1/*-
2 *   BSD LICENSE
3 *
4 *   Copyright(c) 2010-2014 Intel Corporation. All rights reserved.
5 *   All rights reserved.
6 *
7 *   Redistribution and use in source and binary forms, with or without
8 *   modification, are permitted provided that the following conditions
9 *   are met:
10 *
11 *     * Redistributions of source code must retain the above copyright
12 *       notice, this list of conditions and the following disclaimer.
13 *     * Redistributions in binary form must reproduce the above copyright
14 *       notice, this list of conditions and the following disclaimer in
15 *       the documentation and/or other materials provided with the
16 *       distribution.
17 *     * Neither the name of Intel Corporation nor the names of its
18 *       contributors may be used to endorse or promote products derived
19 *       from this software without specific prior written permission.
20 *
21 *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
22 *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
23 *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
24 *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
25 *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
26 *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
27 *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
28 *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
29 *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
30 *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
31 *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
32 */
33
34#ifndef _RTE_LPM_SSE_H_
35#define _RTE_LPM_SSE_H_
36
37#include <rte_branch_prediction.h>
38#include <rte_byteorder.h>
39#include <rte_common.h>
40#include <rte_vect.h>
41#include <rte_lpm.h>
42
43#ifdef __cplusplus
44extern "C" {
45#endif
46
47static inline void
48rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],
49	uint32_t defv)
50{
51	__m128i i24;
52	rte_xmm_t i8;
53	uint32_t tbl[4];
54	uint64_t idx, pt, pt2;
55	const uint32_t *ptbl;
56
57	const __m128i mask8 =
58		_mm_set_epi32(UINT8_MAX, UINT8_MAX, UINT8_MAX, UINT8_MAX);
59
60	/*
61	 * RTE_LPM_VALID_EXT_ENTRY_BITMASK for 2 LPM entries
62	 * as one 64-bit value (0x0300000003000000).
63	 */
64	const uint64_t mask_xv =
65		((uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK |
66		(uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK << 32);
67
68	/*
69	 * RTE_LPM_LOOKUP_SUCCESS for 2 LPM entries
70	 * as one 64-bit value (0x0100000001000000).
71	 */
72	const uint64_t mask_v =
73		((uint64_t)RTE_LPM_LOOKUP_SUCCESS |
74		(uint64_t)RTE_LPM_LOOKUP_SUCCESS << 32);
75
76	/* get 4 indexes for tbl24[]. */
77	i24 = _mm_srli_epi32(ip, CHAR_BIT);
78
79	/* extract values from tbl24[] */
80	idx = _mm_cvtsi128_si64(i24);
81	i24 = _mm_srli_si128(i24, sizeof(uint64_t));
82
83	ptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];
84	tbl[0] = *ptbl;
85	ptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];
86	tbl[1] = *ptbl;
87
88	idx = _mm_cvtsi128_si64(i24);
89
90	ptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];
91	tbl[2] = *ptbl;
92	ptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];
93	tbl[3] = *ptbl;
94
95	/* get 4 indexes for tbl8[]. */
96	i8.x = _mm_and_si128(ip, mask8);
97
98	pt = (uint64_t)tbl[0] |
99		(uint64_t)tbl[1] << 32;
100	pt2 = (uint64_t)tbl[2] |
101		(uint64_t)tbl[3] << 32;
102
103	/* search successfully finished for all 4 IP addresses. */
104	if (likely((pt & mask_xv) == mask_v) &&
105			likely((pt2 & mask_xv) == mask_v)) {
106		*(uint64_t *)hop = pt & RTE_LPM_MASKX4_RES;
107		*(uint64_t *)(hop + 2) = pt2 & RTE_LPM_MASKX4_RES;
108		return;
109	}
110
111	if (unlikely((pt & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
112			RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
113		i8.u32[0] = i8.u32[0] +
114			(uint8_t)tbl[0] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
115		ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[0]];
116		tbl[0] = *ptbl;
117	}
118	if (unlikely((pt >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
119			RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
120		i8.u32[1] = i8.u32[1] +
121			(uint8_t)tbl[1] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
122		ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[1]];
123		tbl[1] = *ptbl;
124	}
125	if (unlikely((pt2 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
126			RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
127		i8.u32[2] = i8.u32[2] +
128			(uint8_t)tbl[2] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
129		ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[2]];
130		tbl[2] = *ptbl;
131	}
132	if (unlikely((pt2 >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
133			RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
134		i8.u32[3] = i8.u32[3] +
135			(uint8_t)tbl[3] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
136		ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[3]];
137		tbl[3] = *ptbl;
138	}
139
140	hop[0] = (tbl[0] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[0] & 0x00FFFFFF : defv;
141	hop[1] = (tbl[1] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[1] & 0x00FFFFFF : defv;
142	hop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[2] & 0x00FFFFFF : defv;
143	hop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF : defv;
144}
145
146#ifdef __cplusplus
147}
148#endif
149
150#endif /* _RTE_LPM_SSE_H_ */
151