| 1 | //===- llvm/ADT/SmallBitVector.h - 'Normally small' bit vectors -*- C++ -*-===// |
| 2 | // |
| 3 | // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions. |
| 4 | // See https://llvm.org/LICENSE.txt for license information. |
| 5 | // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception |
| 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | /// |
| 9 | /// \file |
| 10 | /// This file implements the SmallBitVector class. |
| 11 | /// |
| 12 | //===----------------------------------------------------------------------===// |
| 13 | |
| 14 | #ifndef LLVM_ADT_SMALLBITVECTOR_H |
| 15 | #define LLVM_ADT_SMALLBITVECTOR_H |
| 16 | |
| 17 | #include "llvm/ADT/BitVector.h" |
| 18 | #include "llvm/ADT/iterator_range.h" |
| 19 | #include "llvm/Support/MathExtras.h" |
| 20 | #include <algorithm> |
| 21 | #include <cassert> |
| 22 | #include <climits> |
| 23 | #include <cstddef> |
| 24 | #include <cstdint> |
| 25 | #include <limits> |
| 26 | #include <utility> |
| 27 | |
| 28 | namespace llvm { |
| 29 | |
| 30 | /// This is a 'bitvector' (really, a variable-sized bit array), optimized for |
| 31 | /// the case when the array is small. It contains one pointer-sized field, which |
| 32 | /// is directly used as a plain collection of bits when possible, or as a |
| 33 | /// pointer to a larger heap-allocated array when necessary. This allows normal |
| 34 | /// "small" cases to be fast without losing generality for large inputs. |
| 35 | class SmallBitVector { |
| 36 | // TODO: In "large" mode, a pointer to a BitVector is used, leading to an |
| 37 | // unnecessary level of indirection. It would be more efficient to use a |
| 38 | // pointer to memory containing size, allocation size, and the array of bits. |
| 39 | uintptr_t X = 1; |
| 40 | |
| 41 | enum { |
| 42 | // The number of bits in this class. |
| 43 | NumBaseBits = sizeof(uintptr_t) * CHAR_BIT, |
| 44 | |
| 45 | // One bit is used to discriminate between small and large mode. The |
| 46 | // remaining bits are used for the small-mode representation. |
| 47 | SmallNumRawBits = NumBaseBits - 1, |
| 48 | |
| 49 | // A few more bits are used to store the size of the bit set in small mode. |
| 50 | // Theoretically this is a ceil-log2. These bits are encoded in the most |
| 51 | // significant bits of the raw bits. |
| 52 | SmallNumSizeBits = (NumBaseBits == 32 ? 5 : |
| 53 | NumBaseBits == 64 ? 6 : |
| 54 | SmallNumRawBits), |
| 55 | |
| 56 | // The remaining bits are used to store the actual set in small mode. |
| 57 | SmallNumDataBits = SmallNumRawBits - SmallNumSizeBits |
| 58 | }; |
| 59 | |
| 60 | static_assert(NumBaseBits == 64 || NumBaseBits == 32, |
| 61 | "Unsupported word size" ); |
| 62 | |
| 63 | public: |
| 64 | using size_type = uintptr_t; |
| 65 | |
| 66 | // Encapsulation of a single bit. |
| 67 | class reference { |
| 68 | SmallBitVector &TheVector; |
| 69 | unsigned BitPos; |
| 70 | |
| 71 | public: |
| 72 | reference(SmallBitVector &b, unsigned Idx) : TheVector(b), BitPos(Idx) {} |
| 73 | |
| 74 | reference(const reference&) = default; |
| 75 | |
| 76 | reference& operator=(reference t) { |
| 77 | *this = bool(t); |
| 78 | return *this; |
| 79 | } |
| 80 | |
| 81 | reference& operator=(bool t) { |
| 82 | if (t) |
| 83 | TheVector.set(BitPos); |
| 84 | else |
| 85 | TheVector.reset(Idx: BitPos); |
| 86 | return *this; |
| 87 | } |
| 88 | |
| 89 | operator bool() const { |
| 90 | return const_cast<const SmallBitVector &>(TheVector).operator[](Idx: BitPos); |
| 91 | } |
| 92 | }; |
| 93 | |
| 94 | private: |
| 95 | BitVector *getPointer() const { |
| 96 | assert(!isSmall()); |
| 97 | return reinterpret_cast<BitVector *>(X); |
| 98 | } |
| 99 | |
| 100 | void switchToSmall(uintptr_t NewSmallBits, size_type NewSize) { |
| 101 | X = 1; |
| 102 | setSmallSize(NewSize); |
| 103 | setSmallBits(NewSmallBits); |
| 104 | } |
| 105 | |
| 106 | void switchToLarge(BitVector *BV) { |
| 107 | X = reinterpret_cast<uintptr_t>(BV); |
| 108 | assert(!isSmall() && "Tried to use an unaligned pointer" ); |
| 109 | } |
| 110 | |
| 111 | // Return all the bits used for the "small" representation; this includes |
| 112 | // bits for the size as well as the element bits. |
| 113 | uintptr_t getSmallRawBits() const { |
| 114 | assert(isSmall()); |
| 115 | return X >> 1; |
| 116 | } |
| 117 | |
| 118 | void setSmallRawBits(uintptr_t NewRawBits) { |
| 119 | assert(isSmall()); |
| 120 | X = (NewRawBits << 1) | uintptr_t(1); |
| 121 | } |
| 122 | |
| 123 | // Return the size. |
| 124 | size_type getSmallSize() const { |
| 125 | return getSmallRawBits() >> SmallNumDataBits; |
| 126 | } |
| 127 | |
| 128 | void setSmallSize(size_type Size) { |
| 129 | setSmallRawBits(getSmallBits() | (Size << SmallNumDataBits)); |
| 130 | } |
| 131 | |
| 132 | // Return the element bits. |
| 133 | uintptr_t getSmallBits() const { |
| 134 | return getSmallRawBits() & ~(~uintptr_t(0) << getSmallSize()); |
| 135 | } |
| 136 | |
| 137 | void setSmallBits(uintptr_t NewBits) { |
| 138 | setSmallRawBits((NewBits & ~(~uintptr_t(0) << getSmallSize())) | |
| 139 | (getSmallSize() << SmallNumDataBits)); |
| 140 | } |
| 141 | |
| 142 | public: |
| 143 | /// Creates an empty bitvector. |
| 144 | SmallBitVector() = default; |
| 145 | |
| 146 | /// Creates a bitvector of specified number of bits. All bits are initialized |
| 147 | /// to the specified value. |
| 148 | explicit SmallBitVector(unsigned s, bool t = false) { |
| 149 | if (s <= SmallNumDataBits) |
| 150 | switchToSmall(NewSmallBits: t ? ~uintptr_t(0) : 0, NewSize: s); |
| 151 | else |
| 152 | switchToLarge(BV: new BitVector(s, t)); |
| 153 | } |
| 154 | |
| 155 | /// SmallBitVector copy ctor. |
| 156 | SmallBitVector(const SmallBitVector &RHS) { |
| 157 | if (RHS.isSmall()) |
| 158 | X = RHS.X; |
| 159 | else |
| 160 | switchToLarge(BV: new BitVector(*RHS.getPointer())); |
| 161 | } |
| 162 | |
| 163 | SmallBitVector(SmallBitVector &&RHS) : X(RHS.X) { |
| 164 | RHS.X = 1; |
| 165 | } |
| 166 | |
| 167 | ~SmallBitVector() { |
| 168 | if (!isSmall()) |
| 169 | delete getPointer(); |
| 170 | } |
| 171 | |
| 172 | using const_set_bits_iterator = const_set_bits_iterator_impl<SmallBitVector>; |
| 173 | using set_iterator = const_set_bits_iterator; |
| 174 | |
| 175 | const_set_bits_iterator set_bits_begin() const { |
| 176 | return const_set_bits_iterator(*this); |
| 177 | } |
| 178 | |
| 179 | const_set_bits_iterator set_bits_end() const { |
| 180 | return const_set_bits_iterator(*this, -1); |
| 181 | } |
| 182 | |
| 183 | iterator_range<const_set_bits_iterator> set_bits() const { |
| 184 | return make_range(x: set_bits_begin(), y: set_bits_end()); |
| 185 | } |
| 186 | |
| 187 | bool isSmall() const { return X & uintptr_t(1); } |
| 188 | |
| 189 | /// Tests whether there are no bits in this bitvector. |
| 190 | bool empty() const { |
| 191 | return isSmall() ? getSmallSize() == 0 : getPointer()->empty(); |
| 192 | } |
| 193 | |
| 194 | /// Returns the number of bits in this bitvector. |
| 195 | size_type size() const { |
| 196 | return isSmall() ? getSmallSize() : getPointer()->size(); |
| 197 | } |
| 198 | |
| 199 | /// Returns the number of bits which are set. |
| 200 | size_type count() const { |
| 201 | if (isSmall()) { |
| 202 | uintptr_t Bits = getSmallBits(); |
| 203 | return llvm::popcount(Value: Bits); |
| 204 | } |
| 205 | return getPointer()->count(); |
| 206 | } |
| 207 | |
| 208 | /// Returns true if any bit is set. |
| 209 | bool any() const { |
| 210 | if (isSmall()) |
| 211 | return getSmallBits() != 0; |
| 212 | return getPointer()->any(); |
| 213 | } |
| 214 | |
| 215 | /// Returns true if all bits are set. |
| 216 | bool all() const { |
| 217 | if (isSmall()) |
| 218 | return getSmallBits() == (uintptr_t(1) << getSmallSize()) - 1; |
| 219 | return getPointer()->all(); |
| 220 | } |
| 221 | |
| 222 | /// Returns true if none of the bits are set. |
| 223 | bool none() const { |
| 224 | if (isSmall()) |
| 225 | return getSmallBits() == 0; |
| 226 | return getPointer()->none(); |
| 227 | } |
| 228 | |
| 229 | /// Returns the index of the first set bit, -1 if none of the bits are set. |
| 230 | int find_first() const { |
| 231 | if (isSmall()) { |
| 232 | uintptr_t Bits = getSmallBits(); |
| 233 | if (Bits == 0) |
| 234 | return -1; |
| 235 | return llvm::countr_zero(Val: Bits); |
| 236 | } |
| 237 | return getPointer()->find_first(); |
| 238 | } |
| 239 | |
| 240 | int find_last() const { |
| 241 | if (isSmall()) { |
| 242 | uintptr_t Bits = getSmallBits(); |
| 243 | if (Bits == 0) |
| 244 | return -1; |
| 245 | return NumBaseBits - llvm::countl_zero(Val: Bits) - 1; |
| 246 | } |
| 247 | return getPointer()->find_last(); |
| 248 | } |
| 249 | |
| 250 | /// Returns the index of the first unset bit, -1 if all of the bits are set. |
| 251 | int find_first_unset() const { |
| 252 | if (isSmall()) { |
| 253 | if (count() == getSmallSize()) |
| 254 | return -1; |
| 255 | |
| 256 | uintptr_t Bits = getSmallBits(); |
| 257 | return llvm::countr_one(Value: Bits); |
| 258 | } |
| 259 | return getPointer()->find_first_unset(); |
| 260 | } |
| 261 | |
| 262 | int find_last_unset() const { |
| 263 | if (isSmall()) { |
| 264 | if (count() == getSmallSize()) |
| 265 | return -1; |
| 266 | |
| 267 | uintptr_t Bits = getSmallBits(); |
| 268 | // Set unused bits. |
| 269 | Bits |= ~uintptr_t(0) << getSmallSize(); |
| 270 | return NumBaseBits - llvm::countl_one(Value: Bits) - 1; |
| 271 | } |
| 272 | return getPointer()->find_last_unset(); |
| 273 | } |
| 274 | |
| 275 | /// Returns the index of the next set bit following the "Prev" bit. |
| 276 | /// Returns -1 if the next set bit is not found. |
| 277 | int find_next(unsigned Prev) const { |
| 278 | if (isSmall()) { |
| 279 | uintptr_t Bits = getSmallBits(); |
| 280 | // Mask off previous bits. |
| 281 | Bits &= ~uintptr_t(0) << (Prev + 1); |
| 282 | if (Bits == 0 || Prev + 1 >= getSmallSize()) |
| 283 | return -1; |
| 284 | return llvm::countr_zero(Val: Bits); |
| 285 | } |
| 286 | return getPointer()->find_next(Prev); |
| 287 | } |
| 288 | |
| 289 | /// Returns the index of the next unset bit following the "Prev" bit. |
| 290 | /// Returns -1 if the next unset bit is not found. |
| 291 | int find_next_unset(unsigned Prev) const { |
| 292 | if (isSmall()) { |
| 293 | uintptr_t Bits = getSmallBits(); |
| 294 | // Mask in previous bits. |
| 295 | Bits |= (uintptr_t(1) << (Prev + 1)) - 1; |
| 296 | // Mask in unused bits. |
| 297 | Bits |= ~uintptr_t(0) << getSmallSize(); |
| 298 | |
| 299 | if (Bits == ~uintptr_t(0) || Prev + 1 >= getSmallSize()) |
| 300 | return -1; |
| 301 | return llvm::countr_one(Value: Bits); |
| 302 | } |
| 303 | return getPointer()->find_next_unset(Prev); |
| 304 | } |
| 305 | |
| 306 | /// find_prev - Returns the index of the first set bit that precedes the |
| 307 | /// the bit at \p PriorTo. Returns -1 if all previous bits are unset. |
| 308 | int find_prev(unsigned PriorTo) const { |
| 309 | if (isSmall()) { |
| 310 | if (PriorTo == 0) |
| 311 | return -1; |
| 312 | |
| 313 | --PriorTo; |
| 314 | uintptr_t Bits = getSmallBits(); |
| 315 | Bits &= maskTrailingOnes<uintptr_t>(N: PriorTo + 1); |
| 316 | if (Bits == 0) |
| 317 | return -1; |
| 318 | |
| 319 | return NumBaseBits - llvm::countl_zero(Val: Bits) - 1; |
| 320 | } |
| 321 | return getPointer()->find_prev(PriorTo); |
| 322 | } |
| 323 | |
| 324 | /// Clear all bits. |
| 325 | void clear() { |
| 326 | if (!isSmall()) |
| 327 | delete getPointer(); |
| 328 | switchToSmall(NewSmallBits: 0, NewSize: 0); |
| 329 | } |
| 330 | |
| 331 | /// Grow or shrink the bitvector. |
| 332 | void resize(unsigned N, bool t = false) { |
| 333 | if (!isSmall()) { |
| 334 | getPointer()->resize(N, t); |
| 335 | } else if (SmallNumDataBits >= N) { |
| 336 | uintptr_t NewBits = t ? ~uintptr_t(0) << getSmallSize() : 0; |
| 337 | setSmallSize(N); |
| 338 | setSmallBits(NewBits | getSmallBits()); |
| 339 | } else { |
| 340 | BitVector *BV = new BitVector(N, t); |
| 341 | uintptr_t OldBits = getSmallBits(); |
| 342 | for (size_type I = 0, E = getSmallSize(); I != E; ++I) |
| 343 | (*BV)[I] = (OldBits >> I) & 1; |
| 344 | switchToLarge(BV); |
| 345 | } |
| 346 | } |
| 347 | |
| 348 | void reserve(unsigned N) { |
| 349 | if (isSmall()) { |
| 350 | if (N > SmallNumDataBits) { |
| 351 | uintptr_t OldBits = getSmallRawBits(); |
| 352 | size_type SmallSize = getSmallSize(); |
| 353 | BitVector *BV = new BitVector(SmallSize); |
| 354 | for (size_type I = 0; I < SmallSize; ++I) |
| 355 | if ((OldBits >> I) & 1) |
| 356 | BV->set(I); |
| 357 | BV->reserve(N); |
| 358 | switchToLarge(BV); |
| 359 | } |
| 360 | } else { |
| 361 | getPointer()->reserve(N); |
| 362 | } |
| 363 | } |
| 364 | |
| 365 | // Set, reset, flip |
| 366 | SmallBitVector &set() { |
| 367 | if (isSmall()) |
| 368 | setSmallBits(~uintptr_t(0)); |
| 369 | else |
| 370 | getPointer()->set(); |
| 371 | return *this; |
| 372 | } |
| 373 | |
| 374 | SmallBitVector &set(unsigned Idx) { |
| 375 | if (isSmall()) { |
| 376 | assert(Idx <= static_cast<unsigned>( |
| 377 | std::numeric_limits<uintptr_t>::digits) && |
| 378 | "undefined behavior" ); |
| 379 | setSmallBits(getSmallBits() | (uintptr_t(1) << Idx)); |
| 380 | } |
| 381 | else |
| 382 | getPointer()->set(Idx); |
| 383 | return *this; |
| 384 | } |
| 385 | |
| 386 | /// Efficiently set a range of bits in [I, E) |
| 387 | SmallBitVector &set(unsigned I, unsigned E) { |
| 388 | assert(I <= E && "Attempted to set backwards range!" ); |
| 389 | assert(E <= size() && "Attempted to set out-of-bounds range!" ); |
| 390 | if (I == E) return *this; |
| 391 | if (isSmall()) { |
| 392 | uintptr_t EMask = ((uintptr_t)1) << E; |
| 393 | uintptr_t IMask = ((uintptr_t)1) << I; |
| 394 | uintptr_t Mask = EMask - IMask; |
| 395 | setSmallBits(getSmallBits() | Mask); |
| 396 | } else |
| 397 | getPointer()->set(I, E); |
| 398 | return *this; |
| 399 | } |
| 400 | |
| 401 | SmallBitVector &reset() { |
| 402 | if (isSmall()) |
| 403 | setSmallBits(0); |
| 404 | else |
| 405 | getPointer()->reset(); |
| 406 | return *this; |
| 407 | } |
| 408 | |
| 409 | SmallBitVector &reset(unsigned Idx) { |
| 410 | if (isSmall()) |
| 411 | setSmallBits(getSmallBits() & ~(uintptr_t(1) << Idx)); |
| 412 | else |
| 413 | getPointer()->reset(Idx); |
| 414 | return *this; |
| 415 | } |
| 416 | |
| 417 | /// Efficiently reset a range of bits in [I, E) |
| 418 | SmallBitVector &reset(unsigned I, unsigned E) { |
| 419 | assert(I <= E && "Attempted to reset backwards range!" ); |
| 420 | assert(E <= size() && "Attempted to reset out-of-bounds range!" ); |
| 421 | if (I == E) return *this; |
| 422 | if (isSmall()) { |
| 423 | uintptr_t EMask = ((uintptr_t)1) << E; |
| 424 | uintptr_t IMask = ((uintptr_t)1) << I; |
| 425 | uintptr_t Mask = EMask - IMask; |
| 426 | setSmallBits(getSmallBits() & ~Mask); |
| 427 | } else |
| 428 | getPointer()->reset(I, E); |
| 429 | return *this; |
| 430 | } |
| 431 | |
| 432 | SmallBitVector &flip() { |
| 433 | if (isSmall()) |
| 434 | setSmallBits(~getSmallBits()); |
| 435 | else |
| 436 | getPointer()->flip(); |
| 437 | return *this; |
| 438 | } |
| 439 | |
| 440 | SmallBitVector &flip(unsigned Idx) { |
| 441 | if (isSmall()) |
| 442 | setSmallBits(getSmallBits() ^ (uintptr_t(1) << Idx)); |
| 443 | else |
| 444 | getPointer()->flip(Idx); |
| 445 | return *this; |
| 446 | } |
| 447 | |
| 448 | // No argument flip. |
| 449 | SmallBitVector operator~() const { |
| 450 | return SmallBitVector(*this).flip(); |
| 451 | } |
| 452 | |
| 453 | // Indexing. |
| 454 | reference operator[](unsigned Idx) { |
| 455 | assert(Idx < size() && "Out-of-bounds Bit access." ); |
| 456 | return reference(*this, Idx); |
| 457 | } |
| 458 | |
| 459 | bool operator[](unsigned Idx) const { |
| 460 | assert(Idx < size() && "Out-of-bounds Bit access." ); |
| 461 | if (isSmall()) |
| 462 | return ((getSmallBits() >> Idx) & 1) != 0; |
| 463 | return getPointer()->operator[](Idx); |
| 464 | } |
| 465 | |
| 466 | /// Return the last element in the vector. |
| 467 | bool back() const { |
| 468 | assert(!empty() && "Getting last element of empty vector." ); |
| 469 | return (*this)[size() - 1]; |
| 470 | } |
| 471 | |
| 472 | bool test(unsigned Idx) const { |
| 473 | return (*this)[Idx]; |
| 474 | } |
| 475 | |
| 476 | // Push single bit to end of vector. |
| 477 | void push_back(bool Val) { |
| 478 | resize(N: size() + 1, t: Val); |
| 479 | } |
| 480 | |
| 481 | /// Pop one bit from the end of the vector. |
| 482 | void pop_back() { |
| 483 | assert(!empty() && "Empty vector has no element to pop." ); |
| 484 | resize(N: size() - 1); |
| 485 | } |
| 486 | |
| 487 | /// Test if any common bits are set. |
| 488 | bool anyCommon(const SmallBitVector &RHS) const { |
| 489 | if (isSmall() && RHS.isSmall()) |
| 490 | return (getSmallBits() & RHS.getSmallBits()) != 0; |
| 491 | if (!isSmall() && !RHS.isSmall()) |
| 492 | return getPointer()->anyCommon(RHS: *RHS.getPointer()); |
| 493 | |
| 494 | for (unsigned i = 0, e = std::min(a: size(), b: RHS.size()); i != e; ++i) |
| 495 | if (test(Idx: i) && RHS.test(Idx: i)) |
| 496 | return true; |
| 497 | return false; |
| 498 | } |
| 499 | |
| 500 | // Comparison operators. |
| 501 | bool operator==(const SmallBitVector &RHS) const { |
| 502 | if (size() != RHS.size()) |
| 503 | return false; |
| 504 | if (isSmall() && RHS.isSmall()) |
| 505 | return getSmallBits() == RHS.getSmallBits(); |
| 506 | else if (!isSmall() && !RHS.isSmall()) |
| 507 | return *getPointer() == *RHS.getPointer(); |
| 508 | else { |
| 509 | for (size_type I = 0, E = size(); I != E; ++I) { |
| 510 | if ((*this)[I] != RHS[I]) |
| 511 | return false; |
| 512 | } |
| 513 | return true; |
| 514 | } |
| 515 | } |
| 516 | |
| 517 | bool operator!=(const SmallBitVector &RHS) const { |
| 518 | return !(*this == RHS); |
| 519 | } |
| 520 | |
| 521 | // Intersection, union, disjoint union. |
| 522 | // FIXME BitVector::operator&= does not resize the LHS but this does |
| 523 | SmallBitVector &operator&=(const SmallBitVector &RHS) { |
| 524 | resize(N: std::max(a: size(), b: RHS.size())); |
| 525 | if (isSmall() && RHS.isSmall()) |
| 526 | setSmallBits(getSmallBits() & RHS.getSmallBits()); |
| 527 | else if (!isSmall() && !RHS.isSmall()) |
| 528 | getPointer()->operator&=(RHS: *RHS.getPointer()); |
| 529 | else { |
| 530 | size_type I, E; |
| 531 | for (I = 0, E = std::min(a: size(), b: RHS.size()); I != E; ++I) |
| 532 | (*this)[I] = test(Idx: I) && RHS.test(Idx: I); |
| 533 | for (E = size(); I != E; ++I) |
| 534 | reset(Idx: I); |
| 535 | } |
| 536 | return *this; |
| 537 | } |
| 538 | |
| 539 | /// Reset bits that are set in RHS. Same as *this &= ~RHS. |
| 540 | SmallBitVector &reset(const SmallBitVector &RHS) { |
| 541 | if (isSmall() && RHS.isSmall()) |
| 542 | setSmallBits(getSmallBits() & ~RHS.getSmallBits()); |
| 543 | else if (!isSmall() && !RHS.isSmall()) |
| 544 | getPointer()->reset(RHS: *RHS.getPointer()); |
| 545 | else |
| 546 | for (unsigned i = 0, e = std::min(a: size(), b: RHS.size()); i != e; ++i) |
| 547 | if (RHS.test(Idx: i)) |
| 548 | reset(Idx: i); |
| 549 | |
| 550 | return *this; |
| 551 | } |
| 552 | |
| 553 | /// Check if (This - RHS) is zero. This is the same as reset(RHS) and any(). |
| 554 | bool test(const SmallBitVector &RHS) const { |
| 555 | if (isSmall() && RHS.isSmall()) |
| 556 | return (getSmallBits() & ~RHS.getSmallBits()) != 0; |
| 557 | if (!isSmall() && !RHS.isSmall()) |
| 558 | return getPointer()->test(RHS: *RHS.getPointer()); |
| 559 | |
| 560 | unsigned i, e; |
| 561 | for (i = 0, e = std::min(a: size(), b: RHS.size()); i != e; ++i) |
| 562 | if (test(Idx: i) && !RHS.test(Idx: i)) |
| 563 | return true; |
| 564 | |
| 565 | for (e = size(); i != e; ++i) |
| 566 | if (test(Idx: i)) |
| 567 | return true; |
| 568 | |
| 569 | return false; |
| 570 | } |
| 571 | |
| 572 | SmallBitVector &operator|=(const SmallBitVector &RHS) { |
| 573 | resize(N: std::max(a: size(), b: RHS.size())); |
| 574 | if (isSmall() && RHS.isSmall()) |
| 575 | setSmallBits(getSmallBits() | RHS.getSmallBits()); |
| 576 | else if (!isSmall() && !RHS.isSmall()) |
| 577 | getPointer()->operator|=(RHS: *RHS.getPointer()); |
| 578 | else { |
| 579 | for (size_type I = 0, E = RHS.size(); I != E; ++I) |
| 580 | (*this)[I] = test(Idx: I) || RHS.test(Idx: I); |
| 581 | } |
| 582 | return *this; |
| 583 | } |
| 584 | |
| 585 | SmallBitVector &operator^=(const SmallBitVector &RHS) { |
| 586 | resize(N: std::max(a: size(), b: RHS.size())); |
| 587 | if (isSmall() && RHS.isSmall()) |
| 588 | setSmallBits(getSmallBits() ^ RHS.getSmallBits()); |
| 589 | else if (!isSmall() && !RHS.isSmall()) |
| 590 | getPointer()->operator^=(RHS: *RHS.getPointer()); |
| 591 | else { |
| 592 | for (size_type I = 0, E = RHS.size(); I != E; ++I) |
| 593 | (*this)[I] = test(Idx: I) != RHS.test(Idx: I); |
| 594 | } |
| 595 | return *this; |
| 596 | } |
| 597 | |
| 598 | SmallBitVector &operator<<=(unsigned N) { |
| 599 | if (isSmall()) |
| 600 | setSmallBits(getSmallBits() << N); |
| 601 | else |
| 602 | getPointer()->operator<<=(N); |
| 603 | return *this; |
| 604 | } |
| 605 | |
| 606 | SmallBitVector &operator>>=(unsigned N) { |
| 607 | if (isSmall()) |
| 608 | setSmallBits(getSmallBits() >> N); |
| 609 | else |
| 610 | getPointer()->operator>>=(N); |
| 611 | return *this; |
| 612 | } |
| 613 | |
| 614 | // Assignment operator. |
| 615 | const SmallBitVector &operator=(const SmallBitVector &RHS) { |
| 616 | if (isSmall()) { |
| 617 | if (RHS.isSmall()) |
| 618 | X = RHS.X; |
| 619 | else |
| 620 | switchToLarge(BV: new BitVector(*RHS.getPointer())); |
| 621 | } else { |
| 622 | if (!RHS.isSmall()) |
| 623 | *getPointer() = *RHS.getPointer(); |
| 624 | else { |
| 625 | delete getPointer(); |
| 626 | X = RHS.X; |
| 627 | } |
| 628 | } |
| 629 | return *this; |
| 630 | } |
| 631 | |
| 632 | const SmallBitVector &operator=(SmallBitVector &&RHS) { |
| 633 | if (this != &RHS) { |
| 634 | clear(); |
| 635 | swap(RHS); |
| 636 | } |
| 637 | return *this; |
| 638 | } |
| 639 | |
| 640 | void swap(SmallBitVector &RHS) { |
| 641 | std::swap(a&: X, b&: RHS.X); |
| 642 | } |
| 643 | |
| 644 | /// Add '1' bits from Mask to this vector. Don't resize. |
| 645 | /// This computes "*this |= Mask". |
| 646 | void setBitsInMask(const uint32_t *Mask, unsigned MaskWords = ~0u) { |
| 647 | if (isSmall()) |
| 648 | applyMask<true, false>(Mask, MaskWords); |
| 649 | else |
| 650 | getPointer()->setBitsInMask(Mask, MaskWords); |
| 651 | } |
| 652 | |
| 653 | /// Clear any bits in this vector that are set in Mask. Don't resize. |
| 654 | /// This computes "*this &= ~Mask". |
| 655 | void clearBitsInMask(const uint32_t *Mask, unsigned MaskWords = ~0u) { |
| 656 | if (isSmall()) |
| 657 | applyMask<false, false>(Mask, MaskWords); |
| 658 | else |
| 659 | getPointer()->clearBitsInMask(Mask, MaskWords); |
| 660 | } |
| 661 | |
| 662 | /// Add a bit to this vector for every '0' bit in Mask. Don't resize. |
| 663 | /// This computes "*this |= ~Mask". |
| 664 | void setBitsNotInMask(const uint32_t *Mask, unsigned MaskWords = ~0u) { |
| 665 | if (isSmall()) |
| 666 | applyMask<true, true>(Mask, MaskWords); |
| 667 | else |
| 668 | getPointer()->setBitsNotInMask(Mask, MaskWords); |
| 669 | } |
| 670 | |
| 671 | /// Clear a bit in this vector for every '0' bit in Mask. Don't resize. |
| 672 | /// This computes "*this &= Mask". |
| 673 | void clearBitsNotInMask(const uint32_t *Mask, unsigned MaskWords = ~0u) { |
| 674 | if (isSmall()) |
| 675 | applyMask<false, true>(Mask, MaskWords); |
| 676 | else |
| 677 | getPointer()->clearBitsNotInMask(Mask, MaskWords); |
| 678 | } |
| 679 | |
| 680 | void invalid() { |
| 681 | assert(empty()); |
| 682 | X = (uintptr_t)-1; |
| 683 | } |
| 684 | bool isInvalid() const { return X == (uintptr_t)-1; } |
| 685 | |
| 686 | ArrayRef<uintptr_t> getData(uintptr_t &Store) const { |
| 687 | if (!isSmall()) |
| 688 | return getPointer()->getData(); |
| 689 | Store = getSmallBits(); |
| 690 | return Store; |
| 691 | } |
| 692 | |
| 693 | private: |
| 694 | template <bool AddBits, bool InvertMask> |
| 695 | void applyMask(const uint32_t *Mask, unsigned MaskWords) { |
| 696 | assert(MaskWords <= sizeof(uintptr_t) && "Mask is larger than base!" ); |
| 697 | uintptr_t M = Mask[0]; |
| 698 | if (NumBaseBits == 64) |
| 699 | M |= uint64_t(Mask[1]) << 32; |
| 700 | if (InvertMask) |
| 701 | M = ~M; |
| 702 | if (AddBits) |
| 703 | setSmallBits(getSmallBits() | M); |
| 704 | else |
| 705 | setSmallBits(getSmallBits() & ~M); |
| 706 | } |
| 707 | }; |
| 708 | |
| 709 | inline SmallBitVector |
| 710 | operator&(const SmallBitVector &LHS, const SmallBitVector &RHS) { |
| 711 | SmallBitVector Result(LHS); |
| 712 | Result &= RHS; |
| 713 | return Result; |
| 714 | } |
| 715 | |
| 716 | inline SmallBitVector |
| 717 | operator|(const SmallBitVector &LHS, const SmallBitVector &RHS) { |
| 718 | SmallBitVector Result(LHS); |
| 719 | Result |= RHS; |
| 720 | return Result; |
| 721 | } |
| 722 | |
| 723 | inline SmallBitVector |
| 724 | operator^(const SmallBitVector &LHS, const SmallBitVector &RHS) { |
| 725 | SmallBitVector Result(LHS); |
| 726 | Result ^= RHS; |
| 727 | return Result; |
| 728 | } |
| 729 | |
| 730 | template <> struct DenseMapInfo<SmallBitVector> { |
| 731 | static inline SmallBitVector getEmptyKey() { return SmallBitVector(); } |
| 732 | static inline SmallBitVector getTombstoneKey() { |
| 733 | SmallBitVector V; |
| 734 | V.invalid(); |
| 735 | return V; |
| 736 | } |
| 737 | static unsigned getHashValue(const SmallBitVector &V) { |
| 738 | uintptr_t Store; |
| 739 | return DenseMapInfo< |
| 740 | std::pair<SmallBitVector::size_type, ArrayRef<uintptr_t>>>:: |
| 741 | getHashValue(PairVal: std::make_pair(x: V.size(), y: V.getData(Store))); |
| 742 | } |
| 743 | static bool isEqual(const SmallBitVector &LHS, const SmallBitVector &RHS) { |
| 744 | if (LHS.isInvalid() || RHS.isInvalid()) |
| 745 | return LHS.isInvalid() == RHS.isInvalid(); |
| 746 | return LHS == RHS; |
| 747 | } |
| 748 | }; |
| 749 | } // end namespace llvm |
| 750 | |
| 751 | namespace std { |
| 752 | |
| 753 | /// Implement std::swap in terms of BitVector swap. |
| 754 | inline void |
| 755 | swap(llvm::SmallBitVector &LHS, llvm::SmallBitVector &RHS) { |
| 756 | LHS.swap(RHS); |
| 757 | } |
| 758 | |
| 759 | } // end namespace std |
| 760 | |
| 761 | #endif // LLVM_ADT_SMALLBITVECTOR_H |
| 762 | |