DPDK  16.04.0
rte_lpm_sse.h
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 
42 #ifdef __cplusplus
43 extern "C" {
44 #endif
45 
46 static inline void
47 rte_lpm_lookupx4(const struct rte_lpm *lpm, xmm_t ip, uint32_t hop[4],
48  uint32_t defv)
49 {
50  __m128i i24;
51  rte_xmm_t i8;
52  uint32_t tbl[4];
53  uint64_t idx, pt, pt2;
54  const uint32_t *ptbl;
55 
56  const __m128i mask8 =
57  _mm_set_epi32(UINT8_MAX, UINT8_MAX, UINT8_MAX, UINT8_MAX);
58 
59  /*
60  * RTE_LPM_VALID_EXT_ENTRY_BITMASK for 2 LPM entries
61  * as one 64-bit value (0x0300000003000000).
62  */
63  const uint64_t mask_xv =
64  ((uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK |
65  (uint64_t)RTE_LPM_VALID_EXT_ENTRY_BITMASK << 32);
66 
67  /*
68  * RTE_LPM_LOOKUP_SUCCESS for 2 LPM entries
69  * as one 64-bit value (0x0100000001000000).
70  */
71  const uint64_t mask_v =
72  ((uint64_t)RTE_LPM_LOOKUP_SUCCESS |
73  (uint64_t)RTE_LPM_LOOKUP_SUCCESS << 32);
74 
75  /* get 4 indexes for tbl24[]. */
76  i24 = _mm_srli_epi32(ip, CHAR_BIT);
77 
78  /* extract values from tbl24[] */
79  idx = _mm_cvtsi128_si64(i24);
80  i24 = _mm_srli_si128(i24, sizeof(uint64_t));
81 
82  ptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];
83  tbl[0] = *ptbl;
84  ptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];
85  tbl[1] = *ptbl;
86 
87  idx = _mm_cvtsi128_si64(i24);
88 
89  ptbl = (const uint32_t *)&lpm->tbl24[(uint32_t)idx];
90  tbl[2] = *ptbl;
91  ptbl = (const uint32_t *)&lpm->tbl24[idx >> 32];
92  tbl[3] = *ptbl;
93 
94  /* get 4 indexes for tbl8[]. */
95  i8.x = _mm_and_si128(ip, mask8);
96 
97  pt = (uint64_t)tbl[0] |
98  (uint64_t)tbl[1] << 32;
99  pt2 = (uint64_t)tbl[2] |
100  (uint64_t)tbl[3] << 32;
101 
102  /* search successfully finished for all 4 IP addresses. */
103  if (likely((pt & mask_xv) == mask_v) &&
104  likely((pt2 & mask_xv) == mask_v)) {
105  *(uint64_t *)hop = pt & RTE_LPM_MASKX4_RES;
106  *(uint64_t *)(hop + 2) = pt2 & RTE_LPM_MASKX4_RES;
107  return;
108  }
109 
110  if (unlikely((pt & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
111  RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
112  i8.u32[0] = i8.u32[0] +
113  (uint8_t)tbl[0] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
114  ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[0]];
115  tbl[0] = *ptbl;
116  }
117  if (unlikely((pt >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
118  RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
119  i8.u32[1] = i8.u32[1] +
120  (uint8_t)tbl[1] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
121  ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[1]];
122  tbl[1] = *ptbl;
123  }
124  if (unlikely((pt2 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
125  RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
126  i8.u32[2] = i8.u32[2] +
127  (uint8_t)tbl[2] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
128  ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[2]];
129  tbl[2] = *ptbl;
130  }
131  if (unlikely((pt2 >> 32 & RTE_LPM_VALID_EXT_ENTRY_BITMASK) ==
132  RTE_LPM_VALID_EXT_ENTRY_BITMASK)) {
133  i8.u32[3] = i8.u32[3] +
134  (uint8_t)tbl[3] * RTE_LPM_TBL8_GROUP_NUM_ENTRIES;
135  ptbl = (const uint32_t *)&lpm->tbl8[i8.u32[3]];
136  tbl[3] = *ptbl;
137  }
138 
139  hop[0] = (tbl[0] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[0] & 0x00FFFFFF : defv;
140  hop[1] = (tbl[1] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[1] & 0x00FFFFFF : defv;
141  hop[2] = (tbl[2] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[2] & 0x00FFFFFF : defv;
142  hop[3] = (tbl[3] & RTE_LPM_LOOKUP_SUCCESS) ? tbl[3] & 0x00FFFFFF : defv;
143 }
144 
145 #ifdef __cplusplus
146 }
147 #endif
148 
149 #endif /* _RTE_LPM_SSE_H_ */