[dpdk-dev] [PATCH v3 35/41] table: add exact match SWX table

[Cristian Dumitrescu <cristian.dumitrescu@intel.com>]

Add the exact match table type for the SWX pipeline. Used under the
hood by the SWX pipeline table instruction.

Signed-off-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
---
 lib/librte_table/meson.build           |   6 +-
 lib/librte_table/rte_swx_table_em.c    | 851 +++++++++++++++++++++++++
 lib/librte_table/rte_swx_table_em.h    |  30 +
 lib/librte_table/rte_table_version.map |   7 +
 4 files changed, 892 insertions(+), 2 deletions(-)
 create mode 100644 lib/librte_table/rte_swx_table_em.c
 create mode 100644 lib/librte_table/rte_swx_table_em.h

diff --git a/lib/librte_table/meson.build b/lib/librte_table/meson.build
index b9d4fe3dc..d69678386 100644
--- a/lib/librte_table/meson.build
+++ b/lib/librte_table/meson.build
@@ -11,7 +11,8 @@ sources = files('rte_table_acl.c',
 		'rte_table_hash_ext.c',
 		'rte_table_hash_lru.c',
 		'rte_table_array.c',
-		'rte_table_stub.c')
+		'rte_table_stub.c',
+		'rte_swx_table_em.c',)
 headers = files('rte_table.h',
 		'rte_table_acl.h',
 		'rte_table_lpm.h',
@@ -23,7 +24,8 @@ headers = files('rte_table.h',
 		'rte_lru.h',
 		'rte_table_array.h',
 		'rte_table_stub.h',
-		'rte_swx_table.h',)
+		'rte_swx_table.h',
+		'rte_swx_table_em.h',)
 deps += ['mbuf', 'port', 'lpm', 'hash', 'acl']
 
 if arch_subdir == 'x86'
diff --git a/lib/librte_table/rte_swx_table_em.c b/lib/librte_table/rte_swx_table_em.c
new file mode 100644
index 000000000..85c77ad03
--- /dev/null
+++ b/lib/librte_table/rte_swx_table_em.c
@@ -0,0 +1,851 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2020 Intel Corporation
+ */
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+#include <errno.h>
+
+#include <rte_common.h>
+#include <rte_prefetch.h>
+
+#include "rte_swx_table_em.h"
+
+#define CHECK(condition, err_code)                                             \
+do {                                                                           \
+	if (!(condition))                                                      \
+		return -(err_code);                                            \
+} while (0)
+
+#ifndef RTE_SWX_TABLE_EM_USE_HUGE_PAGES
+#define RTE_SWX_TABLE_EM_USE_HUGE_PAGES 1
+#endif
+
+#if RTE_SWX_TABLE_EM_USE_HUGE_PAGES
+
+#include <rte_malloc.h>
+
+static void *
+env_malloc(size_t size, size_t alignment, int numa_node)
+{
+	return rte_zmalloc_socket(NULL, size, alignment, numa_node);
+}
+
+static void
+env_free(void *start, size_t size __rte_unused)
+{
+	rte_free(start);
+}
+
+#else
+
+#include <numa.h>
+
+static void *
+env_malloc(size_t size, size_t alignment __rte_unused, int numa_node)
+{
+	return numa_alloc_onnode(size, numa_node);
+}
+
+static void
+env_free(void *start, size_t size)
+{
+	numa_free(start, size);
+}
+
+#endif
+
+#if defined(RTE_ARCH_X86_64)
+
+#include <x86intrin.h>
+
+#define crc32_u64(crc, v) _mm_crc32_u64(crc, v)
+
+#else
+
+static inline uint64_t
+crc32_u64_generic(uint64_t crc, uint64_t value)
+{
+	int i;
+
+	crc = (crc & 0xFFFFFFFFLLU) ^ value;
+	for (i = 63; i >= 0; i--) {
+		uint64_t mask;
+
+		mask = -(crc & 1LLU);
+		crc = (crc >> 1LLU) ^ (0x82F63B78LLU & mask);
+	}
+
+	return crc;
+}
+
+#define crc32_u64(crc, v) crc32_u64_generic(crc, v)
+
+#endif
+
+/* Key size needs to be one of: 8, 16, 32 or 64. */
+static inline uint32_t
+hash(void *key, void *key_mask, uint32_t key_size, uint32_t seed)
+{
+	uint64_t *k = key;
+	uint64_t *m = key_mask;
+	uint64_t k0, k2, k5, crc0, crc1, crc2, crc3, crc4, crc5;
+
+	switch (key_size) {
+	case 8:
+		crc0 = crc32_u64(seed, k[0] & m[0]);
+		return crc0;
+
+	case 16:
+		k0 = k[0] & m[0];
+
+		crc0 = crc32_u64(k0, seed);
+		crc1 = crc32_u64(k0 >> 32, k[1] & m[1]);
+
+		crc0 ^= crc1;
+
+		return crc0;
+
+	case 32:
+		k0 = k[0] & m[0];
+		k2 = k[2] & m[2];
+
+		crc0 = crc32_u64(k0, seed);
+		crc1 = crc32_u64(k0 >> 32, k[1] & m[1]);
+
+		crc2 = crc32_u64(k2, k[3] & m[3]);
+		crc3 = k2 >> 32;
+
+		crc0 = crc32_u64(crc0, crc1);
+		crc1 = crc32_u64(crc2, crc3);
+
+		crc0 ^= crc1;
+
+		return crc0;
+
+	case 64:
+		k0 = k[0] & m[0];
+		k2 = k[2] & m[2];
+		k5 = k[5] & m[5];
+
+		crc0 = crc32_u64(k0, seed);
+		crc1 = crc32_u64(k0 >> 32, k[1] & m[1]);
+
+		crc2 = crc32_u64(k2, k[3] & m[3]);
+		crc3 = crc32_u64(k2 >> 32, k[4] & m[4]);
+
+		crc4 = crc32_u64(k5, k[6] & m[6]);
+		crc5 = crc32_u64(k5 >> 32, k[7] & m[7]);
+
+		crc0 = crc32_u64(crc0, (crc1 << 32) ^ crc2);
+		crc1 = crc32_u64(crc3, (crc4 << 32) ^ crc5);
+
+		crc0 ^= crc1;
+
+		return crc0;
+
+	default:
+		crc0 = 0;
+		return crc0;
+	}
+}
+
+/* n_bytes needs to be a multiple of 8 bytes. */
+static void
+keycpy(void *dst, void *src, void *src_mask, uint32_t n_bytes)
+{
+	uint64_t *dst64 = dst, *src64 = src, *src_mask64 = src_mask;
+	uint32_t i;
+
+	for (i = 0; i < n_bytes / sizeof(uint64_t); i++)
+		dst64[i] = src64[i] & src_mask64[i];
+}
+
+/*
+ * Return: 0 = Keys are NOT equal; 1 = Keys are equal.
+ */
+static inline uint32_t
+keycmp(void *a, void *b, void *b_mask, uint32_t n_bytes)
+{
+	uint64_t *a64 = a, *b64 = b, *b_mask64 = b_mask;
+
+	switch (n_bytes) {
+	case 8: {
+		uint64_t xor0 = a64[0] ^ (b64[0] & b_mask64[0]);
+		uint32_t result = 1;
+
+		if (xor0)
+			result = 0;
+		return result;
+	}
+
+	case 16: {
+		uint64_t xor0 = a64[0] ^ (b64[0] & b_mask64[0]);
+		uint64_t xor1 = a64[1] ^ (b64[1] & b_mask64[1]);
+		uint64_t or = xor0 | xor1;
+		uint32_t result = 1;
+
+		if (or)
+			result = 0;
+		return result;
+	}
+
+	case 32: {
+		uint64_t xor0 = a64[0] ^ (b64[0] & b_mask64[0]);
+		uint64_t xor1 = a64[1] ^ (b64[1] & b_mask64[1]);
+		uint64_t xor2 = a64[2] ^ (b64[2] & b_mask64[2]);
+		uint64_t xor3 = a64[3] ^ (b64[3] & b_mask64[3]);
+		uint64_t or = (xor0 | xor1) | (xor2 | xor3);
+		uint32_t result = 1;
+
+		if (or)
+			result = 0;
+		return result;
+	}
+
+	case 64: {
+		uint64_t xor0 = a64[0] ^ (b64[0] & b_mask64[0]);
+		uint64_t xor1 = a64[1] ^ (b64[1] & b_mask64[1]);
+		uint64_t xor2 = a64[2] ^ (b64[2] & b_mask64[2]);
+		uint64_t xor3 = a64[3] ^ (b64[3] & b_mask64[3]);
+		uint64_t xor4 = a64[4] ^ (b64[4] & b_mask64[4]);
+		uint64_t xor5 = a64[5] ^ (b64[5] & b_mask64[5]);
+		uint64_t xor6 = a64[6] ^ (b64[6] & b_mask64[6]);
+		uint64_t xor7 = a64[7] ^ (b64[7] & b_mask64[7]);
+		uint64_t or = ((xor0 | xor1) | (xor2 | xor3)) |
+			      ((xor4 | xor5) | (xor6 | xor7));
+		uint32_t result = 1;
+
+		if (or)
+			result = 0;
+		return result;
+	}
+
+	default: {
+		uint32_t i;
+
+		for (i = 0; i < n_bytes / sizeof(uint64_t); i++)
+			if (a64[i] != (b64[i] & b_mask64[i]))
+				return 0;
+		return 1;
+	}
+	}
+}
+
+#define KEYS_PER_BUCKET 4
+
+struct bucket_extension {
+	struct bucket_extension *next;
+	uint16_t sig[KEYS_PER_BUCKET];
+	uint32_t key_id[KEYS_PER_BUCKET];
+};
+
+struct table {
+	/* Input parameters */
+	struct rte_swx_table_params params;
+
+	/* Internal. */
+	uint32_t key_size;
+	uint32_t data_size;
+	uint32_t key_size_shl;
+	uint32_t data_size_shl;
+	uint32_t n_buckets;
+	uint32_t n_buckets_ext;
+	uint32_t key_stack_tos;
+	uint32_t bkt_ext_stack_tos;
+	uint64_t total_size;
+
+	/* Memory arrays. */
+	uint8_t *key_mask;
+	struct bucket_extension *buckets;
+	struct bucket_extension *buckets_ext;
+	uint8_t *keys;
+	uint32_t *key_stack;
+	uint32_t *bkt_ext_stack;
+	uint8_t *data;
+};
+
+static inline uint8_t *
+table_key(struct table *t, uint32_t key_id)
+{
+	return &t->keys[(uint64_t)key_id << t->key_size_shl];
+}
+
+static inline uint64_t *
+table_key_data(struct table *t, uint32_t key_id)
+{
+	return (uint64_t *)&t->data[(uint64_t)key_id << t->data_size_shl];
+}
+
+static inline int
+bkt_is_empty(struct bucket_extension *bkt)
+{
+	return (!bkt->sig[0] && !bkt->sig[1] && !bkt->sig[2] && !bkt->sig[2]) ?
+		1 : 0;
+}
+
+/* Return:
+ *    0 = Bucket key position is NOT empty;
+ *    1 = Bucket key position is empty.
+ */
+static inline int
+bkt_key_is_empty(struct bucket_extension *bkt, uint32_t bkt_pos)
+{
+	return bkt->sig[bkt_pos] ? 0 : 1;
+}
+
+/* Return: 0 = Keys are NOT equal; 1 = Keys are equal. */
+static inline int
+bkt_keycmp(struct table *t,
+	   struct bucket_extension *bkt,
+	   uint8_t *input_key,
+	   uint32_t bkt_pos,
+	   uint32_t input_sig)
+{
+	uint32_t bkt_key_id;
+	uint8_t *bkt_key;
+
+	/* Key signature comparison. */
+	if (input_sig != bkt->sig[bkt_pos])
+		return 0;
+
+	/* Key comparison. */
+	bkt_key_id = bkt->key_id[bkt_pos];
+	bkt_key = table_key(t, bkt_key_id);
+	return keycmp(bkt_key, input_key, t->key_mask, t->key_size);
+}
+
+static inline void
+bkt_key_install(struct table *t,
+		struct bucket_extension *bkt,
+		struct rte_swx_table_entry *input,
+		uint32_t bkt_pos,
+		uint32_t bkt_key_id,
+		uint32_t input_sig)
+{
+	uint8_t *bkt_key;
+	uint64_t *bkt_data;
+
+	/* Key signature. */
+	bkt->sig[bkt_pos] = (uint16_t)input_sig;
+
+	/* Key. */
+	bkt->key_id[bkt_pos] = bkt_key_id;
+	bkt_key = table_key(t, bkt_key_id);
+	keycpy(bkt_key, input->key, t->key_mask, t->key_size);
+
+	/* Key data. */
+	bkt_data = table_key_data(t, bkt_key_id);
+	bkt_data[0] = input->action_id;
+	if (t->params.action_data_size)
+		memcpy(&bkt_data[1],
+		       input->action_data,
+		       t->params.action_data_size);
+}
+
+static inline void
+bkt_key_data_update(struct table *t,
+		    struct bucket_extension *bkt,
+		    struct rte_swx_table_entry *input,
+		    uint32_t bkt_pos)
+{
+	uint32_t bkt_key_id;
+	uint64_t *bkt_data;
+
+	/* Key. */
+	bkt_key_id = bkt->key_id[bkt_pos];
+
+	/* Key data. */
+	bkt_data = table_key_data(t, bkt_key_id);
+	bkt_data[0] = input->action_id;
+	if (t->params.action_data_size)
+		memcpy(&bkt_data[1],
+		       input->action_data,
+		       t->params.action_data_size);
+}
+
+#define CL RTE_CACHE_LINE_ROUNDUP
+
+static int
+__table_create(struct table **table,
+	       uint64_t *memory_footprint,
+	       struct rte_swx_table_params *params,
+	       const char *args __rte_unused,
+	       int numa_node)
+{
+	struct table *t;
+	uint8_t *memory;
+	size_t table_meta_sz, key_mask_sz, bucket_sz, bucket_ext_sz, key_sz,
+		key_stack_sz, bkt_ext_stack_sz, data_sz, total_size;
+	size_t key_mask_offset, bucket_offset, bucket_ext_offset, key_offset,
+		key_stack_offset, bkt_ext_stack_offset, data_offset;
+	uint32_t key_size, key_data_size, n_buckets, n_buckets_ext, i;
+
+	/* Check input arguments. */
+	CHECK(params, EINVAL);
+	CHECK(params->match_type == RTE_SWX_TABLE_MATCH_EXACT, EINVAL);
+	CHECK(params->key_size, EINVAL);
+	CHECK(params->key_size <= 64, EINVAL);
+	CHECK(params->n_keys_max, EINVAL);
+
+	/* Memory allocation. */
+	key_size = rte_align64pow2(params->key_size);
+	if (key_size < 8)
+		key_size = 8;
+	key_data_size = rte_align64pow2(params->action_data_size + 8);
+	n_buckets = params->n_keys_max / KEYS_PER_BUCKET;
+	n_buckets_ext = params->n_keys_max / KEYS_PER_BUCKET;
+
+	table_meta_sz = CL(sizeof(struct table));
+	key_mask_sz = CL(key_size);
+	bucket_sz = CL(n_buckets * sizeof(struct bucket_extension));
+	bucket_ext_sz = CL(n_buckets_ext * sizeof(struct bucket_extension));
+	key_sz = CL(params->n_keys_max * key_size);
+	key_stack_sz = CL(params->n_keys_max * sizeof(uint32_t));
+	bkt_ext_stack_sz = CL(n_buckets_ext * sizeof(uint32_t));
+	data_sz = CL(params->n_keys_max * key_data_size);
+	total_size = table_meta_sz + key_mask_sz + bucket_sz + bucket_ext_sz +
+		     key_sz + key_stack_sz + bkt_ext_stack_sz + data_sz;
+
+	key_mask_offset = table_meta_sz;
+	bucket_offset = key_mask_offset + key_mask_sz;
+	bucket_ext_offset = bucket_offset + bucket_sz;
+	key_offset = bucket_ext_offset + bucket_ext_sz;
+	key_stack_offset = key_offset + key_sz;
+	bkt_ext_stack_offset = key_stack_offset + key_stack_sz;
+	data_offset = bkt_ext_stack_offset + bkt_ext_stack_sz;
+
+	if (!table) {
+		if (memory_footprint)
+			*memory_footprint = total_size;
+		return 0;
+	}
+
+	memory = env_malloc(total_size, RTE_CACHE_LINE_SIZE, numa_node);
+	CHECK(memory,  ENOMEM);
+	memset(memory, 0, total_size);
+
+	/* Initialization. */
+	t = (struct table *)memory;
+	memcpy(&t->params, params, sizeof(*params));
+
+	t->key_size = key_size;
+	t->data_size = key_data_size;
+	t->key_size_shl = __builtin_ctzl(key_size);
+	t->data_size_shl = __builtin_ctzl(key_data_size);
+	t->n_buckets = n_buckets;
+	t->n_buckets_ext = n_buckets_ext;
+	t->total_size = total_size;
+
+	t->key_mask = &memory[key_mask_offset];
+	t->buckets = (struct bucket_extension *)&memory[bucket_offset];
+	t->buckets_ext = (struct bucket_extension *)&memory[bucket_ext_offset];
+	t->keys = &memory[key_offset];
+	t->key_stack = (uint32_t *)&memory[key_stack_offset];
+	t->bkt_ext_stack = (uint32_t *)&memory[bkt_ext_stack_offset];
+	t->data = &memory[data_offset];
+
+	t->params.key_mask0 = t->key_mask;
+
+	if (!params->key_mask0)
+		memset(t->key_mask, 0xFF, params->key_size);
+	else
+		memcpy(t->key_mask, params->key_mask0, params->key_size);
+
+	for (i = 0; i < t->params.n_keys_max; i++)
+		t->key_stack[i] = t->params.n_keys_max - 1 - i;
+	t->key_stack_tos = t->params.n_keys_max;
+
+	for (i = 0; i < n_buckets_ext; i++)
+		t->bkt_ext_stack[i] = n_buckets_ext - 1 - i;
+	t->bkt_ext_stack_tos = n_buckets_ext;
+
+	*table = t;
+	return 0;
+}
+
+static void
+table_free(void *table)
+{
+	struct table *t = table;
+
+	if (!t)
+		return;
+
+	env_free(t, t->total_size);
+}
+
+static int
+table_add(void *table, struct rte_swx_table_entry *entry)
+{
+	struct table *t = table;
+	struct bucket_extension *bkt0, *bkt, *bkt_prev;
+	uint32_t input_sig, bkt_id, i;
+
+	CHECK(t, EINVAL);
+	CHECK(entry, EINVAL);
+	CHECK(entry->key, EINVAL);
+	CHECK((!t->params.action_data_size && !entry->action_data) ||
+	      (t->params.action_data_size && entry->action_data), EINVAL);
+
+	input_sig = hash(entry->key, t->key_mask, t->key_size, 0);
+	bkt_id = input_sig & (t->n_buckets - 1);
+	bkt0 = &t->buckets[bkt_id];
+	input_sig = (input_sig >> 16) | 1;
+
+	/* Key is present in the bucket. */
+	for (bkt = bkt0; bkt; bkt = bkt->next)
+		for (i = 0; i < KEYS_PER_BUCKET; i++)
+			if (bkt_keycmp(t, bkt, entry->key, i, input_sig)) {
+				bkt_key_data_update(t, bkt, entry, i);
+				return 0;
+			}
+
+	/* Key is not present in the bucket. Bucket not full. */
+	for (bkt = bkt0, bkt_prev = NULL; bkt; bkt_prev = bkt, bkt = bkt->next)
+		for (i = 0; i < KEYS_PER_BUCKET; i++)
+			if (bkt_key_is_empty(bkt, i)) {
+				uint32_t new_bkt_key_id;
+
+				/* Allocate new key & install. */
+				CHECK(t->key_stack_tos, ENOSPC);
+				new_bkt_key_id =
+					t->key_stack[--t->key_stack_tos];
+				bkt_key_install(t, bkt, entry, i,
+						new_bkt_key_id, input_sig);
+				return 0;
+			}
+
+	/* Bucket full: extend bucket. */
+	if (t->bkt_ext_stack_tos && t->key_stack_tos) {
+		struct bucket_extension *new_bkt;
+		uint32_t new_bkt_id, new_bkt_key_id;
+
+		/* Allocate new bucket extension & install. */
+		new_bkt_id = t->bkt_ext_stack[--t->bkt_ext_stack_tos];
+		new_bkt = &t->buckets_ext[new_bkt_id];
+		memset(new_bkt, 0, sizeof(*new_bkt));
+		bkt_prev->next = new_bkt;
+
+		/* Allocate new key & install. */
+		new_bkt_key_id = t->key_stack[--t->key_stack_tos];
+		bkt_key_install(t, new_bkt, entry, 0,
+				new_bkt_key_id, input_sig);
+		return 0;
+	}
+
+	CHECK(0, ENOSPC);
+}
+
+static int
+table_del(void *table, struct rte_swx_table_entry *entry)
+{
+	struct table *t = table;
+	struct bucket_extension *bkt0, *bkt, *bkt_prev;
+	uint32_t input_sig, bkt_id, i;
+
+	CHECK(t, EINVAL);
+	CHECK(entry, EINVAL);
+	CHECK(entry->key, EINVAL);
+
+	input_sig = hash(entry->key, t->key_mask, t->key_size, 0);
+	bkt_id = input_sig & (t->n_buckets - 1);
+	bkt0 = &t->buckets[bkt_id];
+	input_sig = (input_sig >> 16) | 1;
+
+	/* Key is present in the bucket. */
+	for (bkt = bkt0, bkt_prev = NULL; bkt; bkt_prev = bkt, bkt = bkt->next)
+		for (i = 0; i < KEYS_PER_BUCKET; i++)
+			if (bkt_keycmp(t, bkt, entry->key, i, input_sig)) {
+				/* Key free. */
+				bkt->sig[i] = 0;
+				t->key_stack[t->key_stack_tos++] =
+					bkt->key_id[i];
+
+				/* Bucket extension free if empty and not the
+				 * 1st in bucket.
+				 */
+				if (bkt_prev && bkt_is_empty(bkt)) {
+					bkt_prev->next = bkt->next;
+					bkt_id = bkt - t->buckets_ext;
+					t->bkt_ext_stack[t->bkt_ext_stack_tos++]
+						= bkt_id;
+				}
+
+				return 0;
+			}
+
+	return 0;
+}
+
+static uint64_t
+table_mailbox_size_get_unoptimized(void)
+{
+	return 0;
+}
+
+static int
+table_lookup_unoptimized(void *table,
+			 void *mailbox __rte_unused,
+			 uint8_t **key,
+			 uint64_t *action_id,
+			 uint8_t **action_data,
+			 int *hit)
+{
+	struct table *t = table;
+	struct bucket_extension *bkt0, *bkt;
+	uint8_t *input_key;
+	uint32_t input_sig, bkt_id, i;
+
+	input_key = &(*key)[t->params.key_offset];
+
+	input_sig = hash(input_key, t->key_mask, t->key_size, 0);
+	bkt_id = input_sig & (t->n_buckets - 1);
+	bkt0 = &t->buckets[bkt_id];
+	input_sig = (input_sig >> 16) | 1;
+
+	/* Key is present in the bucket. */
+	for (bkt = bkt0; bkt; bkt = bkt->next)
+		for (i = 0; i < KEYS_PER_BUCKET; i++)
+			if (bkt_keycmp(t, bkt, input_key, i, input_sig)) {
+				uint32_t bkt_key_id;
+				uint64_t *bkt_data;
+
+				/* Key. */
+				bkt_key_id = bkt->key_id[i];
+
+				/* Key data. */
+				bkt_data = table_key_data(t, bkt_key_id);
+				*action_id = bkt_data[0];
+				*action_data = (uint8_t *)&bkt_data[1];
+				*hit = 1;
+				return 1;
+			}
+
+	*hit = 0;
+	return 1;
+}
+
+struct mailbox {
+	struct bucket_extension *bkt;
+	uint32_t input_sig;
+	uint32_t bkt_key_id;
+	uint32_t sig_match;
+	uint32_t sig_match_many;
+	int state;
+};
+
+static uint64_t
+table_mailbox_size_get(void)
+{
+	return sizeof(struct mailbox);
+}
+
+/*
+ * mask = match bitmask
+ * match = at least one match
+ * match_many = more than one match
+ * match_pos = position of first match
+ *
+ *+------+-------+------------+-----------+
+ *| mask | match | match_many | match_pos |
+ *+------+-------+------------+-----------+
+ *| 0000 | 0     | 0          | 00        |
+ *| 0001 | 1     | 0          | 00        |
+ *| 0010 | 1     | 0          | 01        |
+ *| 0011 | 1     | 1          | 00        |
+ *+------+-------+------------+-----------+
+ *| 0100 | 1     | 0          | 10        |
+ *| 0101 | 1     | 1          | 00        |
+ *| 0110 | 1     | 1          | 01        |
+ *| 0111 | 1     | 1          | 00        |
+ *+------+-------+------------+-----------+
+ *| 1000 | 1     | 0          | 11        |
+ *| 1001 | 1     | 1          | 00        |
+ *| 1010 | 1     | 1          | 01        |
+ *| 1011 | 1     | 1          | 00        |
+ *+------+-------+------------+-----------+
+ *| 1100 | 1     | 1          | 10        |
+ *| 1101 | 1     | 1          | 00        |
+ *| 1110 | 1     | 1          | 01        |
+ *| 1111 | 1     | 1          | 00        |
+ *+------+-------+------------+-----------+
+ *
+ * match = 1111_1111_1111_1110 = 0xFFFE
+ * match_many = 1111_1110_1110_1000 = 0xFEE8
+ * match_pos = 0001_0010_0001_0011__0001_0010_0001_0000 = 0x12131210
+ *
+ */
+
+#define LUT_MATCH      0xFFFE
+#define LUT_MATCH_MANY 0xFEE8
+#define LUT_MATCH_POS  0x12131210
+
+static int
+table_lookup(void *table,
+	     void *mailbox,
+	     uint8_t **key,
+	     uint64_t *action_id,
+	     uint8_t **action_data,
+	     int *hit)
+{
+	struct table *t = table;
+	struct mailbox *m = mailbox;
+
+	switch (m->state) {
+	case 0: {
+		uint8_t *input_key = &(*key)[t->params.key_offset];
+		struct bucket_extension *bkt;
+		uint32_t input_sig, bkt_id;
+
+		input_sig = hash(input_key, t->key_mask, t->key_size, 0);
+		bkt_id = input_sig & (t->n_buckets - 1);
+		bkt = &t->buckets[bkt_id];
+		rte_prefetch0(bkt);
+
+		m->bkt = bkt;
+		m->input_sig = (input_sig >> 16) | 1;
+		m->state++;
+		return 0;
+	}
+
+	case 1: {
+		struct bucket_extension *bkt = m->bkt;
+		uint32_t input_sig = m->input_sig;
+		uint32_t bkt_sig0, bkt_sig1, bkt_sig2, bkt_sig3;
+		uint32_t mask0 = 0, mask1 = 0, mask2 = 0, mask3 = 0, mask_all;
+		uint32_t sig_match = LUT_MATCH;
+		uint32_t sig_match_many = LUT_MATCH_MANY;
+		uint32_t sig_match_pos = LUT_MATCH_POS;
+		uint32_t bkt_key_id;
+
+		bkt_sig0 = input_sig ^ bkt->sig[0];
+		if (bkt_sig0)
+			mask0 = 1 << 0;
+
+		bkt_sig1 = input_sig ^ bkt->sig[1];
+		if (bkt_sig1)
+			mask1 = 1 << 1;
+
+		bkt_sig2 = input_sig ^ bkt->sig[2];
+		if (bkt_sig2)
+			mask2 = 1 << 2;
+
+		bkt_sig3 = input_sig ^ bkt->sig[3];
+		if (bkt_sig3)
+			mask3 = 1 << 3;
+
+		mask_all = (mask0 | mask1) | (mask2 | mask3);
+		sig_match = (sig_match >> mask_all) & 1;
+		sig_match_many = (sig_match_many >> mask_all) & 1;
+		sig_match_pos = (sig_match_pos >> (mask_all << 1)) & 3;
+
+		bkt_key_id = bkt->key_id[sig_match_pos];
+		rte_prefetch0(table_key(t, bkt_key_id));
+		rte_prefetch0(table_key_data(t, bkt_key_id));
+
+		m->bkt_key_id = bkt_key_id;
+		m->sig_match = sig_match;
+		m->sig_match_many = sig_match_many;
+		m->state++;
+		return 0;
+	}
+
+	case 2: {
+		uint8_t *input_key = &(*key)[t->params.key_offset];
+		struct bucket_extension *bkt = m->bkt;
+		uint32_t bkt_key_id = m->bkt_key_id;
+		uint8_t *bkt_key = table_key(t, bkt_key_id);
+		uint64_t *bkt_data = table_key_data(t, bkt_key_id);
+		uint32_t lkp_hit;
+
+		lkp_hit = keycmp(bkt_key, input_key, t->key_mask, t->key_size);
+		lkp_hit &= m->sig_match;
+		*action_id = bkt_data[0];
+		*action_data = (uint8_t *)&bkt_data[1];
+		*hit = lkp_hit;
+
+		m->state = 0;
+
+		if (!lkp_hit && (m->sig_match_many || bkt->next))
+			return table_lookup_unoptimized(t,
+							m,
+							key,
+							action_id,
+							action_data,
+							hit);
+
+		return 1;
+	}
+
+	default:
+		return 0;
+	}
+}
+
+static void *
+table_create(struct rte_swx_table_params *params,
+	     struct rte_swx_table_entry_list *entries,
+	     const char *args,
+	     int numa_node)
+{
+	struct table *t;
+	struct rte_swx_table_entry *entry;
+	int status;
+
+	/* Table create. */
+	status = __table_create(&t, NULL, params, args, numa_node);
+	if (status)
+		return NULL;
+
+	/* Table add entries. */
+	if (!entries)
+		return t;
+
+	TAILQ_FOREACH(entry, entries, node) {
+		int status;
+
+		status = table_add(t, entry);
+		if (status) {
+			table_free(t);
+			return NULL;
+		}
+	}
+
+	return t;
+}
+
+static uint64_t
+table_footprint(struct rte_swx_table_params *params,
+		struct rte_swx_table_entry_list *entries __rte_unused,
+		const char *args)
+{
+	uint64_t memory_footprint;
+	int status;
+
+	status = __table_create(NULL, &memory_footprint, params, args, 0);
+	if (status)
+		return 0;
+
+	return memory_footprint;
+}
+
+struct rte_swx_table_ops rte_swx_table_exact_match_unoptimized_ops = {
+	.footprint_get = table_footprint,
+	.mailbox_size_get = table_mailbox_size_get_unoptimized,
+	.create = table_create,
+	.add = table_add,
+	.del = table_del,
+	.lkp = table_lookup_unoptimized,
+	.free = table_free,
+};
+
+struct rte_swx_table_ops rte_swx_table_exact_match_ops = {
+	.footprint_get = table_footprint,
+	.mailbox_size_get = table_mailbox_size_get,
+	.create = table_create,
+	.add = table_add,
+	.del = table_del,
+	.lkp = table_lookup,
+	.free = table_free,
+};
diff --git a/lib/librte_table/rte_swx_table_em.h b/lib/librte_table/rte_swx_table_em.h
new file mode 100644
index 000000000..909ada483
--- /dev/null
+++ b/lib/librte_table/rte_swx_table_em.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright(c) 2020 Intel Corporation
+ */
+#ifndef __INCLUDE_RTE_SWX_TABLE_EM_H__
+#define __INCLUDE_RTE_SWX_TABLE_EM_H__
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * @file
+ * RTE SWX Exact Match Table
+ */
+
+#include <stdint.h>
+
+#include <rte_swx_table.h>
+
+/** Exact match table operations - unoptimized. */
+extern struct rte_swx_table_ops rte_swx_table_exact_match_unoptimized_ops;
+
+/** Exact match table operations. */
+extern struct rte_swx_table_ops rte_swx_table_exact_match_ops;
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
diff --git a/lib/librte_table/rte_table_version.map b/lib/librte_table/rte_table_version.map
index 568a6c6a8..81c554b63 100644
--- a/lib/librte_table/rte_table_version.map
+++ b/lib/librte_table/rte_table_version.map
@@ -18,3 +18,10 @@ DPDK_21 {
 
 	local: *;
 };
+
+EXPERIMENTAL {
+	global:
+
+	rte_swx_table_exact_match_unoptimized_ops;
+	rte_swx_table_exact_match_ops;
+};
-- 
2.17.1