| 1 | // Copyright (c) 2012, the Dart project authors. Please see the AUTHORS file |
| 2 | // for details. All rights reserved. Use of this source code is governed by a |
| 3 | // BSD-style license that can be found in the LICENSE file. |
| 4 | |
| 5 | #ifndef RUNTIME_PLATFORM_UTILS_H_ |
| 6 | #define RUNTIME_PLATFORM_UTILS_H_ |
| 7 | |
| 8 | #include <limits> |
| 9 | #include <memory> |
| 10 | #include <type_traits> |
| 11 | |
| 12 | #include "platform/assert.h" |
| 13 | #include "platform/globals.h" |
| 14 | |
| 15 | namespace dart { |
| 16 | |
| 17 | class Utils { |
| 18 | public: |
| 19 | template <typename T> |
| 20 | static inline T Minimum(T x, T y) { |
| 21 | return x < y ? x : y; |
| 22 | } |
| 23 | |
| 24 | template <typename T> |
| 25 | static constexpr inline T Maximum(T x, T y) { |
| 26 | return x > y ? x : y; |
| 27 | } |
| 28 | |
| 29 | // Calculates absolute value of a given signed integer. |
| 30 | // `x` must not be equal to minimum value representable by `T` |
| 31 | // as its absolute value is out of range. |
| 32 | template <typename T> |
| 33 | static inline T Abs(T x) { |
| 34 | // Note: as a general rule, it is not OK to use STL in Dart VM. |
| 35 | // However, std::numeric_limits<T>::min() and max() are harmless |
| 36 | // and worthwhile exception from this rule. |
| 37 | ASSERT(x != std::numeric_limits<T>::min()); |
| 38 | if (x < 0) return -x; |
| 39 | return x; |
| 40 | } |
| 41 | |
| 42 | // Calculates absolute value of a given signed integer with saturation. |
| 43 | // If `x` equals to minimum value representable by `T`, then |
| 44 | // absolute value is saturated to the maximum value representable by `T`. |
| 45 | template <typename T> |
| 46 | static inline T AbsWithSaturation(T x) { |
| 47 | if (x < 0) { |
| 48 | // Note: as a general rule, it is not OK to use STL in Dart VM. |
| 49 | // However, std::numeric_limits<T>::min() and max() are harmless |
| 50 | // and worthwhile exception from this rule. |
| 51 | if (x == std::numeric_limits<T>::min()) { |
| 52 | return std::numeric_limits<T>::max(); |
| 53 | } |
| 54 | return -x; |
| 55 | } |
| 56 | return x; |
| 57 | } |
| 58 | |
| 59 | template <typename T> |
| 60 | static inline bool IsPowerOfTwo(T x) { |
| 61 | return ((x & (x - 1)) == 0) && (x != 0); |
| 62 | } |
| 63 | |
| 64 | template <typename T> |
| 65 | static inline int ShiftForPowerOfTwo(T x) { |
| 66 | ASSERT(IsPowerOfTwo(x)); |
| 67 | int num_shifts = 0; |
| 68 | while (x > 1) { |
| 69 | num_shifts++; |
| 70 | x = x >> 1; |
| 71 | } |
| 72 | return num_shifts; |
| 73 | } |
| 74 | |
| 75 | template <typename T> |
| 76 | static inline bool IsAligned(T x, intptr_t n) { |
| 77 | ASSERT(IsPowerOfTwo(n)); |
| 78 | return (x & (n - 1)) == 0; |
| 79 | } |
| 80 | |
| 81 | template <typename T> |
| 82 | static inline bool IsAligned(T* x, intptr_t n) { |
| 83 | return IsAligned(reinterpret_cast<uword>(x), n); |
| 84 | } |
| 85 | |
| 86 | template <typename T> |
| 87 | static inline T RoundDown(T x, intptr_t n) { |
| 88 | ASSERT(IsPowerOfTwo(n)); |
| 89 | return (x & -n); |
| 90 | } |
| 91 | |
| 92 | template <typename T> |
| 93 | static inline T* RoundDown(T* x, intptr_t n) { |
| 94 | return reinterpret_cast<T*>(RoundDown(reinterpret_cast<uword>(x), n)); |
| 95 | } |
| 96 | |
| 97 | template <typename T> |
| 98 | static inline T RoundUp(T x, intptr_t n) { |
| 99 | return RoundDown(x + n - 1, n); |
| 100 | } |
| 101 | |
| 102 | template <typename T> |
| 103 | static inline T* RoundUp(T* x, intptr_t n) { |
| 104 | return reinterpret_cast<T*>(RoundUp(reinterpret_cast<uword>(x), n)); |
| 105 | } |
| 106 | |
| 107 | static uintptr_t RoundUpToPowerOfTwo(uintptr_t x); |
| 108 | |
| 109 | static int CountOneBits64(uint64_t x); |
| 110 | static int CountOneBits32(uint32_t x); |
| 111 | |
| 112 | static int CountOneBitsWord(uword x) { |
| 113 | #ifdef ARCH_IS_64_BIT |
| 114 | return CountOneBits64(x); |
| 115 | #else |
| 116 | return CountOneBits32(x); |
| 117 | #endif |
| 118 | } |
| 119 | |
| 120 | static int HighestBit(int64_t v); |
| 121 | |
| 122 | static int BitLength(int64_t value) { |
| 123 | // Flip bits if negative (-1 becomes 0). |
| 124 | value ^= value >> (8 * sizeof(value) - 1); |
| 125 | return (value == 0) ? 0 : (Utils::HighestBit(value) + 1); |
| 126 | } |
| 127 | |
| 128 | static int CountLeadingZeros64(uint64_t x); |
| 129 | static int CountLeadingZeros32(uint32_t x); |
| 130 | |
| 131 | static int CountLeadingZerosWord(uword x) { |
| 132 | #ifdef ARCH_IS_64_BIT |
| 133 | return CountLeadingZeros64(x); |
| 134 | #else |
| 135 | return CountLeadingZeros32(x); |
| 136 | #endif |
| 137 | } |
| 138 | |
| 139 | static int CountTrailingZeros64(uint64_t x); |
| 140 | static int CountTrailingZeros32(uint32_t x); |
| 141 | |
| 142 | static int CountTrailingZerosWord(uword x) { |
| 143 | #ifdef ARCH_IS_64_BIT |
| 144 | return CountTrailingZeros64(x); |
| 145 | #else |
| 146 | return CountTrailingZeros32(x); |
| 147 | #endif |
| 148 | } |
| 149 | |
| 150 | static uint64_t ReverseBits64(uint64_t x); |
| 151 | static uint32_t ReverseBits32(uint32_t x); |
| 152 | |
| 153 | static uword ReverseBitsWord(uword x) { |
| 154 | #ifdef ARCH_IS_64_BIT |
| 155 | return ReverseBits64(x); |
| 156 | #else |
| 157 | return ReverseBits32(x); |
| 158 | #endif |
| 159 | } |
| 160 | |
| 161 | // Computes magic numbers to implement DIV or MOD operator. |
| 162 | static void CalculateMagicAndShiftForDivRem(int64_t divisor, |
| 163 | int64_t* magic, |
| 164 | int64_t* shift); |
| 165 | |
| 166 | // Computes a hash value for the given string. |
| 167 | static uint32_t StringHash(const char* data, int length); |
| 168 | |
| 169 | // Computes a hash value for the given word. |
| 170 | static uint32_t WordHash(intptr_t key); |
| 171 | |
| 172 | // Check whether an N-bit two's-complement representation can hold value. |
| 173 | template <typename T> |
| 174 | static inline bool IsInt(int N, T value) { |
| 175 | ASSERT((0 < N) && |
| 176 | (static_cast<unsigned int>(N) < (kBitsPerByte * sizeof(value)))); |
| 177 | T limit = static_cast<T>(1) << (N - 1); |
| 178 | return (-limit <= value) && (value < limit); |
| 179 | } |
| 180 | |
| 181 | template <typename T> |
| 182 | static inline bool IsUint(int N, T value) { |
| 183 | ASSERT((0 < N) && |
| 184 | (static_cast<unsigned int>(N) < (kBitsPerByte * sizeof(value)))); |
| 185 | const auto limit = |
| 186 | (static_cast<typename std::make_unsigned<T>::type>(1) << N) - 1; |
| 187 | return (0 <= value) && |
| 188 | (static_cast<typename std::make_unsigned<T>::type>(value) <= limit); |
| 189 | } |
| 190 | |
| 191 | // Check whether the magnitude of value fits in N bits, i.e., whether an |
| 192 | // (N+1)-bit sign-magnitude representation can hold value. |
| 193 | template <typename T> |
| 194 | static inline bool IsAbsoluteUint(int N, T value) { |
| 195 | ASSERT((0 < N) && |
| 196 | (static_cast<unsigned int>(N) < (kBitsPerByte * sizeof(value)))); |
| 197 | if (value < 0) value = -value; |
| 198 | return IsUint(N, value); |
| 199 | } |
| 200 | |
| 201 | static inline int32_t Low16Bits(int32_t value) { |
| 202 | return static_cast<int32_t>(value & 0xffff); |
| 203 | } |
| 204 | |
| 205 | static inline int32_t High16Bits(int32_t value) { |
| 206 | return static_cast<int32_t>(value >> 16); |
| 207 | } |
| 208 | |
| 209 | static inline int32_t Low32Bits(int64_t value) { |
| 210 | return static_cast<int32_t>(value); |
| 211 | } |
| 212 | |
| 213 | static inline int32_t High32Bits(int64_t value) { |
| 214 | return static_cast<int32_t>(value >> 32); |
| 215 | } |
| 216 | |
| 217 | static inline int64_t LowHighTo64Bits(uint32_t low, int32_t high) { |
| 218 | return (static_cast<uint64_t>(high) << 32) | (low & 0x0ffffffffLL); |
| 219 | } |
| 220 | |
| 221 | static inline constexpr bool IsAlphaNumeric(uint32_t c) { |
| 222 | return (c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z') || |
| 223 | IsDecimalDigit(c); |
| 224 | } |
| 225 | |
| 226 | static inline constexpr bool IsDecimalDigit(uint32_t c) { |
| 227 | return ('0' <= c) && (c <= '9'); |
| 228 | } |
| 229 | |
| 230 | static bool IsHexDigit(char c) { |
| 231 | return IsDecimalDigit(c) || (('A' <= c) && (c <= 'F')) || |
| 232 | (('a' <= c) && (c <= 'f')); |
| 233 | } |
| 234 | |
| 235 | static int HexDigitToInt(char c) { |
| 236 | ASSERT(IsHexDigit(c)); |
| 237 | if (IsDecimalDigit(c)) return c - '0'; |
| 238 | if (('A' <= c) && (c <= 'F')) return 10 + (c - 'A'); |
| 239 | return 10 + (c - 'a'); |
| 240 | } |
| 241 | |
| 242 | static char IntToHexDigit(int i) { |
| 243 | ASSERT(0 <= i && i < 16); |
| 244 | if (i < 10) return static_cast<char>('0' + i); |
| 245 | return static_cast<char>('A' + (i - 10)); |
| 246 | } |
| 247 | |
| 248 | // Perform a range check, checking if |
| 249 | // offset + count <= length |
| 250 | // without the risk of integer overflow. |
| 251 | static inline bool RangeCheck(intptr_t offset, |
| 252 | intptr_t count, |
| 253 | intptr_t length) { |
| 254 | return offset >= 0 && count >= 0 && length >= 0 && |
| 255 | count <= (length - offset); |
| 256 | } |
| 257 | |
| 258 | static inline bool WillAddOverflow(int64_t a, int64_t b) { |
| 259 | return ((b > 0) && (a > (kMaxInt64 - b))) || |
| 260 | ((b < 0) && (a < (kMinInt64 - b))); |
| 261 | } |
| 262 | |
| 263 | static inline bool WillSubOverflow(int64_t a, int64_t b) { |
| 264 | return ((b > 0) && (a < (kMinInt64 + b))) || |
| 265 | ((b < 0) && (a > (kMaxInt64 + b))); |
| 266 | } |
| 267 | |
| 268 | // Adds two int64_t values with wrapping around |
| 269 | // (two's complement arithmetic). |
| 270 | static inline int64_t AddWithWrapAround(int64_t a, int64_t b) { |
| 271 | // Avoid undefined behavior by doing arithmetic in the unsigned type. |
| 272 | return static_cast<int64_t>(static_cast<uint64_t>(a) + |
| 273 | static_cast<uint64_t>(b)); |
| 274 | } |
| 275 | |
| 276 | // Subtracts two int64_t values with wrapping around |
| 277 | // (two's complement arithmetic). |
| 278 | static inline int64_t SubWithWrapAround(int64_t a, int64_t b) { |
| 279 | // Avoid undefined behavior by doing arithmetic in the unsigned type. |
| 280 | return static_cast<int64_t>(static_cast<uint64_t>(a) - |
| 281 | static_cast<uint64_t>(b)); |
| 282 | } |
| 283 | |
| 284 | // Multiplies two int64_t values with wrapping around |
| 285 | // (two's complement arithmetic). |
| 286 | static inline int64_t MulWithWrapAround(int64_t a, int64_t b) { |
| 287 | // Avoid undefined behavior by doing arithmetic in the unsigned type. |
| 288 | return static_cast<int64_t>(static_cast<uint64_t>(a) * |
| 289 | static_cast<uint64_t>(b)); |
| 290 | } |
| 291 | |
| 292 | // Shifts int64_t value left. Supports any non-negative number of bits and |
| 293 | // silently discards shifted out bits. |
| 294 | static inline int64_t ShiftLeftWithTruncation(int64_t a, int64_t b) { |
| 295 | ASSERT(b >= 0); |
| 296 | if (b >= kBitsPerInt64) { |
| 297 | return 0; |
| 298 | } |
| 299 | // Avoid undefined behavior by doing arithmetic in the unsigned type. |
| 300 | return static_cast<int64_t>(static_cast<uint64_t>(a) << b); |
| 301 | } |
| 302 | |
| 303 | template <typename T> |
| 304 | static inline T RotateLeft(T value, uint8_t rotate) { |
| 305 | const uint8_t width = sizeof(T) * kBitsPerByte; |
| 306 | ASSERT(0 <= rotate); |
| 307 | ASSERT(rotate <= width); |
| 308 | using Unsigned = typename std::make_unsigned<T>::type; |
| 309 | return (static_cast<Unsigned>(value) << rotate) | |
| 310 | (static_cast<T>(value) >> ((width - rotate) & (width - 1))); |
| 311 | } |
| 312 | template <typename T> |
| 313 | static inline T RotateRight(T value, uint8_t rotate) { |
| 314 | const uint8_t width = sizeof(T) * kBitsPerByte; |
| 315 | ASSERT(0 <= rotate); |
| 316 | ASSERT(rotate <= width); |
| 317 | using Unsigned = typename std::make_unsigned<T>::type; |
| 318 | return (static_cast<T>(value) >> rotate) | |
| 319 | (static_cast<Unsigned>(value) << ((width - rotate) & (width - 1))); |
| 320 | } |
| 321 | |
| 322 | // Utility functions for converting values from host endianness to |
| 323 | // big or little endian values. |
| 324 | static uint16_t HostToBigEndian16(uint16_t host_value); |
| 325 | static uint32_t HostToBigEndian32(uint32_t host_value); |
| 326 | static uint64_t HostToBigEndian64(uint64_t host_value); |
| 327 | static uint16_t HostToLittleEndian16(uint16_t host_value); |
| 328 | static uint32_t HostToLittleEndian32(uint32_t host_value); |
| 329 | static uint64_t HostToLittleEndian64(uint64_t host_value); |
| 330 | |
| 331 | // Going between Host <-> LE/BE is the same operation for all practical |
| 332 | // purposes. |
| 333 | static inline uint32_t BigEndianToHost32(uint32_t be_value) { |
| 334 | return HostToBigEndian32(be_value); |
| 335 | } |
| 336 | static inline uint64_t LittleEndianToHost64(uint64_t le_value) { |
| 337 | return HostToLittleEndian64(le_value); |
| 338 | } |
| 339 | |
| 340 | static bool DoublesBitEqual(const double a, const double b) { |
| 341 | return bit_cast<int64_t, double>(a) == bit_cast<int64_t, double>(b); |
| 342 | } |
| 343 | |
| 344 | // A double-to-integer conversion that avoids undefined behavior. |
| 345 | // Out of range values and NaNs are converted to minimum value |
| 346 | // for type T. |
| 347 | template <typename T> |
| 348 | static T SafeDoubleToInt(double v) { |
| 349 | const double min = static_cast<double>(std::numeric_limits<T>::min()); |
| 350 | const double max = static_cast<double>(std::numeric_limits<T>::max()); |
| 351 | return (min <= v && v <= max) ? static_cast<T>(v) |
| 352 | : std::numeric_limits<T>::min(); |
| 353 | } |
| 354 | |
| 355 | // dart2js represents integers as double precision floats, which can |
| 356 | // represent anything in the range -2^53 ... 2^53. |
| 357 | static bool IsJavascriptInt(int64_t value) { |
| 358 | return ((-0x20000000000000LL <= value) && (value <= 0x20000000000000LL)); |
| 359 | } |
| 360 | |
| 361 | // The lowest n bits are 1, the others are 0. |
| 362 | static uword NBitMask(uint32_t n) { |
| 363 | ASSERT(n <= kBitsPerWord); |
| 364 | if (n == kBitsPerWord) { |
| 365 | static_assert((sizeof(uword) * kBitsPerByte) == kBitsPerWord, |
| 366 | "Unexpected uword size" ); |
| 367 | return std::numeric_limits<uword>::max(); |
| 368 | } |
| 369 | return (static_cast<uword>(1) << n) - 1; |
| 370 | } |
| 371 | |
| 372 | static word SignedNBitMask(uint32_t n) { |
| 373 | uword mask = NBitMask(n); |
| 374 | return bit_cast<word>(mask); |
| 375 | } |
| 376 | |
| 377 | template <typename T = uword> |
| 378 | static T Bit(uint32_t n) { |
| 379 | ASSERT(n < sizeof(T) * kBitsPerByte); |
| 380 | T bit = 1; |
| 381 | return bit << n; |
| 382 | } |
| 383 | |
| 384 | template <typename T> |
| 385 | DART_FORCE_INLINE static bool TestBit(T mask, intptr_t position) { |
| 386 | ASSERT(position < static_cast<intptr_t>(sizeof(T) * kBitsPerByte)); |
| 387 | return ((mask >> position) & 1) != 0; |
| 388 | } |
| 389 | |
| 390 | // Decode integer in SLEB128 format from |data| and update |byte_index|. |
| 391 | template <typename ValueType> |
| 392 | static ValueType DecodeSLEB128(const uint8_t* data, |
| 393 | const intptr_t data_length, |
| 394 | intptr_t* byte_index) { |
| 395 | using Unsigned = typename std::make_unsigned<ValueType>::type; |
| 396 | ASSERT(*byte_index < data_length); |
| 397 | uword shift = 0; |
| 398 | Unsigned value = 0; |
| 399 | uint8_t part = 0; |
| 400 | do { |
| 401 | part = data[(*byte_index)++]; |
| 402 | value |= static_cast<Unsigned>(part & 0x7f) << shift; |
| 403 | shift += 7; |
| 404 | } while ((part & 0x80) != 0); |
| 405 | |
| 406 | if ((shift < (sizeof(ValueType) * CHAR_BIT)) && ((part & 0x40) != 0)) { |
| 407 | const Unsigned kMax = std::numeric_limits<Unsigned>::max(); |
| 408 | value |= static_cast<Unsigned>(kMax << shift); |
| 409 | } |
| 410 | return static_cast<ValueType>(value); |
| 411 | } |
| 412 | |
| 413 | static char* StrError(int err, char* buffer, size_t bufsize); |
| 414 | |
| 415 | // Not all platforms support strndup. |
| 416 | static char* StrNDup(const char* s, intptr_t n); |
| 417 | static char* StrDup(const char* s); |
| 418 | static intptr_t StrNLen(const char* s, intptr_t n); |
| 419 | |
| 420 | static int Close(int fildes); |
| 421 | static size_t Read(int filedes, void* buf, size_t nbyte); |
| 422 | static int Unlink(const char* path); |
| 423 | |
| 424 | // Print formatted output info a buffer. |
| 425 | // |
| 426 | // Does not write more than size characters (including the trailing '\0'). |
| 427 | // |
| 428 | // Returns the number of characters (excluding the trailing '\0') |
| 429 | // that would been written if the buffer had been big enough. If |
| 430 | // the return value is greater or equal than the given size then the |
| 431 | // output has been truncated. The return value is never negative. |
| 432 | // |
| 433 | // The buffer will always be terminated by a '\0', unless the buffer |
| 434 | // is of size 0. The buffer might be NULL if the size is 0. |
| 435 | // |
| 436 | // This specification conforms to C99 standard which is implemented |
| 437 | // by glibc 2.1+ with one exception: the C99 standard allows a |
| 438 | // negative return value. We will terminate the vm rather than let |
| 439 | // that occur. |
| 440 | static int SNPrint(char* str, size_t size, const char* format, ...) |
| 441 | PRINTF_ATTRIBUTE(3, 4); |
| 442 | static int VSNPrint(char* str, size_t size, const char* format, va_list args); |
| 443 | |
| 444 | // Allocate a string and print formatted output into a malloc'd buffer. |
| 445 | static char* SCreate(const char* format, ...) PRINTF_ATTRIBUTE(1, 2); |
| 446 | static char* VSCreate(const char* format, va_list args); |
| 447 | |
| 448 | typedef std::unique_ptr<char, decltype(std::free)*> CStringUniquePtr; |
| 449 | |
| 450 | // Returns str in a unique_ptr with free used as its deleter. |
| 451 | static CStringUniquePtr CreateCStringUniquePtr(char* str); |
| 452 | }; |
| 453 | |
| 454 | } // namespace dart |
| 455 | |
| 456 | #if defined(HOST_OS_ANDROID) |
| 457 | #include "platform/utils_android.h" |
| 458 | #elif defined(HOST_OS_FUCHSIA) |
| 459 | #include "platform/utils_fuchsia.h" |
| 460 | #elif defined(HOST_OS_LINUX) |
| 461 | #include "platform/utils_linux.h" |
| 462 | #elif defined(HOST_OS_MACOS) |
| 463 | #include "platform/utils_macos.h" |
| 464 | #elif defined(HOST_OS_WINDOWS) |
| 465 | #include "platform/utils_win.h" |
| 466 | #else |
| 467 | #error Unknown target os. |
| 468 | #endif |
| 469 | |
| 470 | #endif // RUNTIME_PLATFORM_UTILS_H_ |
| 471 | |