| 1 |
/* This file contains input data static definitions, shared by most of |
| 2 |
the tests. */ |
| 3 |
|
| 4 |
#include <arm_neon.h> |
| 5 |
#include "arm-neon-ref.h" |
| 6 |
|
| 7 |
/* Initialization helpers; 4 slices are needed for vld2, vld3 and |
| 8 |
vld4. */ |
| 9 |
#define MY_INIT_TABLE(T,W,N) xNAME(INIT_TABLE,N)(T##W##_t) |
| 10 |
#define MY_INIT_TABLE2(T,W,N) xNAME(INIT_TABLE2,N)(T##W##_t) |
| 11 |
#define MY_INIT_TABLE3(T,W,N) xNAME(INIT_TABLE3,N)(T##W##_t) |
| 12 |
#define MY_INIT_TABLE4(T,W,N) xNAME(INIT_TABLE4,N)(T##W##_t) |
| 13 |
|
| 14 |
/* Initialized input buffers. */ |
| 15 |
#define VECT_VAR_DECL_INIT(V, T, W, N) \ |
| 16 |
VECT_VAR_DECL(V,T,W,N) [] = { MY_INIT_TABLE(T,W,N) } |
| 17 |
|
| 18 |
/* Specialized initializer with 4 entries, as used by vldX_dup and |
| 19 |
vdup tests, which iterate 4 times on input buffers. */ |
| 20 |
#define VECT_VAR_DECL_INIT4(V, T, W, N) \ |
| 21 |
VECT_VAR_DECL(V,T,W,N) [] = { MY_INIT_TABLE(T,W,4) }; |
| 22 |
|
| 23 |
/* Initializers for arrays of vectors. */ |
| 24 |
#define VECT_ARRAY_INIT2(V, T, W, N) \ |
| 25 |
T##W##_t VECT_ARRAY_VAR(V,T,W,N,2)[] = \ |
| 26 |
{ MY_INIT_TABLE(T,W,N) \ |
| 27 |
MY_INIT_TABLE2(T,W,N) } |
| 28 |
|
| 29 |
#define VECT_ARRAY_INIT3(V, T, W, N) \ |
| 30 |
T##W##_t VECT_ARRAY_VAR(V,T,W,N,3)[] = \ |
| 31 |
{ MY_INIT_TABLE(T,W,N) \ |
| 32 |
MY_INIT_TABLE2(T,W,N) \ |
| 33 |
MY_INIT_TABLE3(T,W,N) } |
| 34 |
|
| 35 |
#define VECT_ARRAY_INIT4(V, T, W, N) \ |
| 36 |
T##W##_t VECT_ARRAY_VAR(V,T,W,N,4)[] = \ |
| 37 |
{ MY_INIT_TABLE(T,W,N) \ |
| 38 |
MY_INIT_TABLE2(T,W,N) \ |
| 39 |
MY_INIT_TABLE3(T,W,N) \ |
| 40 |
MY_INIT_TABLE4(T,W,N) } |
| 41 |
|
| 42 |
/* Sample initialization vectors. */ |
| 43 |
#define INIT_TABLE_1(T) \ |
| 44 |
(T)-16, |
| 45 |
#define INIT_TABLE2_1(T) \ |
| 46 |
(T)-15, |
| 47 |
#define INIT_TABLE3_1(T) \ |
| 48 |
(T)-14, |
| 49 |
#define INIT_TABLE4_1(T) \ |
| 50 |
(T)-13, |
| 51 |
|
| 52 |
#define INIT_TABLE_2(T) \ |
| 53 |
(T)-16, (T)-15, |
| 54 |
#define INIT_TABLE2_2(T) \ |
| 55 |
(T)-14, (T)-13, |
| 56 |
#define INIT_TABLE3_2(T) \ |
| 57 |
(T)-12, (T)-11, |
| 58 |
#define INIT_TABLE4_2(T) \ |
| 59 |
(T)-10, (T)-9, |
| 60 |
|
| 61 |
/* Initializer for vld3_lane tests. */ |
| 62 |
#define INIT_TABLE_3(T) \ |
| 63 |
(T)-16, (T)-15, (T)-14, |
| 64 |
|
| 65 |
#define INIT_TABLE_4(T) \ |
| 66 |
(T)-16, (T)-15, (T)-14, (T)-13, |
| 67 |
#define INIT_TABLE2_4(T) \ |
| 68 |
(T)-12, (T)-11, (T)-10, (T)-9, |
| 69 |
#define INIT_TABLE3_4(T) \ |
| 70 |
(T)-8, (T)-7, (T)-6, (T)-5, |
| 71 |
#define INIT_TABLE4_4(T) \ |
| 72 |
(T)-4, (T)-3, (T)-2, (T)-1, |
| 73 |
|
| 74 |
#define INIT_TABLE_8(T) \ |
| 75 |
(T)-16, (T)-15, (T)-14, (T)-13, (T)-12, (T)-11, (T)-10, (T)-9, |
| 76 |
#define INIT_TABLE2_8(T) \ |
| 77 |
(T)-8, (T)-7, (T)-6, (T)-5, (T)-4, (T)-3, (T)-2, (T)-1, |
| 78 |
#define INIT_TABLE3_8(T) \ |
| 79 |
(T)0, (T)1, (T)2, (T)3, (T)4, (T)5, (T)6, (T)7, |
| 80 |
#define INIT_TABLE4_8(T) \ |
| 81 |
(T)8, (T)9, (T)10, (T)11, (T)12, (T)13, (T)14, (T)15, |
| 82 |
|
| 83 |
#define INIT_TABLE_16(T) \ |
| 84 |
(T)-16, (T)-15, (T)-14, (T)-13, (T)-12, (T)-11, (T)-10, (T)-9, \ |
| 85 |
(T)-8, (T)-7, (T)-6, (T)-5, (T)-4, (T)-3, (T)-2, (T)-1, |
| 86 |
#define INIT_TABLE2_16(T) \ |
| 87 |
(T)0, (T)1, (T)2, (T)3, (T)4, (T)5, (T)6, (T)7, \ |
| 88 |
(T)8, (T)9, (T)10, (T)11, (T)12, (T)13, (T)14, (T)15, |
| 89 |
#define INIT_TABLE3_16(T) \ |
| 90 |
(T)16, (T)17, (T)18, (T)19, (T)20, (T)21, (T)22, (T)23, \ |
| 91 |
(T)24, (T)25, (T)26, (T)27, (T)28, (T)29, (T)30, (T)31, |
| 92 |
#define INIT_TABLE4_16(T) \ |
| 93 |
(T)32, (T)33, (T)34, (T)35, (T)36, (T)37, (T)38, (T)39, \ |
| 94 |
(T)40, (T)41, (T)42, (T)43, (T)44, (T)45, (T)46, (T)47, |
| 95 |
|
| 96 |
/* This one is used for padding between input buffers. */ |
| 97 |
#define PAD(V, T, W, N) char VECT_VAR(V,T,W,N)=42 |
| 98 |
|
| 99 |
/* Input buffers, one of each size. */ |
| 100 |
/* Insert some padding to try to exhibit out of bounds accesses. */ |
| 101 |
VECT_VAR_DECL_INIT(buffer, int, 8, 8); |
| 102 |
PAD(buffer_pad, int, 8, 8); |
| 103 |
VECT_VAR_DECL_INIT(buffer, int, 16, 4); |
| 104 |
PAD(buffer_pad, int, 16, 4); |
| 105 |
VECT_VAR_DECL_INIT(buffer, int, 32, 2); |
| 106 |
PAD(buffer_pad, int, 32, 2); |
| 107 |
VECT_VAR_DECL_INIT(buffer, int, 64, 1); |
| 108 |
PAD(buffer_pad, int, 64, 1); |
| 109 |
VECT_VAR_DECL_INIT(buffer, uint, 8, 8); |
| 110 |
PAD(buffer_pad, uint, 8, 8); |
| 111 |
VECT_VAR_DECL_INIT(buffer, poly, 8, 8); |
| 112 |
PAD(buffer_pad, poly, 8, 8); |
| 113 |
VECT_VAR_DECL_INIT(buffer, poly, 16, 4); |
| 114 |
PAD(buffer_pad, poly, 16, 4); |
| 115 |
VECT_VAR_DECL_INIT(buffer, uint, 16, 4); |
| 116 |
PAD(buffer_pad, uint, 16, 4); |
| 117 |
VECT_VAR_DECL_INIT(buffer, uint, 32, 2); |
| 118 |
PAD(buffer_pad, uint, 32, 2); |
| 119 |
VECT_VAR_DECL_INIT(buffer, uint, 64, 1); |
| 120 |
PAD(buffer_pad, uint, 64, 1); |
| 121 |
#if defined (__ARM_FEATURE_CRYPTO) |
| 122 |
VECT_VAR_DECL_INIT(buffer, poly, 64, 1); |
| 123 |
PAD(buffer_pad, poly, 64, 1); |
| 124 |
#endif |
| 125 |
#if MFLOAT8_SUPPORTED |
| 126 |
VECT_VAR_DECL(buffer, mfloat, 8, 8)[8]; |
| 127 |
PAD(buffer_pad, mfloat, 8, 8); |
| 128 |
#endif |
| 129 |
#if defined (__ARM_FP16_FORMAT_IEEE) || defined (__ARM_FP16_FORMAT_ALTERNATIVE) |
| 130 |
VECT_VAR_DECL_INIT(buffer, float, 16, 4); |
| 131 |
PAD(buffer_pad, float, 16, 4); |
| 132 |
#endif |
| 133 |
VECT_VAR_DECL_INIT(buffer, float, 32, 2); |
| 134 |
PAD(buffer_pad, float, 32, 2); |
| 135 |
VECT_VAR_DECL_INIT(buffer, int, 8, 16); |
| 136 |
PAD(buffer_pad, int, 8, 16); |
| 137 |
VECT_VAR_DECL_INIT(buffer, int, 16, 8); |
| 138 |
PAD(buffer_pad, int, 16, 8); |
| 139 |
VECT_VAR_DECL_INIT(buffer, int, 32, 4); |
| 140 |
PAD(buffer_pad, int, 32, 4); |
| 141 |
VECT_VAR_DECL_INIT(buffer, int, 64, 2); |
| 142 |
PAD(buffer_pad, int, 64, 2); |
| 143 |
VECT_VAR_DECL_INIT(buffer, uint, 8, 16); |
| 144 |
PAD(buffer_pad, uint, 8, 16); |
| 145 |
VECT_VAR_DECL_INIT(buffer, uint, 16, 8); |
| 146 |
PAD(buffer_pad, uint, 16, 8); |
| 147 |
VECT_VAR_DECL_INIT(buffer, uint, 32, 4); |
| 148 |
PAD(buffer_pad, uint, 32, 4); |
| 149 |
VECT_VAR_DECL_INIT(buffer, uint, 64, 2); |
| 150 |
PAD(buffer_pad, uint, 64, 2); |
| 151 |
VECT_VAR_DECL_INIT(buffer, poly, 8, 16); |
| 152 |
PAD(buffer_pad, poly, 8, 16); |
| 153 |
VECT_VAR_DECL_INIT(buffer, poly, 16, 8); |
| 154 |
PAD(buffer_pad, poly, 16, 8); |
| 155 |
#if defined (__ARM_FEATURE_CRYPTO) |
| 156 |
VECT_VAR_DECL_INIT(buffer, poly, 64, 2); |
| 157 |
PAD(buffer_pad, poly, 64, 2); |
| 158 |
#endif |
| 159 |
#if MFLOAT8_SUPPORTED |
| 160 |
VECT_VAR_DECL(buffer, mfloat, 8, 16)[16]; |
| 161 |
PAD(buffer_pad, mfloat, 8, 16); |
| 162 |
#endif |
| 163 |
#if defined (__ARM_FP16_FORMAT_IEEE) || defined (__ARM_FP16_FORMAT_ALTERNATIVE) |
| 164 |
VECT_VAR_DECL_INIT(buffer, float, 16, 8); |
| 165 |
PAD(buffer_pad, float, 16, 8); |
| 166 |
#endif |
| 167 |
VECT_VAR_DECL_INIT(buffer, float, 32, 4); |
| 168 |
PAD(buffer_pad, float, 32, 4); |
| 169 |
#ifdef __aarch64__ |
| 170 |
VECT_VAR_DECL_INIT(buffer, float, 64, 2); |
| 171 |
PAD(buffer_pad, float, 64, 2); |
| 172 |
#endif |
| 173 |
|
| 174 |
/* The tests for vld1_dup and vdup expect at least 4 entries in the |
| 175 |
input buffer, so force 1- and 2-elements initializers to have 4 |
| 176 |
entries (using VECT_VAR_DECL_INIT4). */ |
| 177 |
VECT_VAR_DECL_INIT(buffer_dup, int, 8, 8); |
| 178 |
VECT_VAR_DECL(buffer_dup_pad, int, 8, 8); |
| 179 |
VECT_VAR_DECL_INIT(buffer_dup, int, 16, 4); |
| 180 |
VECT_VAR_DECL(buffer_dup_pad, int, 16, 4); |
| 181 |
VECT_VAR_DECL_INIT4(buffer_dup, int, 32, 2); |
| 182 |
VECT_VAR_DECL(buffer_dup_pad, int, 32, 2); |
| 183 |
VECT_VAR_DECL_INIT4(buffer_dup, int, 64, 1); |
| 184 |
VECT_VAR_DECL(buffer_dup_pad, int, 64, 1); |
| 185 |
VECT_VAR_DECL_INIT(buffer_dup, uint, 8, 8); |
| 186 |
VECT_VAR_DECL(buffer_dup_pad, uint, 8, 8); |
| 187 |
VECT_VAR_DECL_INIT(buffer_dup, uint, 16, 4); |
| 188 |
VECT_VAR_DECL(buffer_dup_pad, uint, 16, 4); |
| 189 |
VECT_VAR_DECL_INIT4(buffer_dup, uint, 32, 2); |
| 190 |
VECT_VAR_DECL(buffer_dup_pad, uint, 32, 2); |
| 191 |
VECT_VAR_DECL_INIT4(buffer_dup, uint, 64, 1); |
| 192 |
VECT_VAR_DECL(buffer_dup_pad, uint, 64, 1); |
| 193 |
VECT_VAR_DECL_INIT(buffer_dup, poly, 8, 8); |
| 194 |
VECT_VAR_DECL(buffer_dup_pad, poly, 8, 8); |
| 195 |
VECT_VAR_DECL_INIT(buffer_dup, poly, 16, 4); |
| 196 |
VECT_VAR_DECL(buffer_dup_pad, poly, 16, 4); |
| 197 |
#if defined (__ARM_FEATURE_CRYPTO) |
| 198 |
VECT_VAR_DECL_INIT4(buffer_dup, poly, 64, 1); |
| 199 |
VECT_VAR_DECL(buffer_dup_pad, poly, 64, 1); |
| 200 |
#endif |
| 201 |
#if MFLOAT8_SUPPORTED |
| 202 |
VECT_VAR_DECL(buffer_dup, mfloat, 8, 8)[8]; |
| 203 |
PAD(buffer_dup_pad, mfloat, 8, 8); |
| 204 |
#endif |
| 205 |
#if defined (__ARM_FP16_FORMAT_IEEE) || defined (__ARM_FP16_FORMAT_ALTERNATIVE) |
| 206 |
VECT_VAR_DECL_INIT4(buffer_dup, float, 16, 4); |
| 207 |
VECT_VAR_DECL(buffer_dup_pad, float, 16, 4); |
| 208 |
#endif |
| 209 |
VECT_VAR_DECL_INIT4(buffer_dup, float, 32, 2); |
| 210 |
VECT_VAR_DECL(buffer_dup_pad, float, 32, 2); |
| 211 |
|
| 212 |
VECT_VAR_DECL_INIT(buffer_dup, int, 8, 16); |
| 213 |
VECT_VAR_DECL(buffer_dup_pad, int, 8, 16); |
| 214 |
VECT_VAR_DECL_INIT(buffer_dup, int, 16, 8); |
| 215 |
VECT_VAR_DECL(buffer_dup_pad, int, 16, 8); |
| 216 |
VECT_VAR_DECL_INIT(buffer_dup, int, 32, 4); |
| 217 |
VECT_VAR_DECL(buffer_dup_pad, int, 32, 4); |
| 218 |
VECT_VAR_DECL_INIT4(buffer_dup, int, 64, 2); |
| 219 |
VECT_VAR_DECL(buffer_dup_pad, int, 64, 2); |
| 220 |
VECT_VAR_DECL_INIT(buffer_dup, uint, 8, 16); |
| 221 |
VECT_VAR_DECL(buffer_dup_pad, uint, 8, 16); |
| 222 |
VECT_VAR_DECL_INIT(buffer_dup, uint, 16, 8); |
| 223 |
VECT_VAR_DECL(buffer_dup_pad, uint, 16, 8); |
| 224 |
VECT_VAR_DECL_INIT(buffer_dup, uint, 32, 4); |
| 225 |
VECT_VAR_DECL(buffer_dup_pad, uint, 32, 4); |
| 226 |
VECT_VAR_DECL_INIT4(buffer_dup, uint, 64, 2); |
| 227 |
VECT_VAR_DECL(buffer_dup_pad, uint, 64, 2); |
| 228 |
VECT_VAR_DECL_INIT(buffer_dup, poly, 8, 16); |
| 229 |
VECT_VAR_DECL(buffer_dup_pad, poly, 8, 16); |
| 230 |
VECT_VAR_DECL_INIT(buffer_dup, poly, 16, 8); |
| 231 |
VECT_VAR_DECL(buffer_dup_pad, poly, 16, 8); |
| 232 |
#if defined (__ARM_FEATURE_CRYPTO) |
| 233 |
VECT_VAR_DECL_INIT4(buffer_dup, poly, 64, 2); |
| 234 |
VECT_VAR_DECL(buffer_dup_pad, poly, 64, 2); |
| 235 |
#endif |
| 236 |
#if MFLOAT8_SUPPORTED |
| 237 |
VECT_VAR_DECL(buffer_dup, mfloat, 8, 16)[16]; |
| 238 |
PAD(buffer_dup_pad, mfloat, 8, 16); |
| 239 |
#endif |
| 240 |
#if defined (__ARM_FP16_FORMAT_IEEE) || defined (__ARM_FP16_FORMAT_ALTERNATIVE) |
| 241 |
VECT_VAR_DECL_INIT(buffer_dup, float, 16, 8); |
| 242 |
VECT_VAR_DECL(buffer_dup_pad, float, 16, 8); |
| 243 |
#endif |
| 244 |
VECT_VAR_DECL_INIT(buffer_dup, float, 32, 4); |
| 245 |
VECT_VAR_DECL(buffer_dup_pad, float, 32, 4); |
| 246 |
|
| 247 |
#if MFLOAT8_SUPPORTED |
| 248 |
static void __attribute__((constructor)) |
| 249 |
copy_mfloat8 () |
| 250 |
{ |
| 251 |
memcpy (VECT_VAR(buffer, mfloat, 8, 8), VECT_VAR(buffer, uint, 8, 8), 8); |
| 252 |
memcpy (VECT_VAR(buffer, mfloat, 8, 16), VECT_VAR(buffer, uint, 8, 16), 16); |
| 253 |
memcpy (VECT_VAR(buffer_dup, mfloat, 8, 8), |
| 254 |
VECT_VAR(buffer_dup, uint, 8, 8), 8); |
| 255 |
memcpy (VECT_VAR(buffer_dup, mfloat, 8, 16), |
| 256 |
VECT_VAR(buffer_dup, uint, 8, 16), 16); |
| 257 |
} |
| 258 |
#endif |