MAX_PUSHES) + return -ENOSPC; + + next_bucket_idx = (bkt->sigs[i] ^ bkt_num) & ss->bucket_mask; + /* Set flag to indicate that this entry is going to be pushed */ + bkt->sets[i] |= flag_mask; + + nr_pushes++; + /* Need room in alternative bucket to insert the pushed entry */ + ret = make_space_bucket(ss, next_bucket_idx); + /* + * After recursive function. + * Clear flags and insert the pushed entry + * in its alternative location if successful, + * or return error + */ + bkt->sets[i] &= ~flag_mask; + nr_pushes = 0; + if (ret >= 0) { + next_bkt[i]->sigs[ret] = bkt->sigs[i]; + next_bkt[i]->sets[ret] = bkt->sets[i]; + return i; + } else + return ret; +} + + +int +rte_member_add_ht(const struct rte_member_setsum *ss, + const void *key, MEMBER_SET_TYPE set_id) +{ + int ret; + uint32_t prim_bucket, sec_bucket; + SIG_TYPE tmp_sig; + struct member_ht_bucket *buckets = ss->table; + MEMBER_SET_TYPE flag_mask = 1U << (sizeof(MEMBER_SET_TYPE) * 8 - 1); + + if (set_id == RTE_MEMBER_NO_MATCH || (set_id & flag_mask) != 0) + return -EINVAL; + + get_buckets_index(ss, key, &prim_bucket, &sec_bucket, &tmp_sig); + + /* if it is cache based filter, we try overwriting existing entry */ + if (ss->cache) { + ret = try_overwrite(buckets, prim_bucket, sec_bucket, tmp_sig, + set_id); + if (ret != -1) + return ret; + } + /* If not full then insert into one slot*/ + ret = try_insert(buckets, prim_bucket, sec_bucket, tmp_sig, set_id); + if (ret != -1) + return ret; + + /* random pick prim or sec for recursive displacement */ + + uint32_t select_bucket = (tmp_sig && 1U) ? prim_bucket : sec_bucket; + if (ss->cache) { + ret = evict_from_bucket(); + buckets[select_bucket].sigs[ret] = tmp_sig; + buckets[select_bucket].sets[ret] = set_id; + return 1; + } + + ret = make_space_bucket(ss, select_bucket); + if (ret >= 0) { + buckets[select_bucket].sigs[ret] = tmp_sig; + buckets[select_bucket].sets[ret] = set_id; + ret = 1; + } + + return ret; +} + + +void +rte_member_free_ht(struct rte_member_setsum *ss) +{ + rte_free(ss->table); +} + + +int +rte_member_delete_ht(struct rte_member_setsum *ss, const void *key, + MEMBER_SET_TYPE set_id) +{ + int i; + uint32_t prim_bucket, sec_bucket; + SIG_TYPE tmp_sig; + struct member_ht_bucket *buckets = ss->table; + + get_buckets_index(ss, key, &prim_bucket, &sec_bucket, &tmp_sig); + + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + if (tmp_sig == buckets[prim_bucket].sigs[i] && + set_id == buckets[prim_bucket].sets[i]) { + buckets[prim_bucket].sets[i] = RTE_MEMBER_NO_MATCH; + return 0; + } + } + + for (i = 0; i < RTE_MEMBER_BUCKET_ENTRIES; i++) { + if (tmp_sig == buckets[sec_bucket].sigs[i] && + set_id == buckets[sec_bucket].sets[i]) { + buckets[sec_bucket].sets[i] = RTE_MEMBER_NO_MATCH; + return 0; + } + } + return -ENOENT; +} + + +void +rte_member_reset_ht(const struct rte_member_setsum *ss) +{ + uint32_t i, j; + struct member_ht_bucket *buckets = ss->table; + for (i = 0; i < ss->bucket_cnt; i++) { + for (j = 0; j < RTE_MEMBER_BUCKET_ENTRIES; j++) + buckets[i].sets[j] = RTE_MEMBER_NO_MATCH; + } +} diff --git a/lib/librte_member/rte_member_ht.h b/lib/librte_member/rte_member_ht.h new file mode 100644 index 0000000..aebf1db --- /dev/null +++ b/lib/librte_member/rte_member_ht.h @@ -0,0 +1,115 @@ +/*- + * BSD LICENSE + * + * Copyright(c) 2017 Intel Corporation. All rights reserved. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in + * the documentation and/or other materials provided with the + * distribution. + * * Neither the name of Intel Corporation nor the names of its + * contributors may be used to endorse or promote products derived + * from this software without specific prior written permission. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + */ + +#ifndef _RTE_MEMBER_HT_H_ +#define _RTE_MEMBER_HT_H_ + +#ifdef __cplusplus +extern "C" { +#endif + +/* Maximum number of pushes for cuckoo path in HT mode. */ +#define RTE_MEMBER_MAX_PUSHES 100 + +typedef uint16_t SIG_TYPE; /* signature size is 16 bit */ +#define SIG_BITMASK 0xFFFF /* signature mask, 16 bit */ + + +/* The bucket struct for ht setsum */ +struct member_ht_bucket { + SIG_TYPE sigs[RTE_MEMBER_BUCKET_ENTRIES]; /* 2-byte signature */ + MEMBER_SET_TYPE sets[RTE_MEMBER_BUCKET_ENTRIES]; /* 2-byte set */ +} __rte_cache_aligned; + + + +int +rte_member_create_ht(struct rte_member_setsum *ss, + const struct rte_member_parameters *params); + +int +rte_member_lookup_ht(const struct rte_member_setsum *setsum, + const void *key, MEMBER_SET_TYPE *set_id); + + + +uint32_t +rte_member_lookup_bulk_ht(const struct rte_member_setsum *setsum, + const void **keys, uint32_t num_keys, + MEMBER_SET_TYPE *set_ids); + + + + +uint32_t +rte_member_lookup_multi_ht(const struct rte_member_setsum *setsum, + const void *key, uint32_t match_per_key, + MEMBER_SET_TYPE *set_id); + + + +uint32_t +rte_member_lookup_multi_bulk_ht(const struct rte_member_setsum *setsum, + const void **keys, uint32_t num_keys, uint32_t match_per_key, + uint32_t *match_count, + MEMBER_SET_TYPE *set_ids); + + + + +int +rte_member_add_ht(const struct rte_member_setsum *setsum, + const void *key, MEMBER_SET_TYPE set_id); + +void +rte_member_free_ht(struct rte_member_setsum *setsum); + + + + +int +rte_member_delete_ht(struct rte_member_setsum *ss, const void *key, + MEMBER_SET_TYPE set_id); + + +void +rte_member_reset_ht(const struct rte_member_setsum *setsum); + + +#ifdef __cplusplus +} +#endif + + + +#endif /* _RTE_MEMBER_HT_H_ */ -- 2.7.4