1 //===- ConstantRange.cpp - ConstantRange implementation -------------------===// 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 // Represent a range of possible values that may occur when the program is run 10 // for an integral value. This keeps track of a lower and upper bound for the 11 // constant, which MAY wrap around the end of the numeric range. To do this, it 12 // keeps track of a [lower, upper) bound, which specifies an interval just like 13 // STL iterators. When used with boolean values, the following are important 14 // ranges (other integral ranges use min/max values for special range values): 15 // 16 // [F, F) = {} = Empty set 17 // [T, F) = {T} 18 // [F, T) = {F} 19 // [T, T) = {F, T} = Full set 20 // 21 //===----------------------------------------------------------------------===// 22 23 #include "llvm/ADT/APInt.h" 24 #include "llvm/Config/llvm-config.h" 25 #include "llvm/IR/ConstantRange.h" 26 #include "llvm/IR/Constants.h" 27 #include "llvm/IR/InstrTypes.h" 28 #include "llvm/IR/Instruction.h" 29 #include "llvm/IR/Metadata.h" 30 #include "llvm/IR/Operator.h" 31 #include "llvm/Support/Compiler.h" 32 #include "llvm/Support/Debug.h" 33 #include "llvm/Support/ErrorHandling.h" 34 #include "llvm/Support/KnownBits.h" 35 #include "llvm/Support/raw_ostream.h" 36 #include <algorithm> 37 #include <cassert> 38 #include <cstdint> 39 40 using namespace llvm; 41 42 ConstantRange::ConstantRange(uint32_t BitWidth, bool Full) 43 : Lower(Full ? APInt::getMaxValue(BitWidth) : APInt::getMinValue(BitWidth)), 44 Upper(Lower) {} 45 46 ConstantRange::ConstantRange(APInt V) 47 : Lower(std::move(V)), Upper(Lower + 1) {} 48 49 ConstantRange::ConstantRange(APInt L, APInt U) 50 : Lower(std::move(L)), Upper(std::move(U)) { 51 assert(Lower.getBitWidth() == Upper.getBitWidth() && 52 "ConstantRange with unequal bit widths"); 53 assert((Lower != Upper || (Lower.isMaxValue() || Lower.isMinValue())) && 54 "Lower == Upper, but they aren't min or max value!"); 55 } 56 57 ConstantRange ConstantRange::fromKnownBits(const KnownBits &Known, 58 bool IsSigned) { 59 assert(!Known.hasConflict() && "Expected valid KnownBits"); 60 61 if (Known.isUnknown()) 62 return getFull(Known.getBitWidth()); 63 64 // For unsigned ranges, or signed ranges with known sign bit, create a simple 65 // range between the smallest and largest possible value. 66 if (!IsSigned || Known.isNegative() || Known.isNonNegative()) 67 return ConstantRange(Known.getMinValue(), Known.getMaxValue() + 1); 68 69 // If we don't know the sign bit, pick the lower bound as a negative number 70 // and the upper bound as a non-negative one. 71 APInt Lower = Known.getMinValue(), Upper = Known.getMaxValue(); 72 Lower.setSignBit(); 73 Upper.clearSignBit(); 74 return ConstantRange(Lower, Upper + 1); 75 } 76 77 ConstantRange ConstantRange::makeAllowedICmpRegion(CmpInst::Predicate Pred, 78 const ConstantRange &CR) { 79 if (CR.isEmptySet()) 80 return CR; 81 82 uint32_t W = CR.getBitWidth(); 83 switch (Pred) { 84 default: 85 llvm_unreachable("Invalid ICmp predicate to makeAllowedICmpRegion()"); 86 case CmpInst::ICMP_EQ: 87 return CR; 88 case CmpInst::ICMP_NE: 89 if (CR.isSingleElement()) 90 return ConstantRange(CR.getUpper(), CR.getLower()); 91 return getFull(W); 92 case CmpInst::ICMP_ULT: { 93 APInt UMax(CR.getUnsignedMax()); 94 if (UMax.isMinValue()) 95 return getEmpty(W); 96 return ConstantRange(APInt::getMinValue(W), std::move(UMax)); 97 } 98 case CmpInst::ICMP_SLT: { 99 APInt SMax(CR.getSignedMax()); 100 if (SMax.isMinSignedValue()) 101 return getEmpty(W); 102 return ConstantRange(APInt::getSignedMinValue(W), std::move(SMax)); 103 } 104 case CmpInst::ICMP_ULE: 105 return getNonEmpty(APInt::getMinValue(W), CR.getUnsignedMax() + 1); 106 case CmpInst::ICMP_SLE: 107 return getNonEmpty(APInt::getSignedMinValue(W), CR.getSignedMax() + 1); 108 case CmpInst::ICMP_UGT: { 109 APInt UMin(CR.getUnsignedMin()); 110 if (UMin.isMaxValue()) 111 return getEmpty(W); 112 return ConstantRange(std::move(UMin) + 1, APInt::getNullValue(W)); 113 } 114 case CmpInst::ICMP_SGT: { 115 APInt SMin(CR.getSignedMin()); 116 if (SMin.isMaxSignedValue()) 117 return getEmpty(W); 118 return ConstantRange(std::move(SMin) + 1, APInt::getSignedMinValue(W)); 119 } 120 case CmpInst::ICMP_UGE: 121 return getNonEmpty(CR.getUnsignedMin(), APInt::getNullValue(W)); 122 case CmpInst::ICMP_SGE: 123 return getNonEmpty(CR.getSignedMin(), APInt::getSignedMinValue(W)); 124 } 125 } 126 127 ConstantRange ConstantRange::makeSatisfyingICmpRegion(CmpInst::Predicate Pred, 128 const ConstantRange &CR) { 129 // Follows from De-Morgan's laws: 130 // 131 // ~(~A union ~B) == A intersect B. 132 // 133 return makeAllowedICmpRegion(CmpInst::getInversePredicate(Pred), CR) 134 .inverse(); 135 } 136 137 ConstantRange ConstantRange::makeExactICmpRegion(CmpInst::Predicate Pred, 138 const APInt &C) { 139 // Computes the exact range that is equal to both the constant ranges returned 140 // by makeAllowedICmpRegion and makeSatisfyingICmpRegion. This is always true 141 // when RHS is a singleton such as an APInt and so the assert is valid. 142 // However for non-singleton RHS, for example ult [2,5) makeAllowedICmpRegion 143 // returns [0,4) but makeSatisfyICmpRegion returns [0,2). 144 // 145 assert(makeAllowedICmpRegion(Pred, C) == makeSatisfyingICmpRegion(Pred, C)); 146 return makeAllowedICmpRegion(Pred, C); 147 } 148 149 bool ConstantRange::getEquivalentICmp(CmpInst::Predicate &Pred, 150 APInt &RHS) const { 151 bool Success = false; 152 153 if (isFullSet() || isEmptySet()) { 154 Pred = isEmptySet() ? CmpInst::ICMP_ULT : CmpInst::ICMP_UGE; 155 RHS = APInt(getBitWidth(), 0); 156 Success = true; 157 } else if (auto *OnlyElt = getSingleElement()) { 158 Pred = CmpInst::ICMP_EQ; 159 RHS = *OnlyElt; 160 Success = true; 161 } else if (auto *OnlyMissingElt = getSingleMissingElement()) { 162 Pred = CmpInst::ICMP_NE; 163 RHS = *OnlyMissingElt; 164 Success = true; 165 } else if (getLower().isMinSignedValue() || getLower().isMinValue()) { 166 Pred = 167 getLower().isMinSignedValue() ? CmpInst::ICMP_SLT : CmpInst::ICMP_ULT; 168 RHS = getUpper(); 169 Success = true; 170 } else if (getUpper().isMinSignedValue() || getUpper().isMinValue()) { 171 Pred = 172 getUpper().isMinSignedValue() ? CmpInst::ICMP_SGE : CmpInst::ICMP_UGE; 173 RHS = getLower(); 174 Success = true; 175 } 176 177 assert((!Success || ConstantRange::makeExactICmpRegion(Pred, RHS) == *this) && 178 "Bad result!"); 179 180 return Success; 181 } 182 183 /// Exact mul nuw region for single element RHS. 184 static ConstantRange makeExactMulNUWRegion(const APInt &V) { 185 unsigned BitWidth = V.getBitWidth(); 186 if (V == 0) 187 return ConstantRange::getFull(V.getBitWidth()); 188 189 return ConstantRange::getNonEmpty( 190 APIntOps::RoundingUDiv(APInt::getMinValue(BitWidth), V, 191 APInt::Rounding::UP), 192 APIntOps::RoundingUDiv(APInt::getMaxValue(BitWidth), V, 193 APInt::Rounding::DOWN) + 1); 194 } 195 196 /// Exact mul nsw region for single element RHS. 197 static ConstantRange makeExactMulNSWRegion(const APInt &V) { 198 // Handle special case for 0, -1 and 1. See the last for reason why we 199 // specialize -1 and 1. 200 unsigned BitWidth = V.getBitWidth(); 201 if (V == 0 || V.isOneValue()) 202 return ConstantRange::getFull(BitWidth); 203 204 APInt MinValue = APInt::getSignedMinValue(BitWidth); 205 APInt MaxValue = APInt::getSignedMaxValue(BitWidth); 206 // e.g. Returning [-127, 127], represented as [-127, -128). 207 if (V.isAllOnesValue()) 208 return ConstantRange(-MaxValue, MinValue); 209 210 APInt Lower, Upper; 211 if (V.isNegative()) { 212 Lower = APIntOps::RoundingSDiv(MaxValue, V, APInt::Rounding::UP); 213 Upper = APIntOps::RoundingSDiv(MinValue, V, APInt::Rounding::DOWN); 214 } else { 215 Lower = APIntOps::RoundingSDiv(MinValue, V, APInt::Rounding::UP); 216 Upper = APIntOps::RoundingSDiv(MaxValue, V, APInt::Rounding::DOWN); 217 } 218 // ConstantRange ctor take a half inclusive interval [Lower, Upper + 1). 219 // Upper + 1 is guaranteed not to overflow, because |divisor| > 1. 0, -1, 220 // and 1 are already handled as special cases. 221 return ConstantRange(Lower, Upper + 1); 222 } 223 224 ConstantRange 225 ConstantRange::makeGuaranteedNoWrapRegion(Instruction::BinaryOps BinOp, 226 const ConstantRange &Other, 227 unsigned NoWrapKind) { 228 using OBO = OverflowingBinaryOperator; 229 230 assert(Instruction::isBinaryOp(BinOp) && "Binary operators only!"); 231 232 assert((NoWrapKind == OBO::NoSignedWrap || 233 NoWrapKind == OBO::NoUnsignedWrap) && 234 "NoWrapKind invalid!"); 235 236 bool Unsigned = NoWrapKind == OBO::NoUnsignedWrap; 237 unsigned BitWidth = Other.getBitWidth(); 238 239 switch (BinOp) { 240 default: 241 llvm_unreachable("Unsupported binary op"); 242 243 case Instruction::Add: { 244 if (Unsigned) 245 return getNonEmpty(APInt::getNullValue(BitWidth), 246 -Other.getUnsignedMax()); 247 248 APInt SignedMinVal = APInt::getSignedMinValue(BitWidth); 249 APInt SMin = Other.getSignedMin(), SMax = Other.getSignedMax(); 250 return getNonEmpty( 251 SMin.isNegative() ? SignedMinVal - SMin : SignedMinVal, 252 SMax.isStrictlyPositive() ? SignedMinVal - SMax : SignedMinVal); 253 } 254 255 case Instruction::Sub: { 256 if (Unsigned) 257 return getNonEmpty(Other.getUnsignedMax(), APInt::getMinValue(BitWidth)); 258 259 APInt SignedMinVal = APInt::getSignedMinValue(BitWidth); 260 APInt SMin = Other.getSignedMin(), SMax = Other.getSignedMax(); 261 return getNonEmpty( 262 SMax.isStrictlyPositive() ? SignedMinVal + SMax : SignedMinVal, 263 SMin.isNegative() ? SignedMinVal + SMin : SignedMinVal); 264 } 265 266 case Instruction::Mul: 267 if (Unsigned) 268 return makeExactMulNUWRegion(Other.getUnsignedMax()); 269 270 return makeExactMulNSWRegion(Other.getSignedMin()) 271 .intersectWith(makeExactMulNSWRegion(Other.getSignedMax())); 272 273 case Instruction::Shl: { 274 // For given range of shift amounts, if we ignore all illegal shift amounts 275 // (that always produce poison), what shift amount range is left? 276 ConstantRange ShAmt = Other.intersectWith( 277 ConstantRange(APInt(BitWidth, 0), APInt(BitWidth, (BitWidth - 1) + 1))); 278 if (ShAmt.isEmptySet()) { 279 // If the entire range of shift amounts is already poison-producing, 280 // then we can freely add more poison-producing flags ontop of that. 281 return getFull(BitWidth); 282 } 283 // There are some legal shift amounts, we can compute conservatively-correct 284 // range of no-wrap inputs. Note that by now we have clamped the ShAmtUMax 285 // to be at most bitwidth-1, which results in most conservative range. 286 APInt ShAmtUMax = ShAmt.getUnsignedMax(); 287 if (Unsigned) 288 return getNonEmpty(APInt::getNullValue(BitWidth), 289 APInt::getMaxValue(BitWidth).lshr(ShAmtUMax) + 1); 290 return getNonEmpty(APInt::getSignedMinValue(BitWidth).ashr(ShAmtUMax), 291 APInt::getSignedMaxValue(BitWidth).ashr(ShAmtUMax) + 1); 292 } 293 } 294 } 295 296 ConstantRange ConstantRange::makeExactNoWrapRegion(Instruction::BinaryOps BinOp, 297 const APInt &Other, 298 unsigned NoWrapKind) { 299 // makeGuaranteedNoWrapRegion() is exact for single-element ranges, as 300 // "for all" and "for any" coincide in this case. 301 return makeGuaranteedNoWrapRegion(BinOp, ConstantRange(Other), NoWrapKind); 302 } 303 304 bool ConstantRange::isFullSet() const { 305 return Lower == Upper && Lower.isMaxValue(); 306 } 307 308 bool ConstantRange::isEmptySet() const { 309 return Lower == Upper && Lower.isMinValue(); 310 } 311 312 bool ConstantRange::isWrappedSet() const { 313 return Lower.ugt(Upper) && !Upper.isNullValue(); 314 } 315 316 bool ConstantRange::isUpperWrapped() const { 317 return Lower.ugt(Upper); 318 } 319 320 bool ConstantRange::isSignWrappedSet() const { 321 return Lower.sgt(Upper) && !Upper.isMinSignedValue(); 322 } 323 324 bool ConstantRange::isUpperSignWrapped() const { 325 return Lower.sgt(Upper); 326 } 327 328 bool 329 ConstantRange::isSizeStrictlySmallerThan(const ConstantRange &Other) const { 330 assert(getBitWidth() == Other.getBitWidth()); 331 if (isFullSet()) 332 return false; 333 if (Other.isFullSet()) 334 return true; 335 return (Upper - Lower).ult(Other.Upper - Other.Lower); 336 } 337 338 bool 339 ConstantRange::isSizeLargerThan(uint64_t MaxSize) const { 340 assert(MaxSize && "MaxSize can't be 0."); 341 // If this a full set, we need special handling to avoid needing an extra bit 342 // to represent the size. 343 if (isFullSet()) 344 return APInt::getMaxValue(getBitWidth()).ugt(MaxSize - 1); 345 346 return (Upper - Lower).ugt(MaxSize); 347 } 348 349 bool ConstantRange::isAllNegative() const { 350 // Empty set is all negative, full set is not. 351 if (isEmptySet()) 352 return true; 353 if (isFullSet()) 354 return false; 355 356 return !isUpperSignWrapped() && !Upper.isStrictlyPositive(); 357 } 358 359 bool ConstantRange::isAllNonNegative() const { 360 // Empty and full set are automatically treated correctly. 361 return !isSignWrappedSet() && Lower.isNonNegative(); 362 } 363 364 APInt ConstantRange::getUnsignedMax() const { 365 if (isFullSet() || isUpperWrapped()) 366 return APInt::getMaxValue(getBitWidth()); 367 return getUpper() - 1; 368 } 369 370 APInt ConstantRange::getUnsignedMin() const { 371 if (isFullSet() || isWrappedSet()) 372 return APInt::getMinValue(getBitWidth()); 373 return getLower(); 374 } 375 376 APInt ConstantRange::getSignedMax() const { 377 if (isFullSet() || isUpperSignWrapped()) 378 return APInt::getSignedMaxValue(getBitWidth()); 379 return getUpper() - 1; 380 } 381 382 APInt ConstantRange::getSignedMin() const { 383 if (isFullSet() || isSignWrappedSet()) 384 return APInt::getSignedMinValue(getBitWidth()); 385 return getLower(); 386 } 387 388 bool ConstantRange::contains(const APInt &V) const { 389 if (Lower == Upper) 390 return isFullSet(); 391 392 if (!isUpperWrapped()) 393 return Lower.ule(V) && V.ult(Upper); 394 return Lower.ule(V) || V.ult(Upper); 395 } 396 397 bool ConstantRange::contains(const ConstantRange &Other) const { 398 if (isFullSet() || Other.isEmptySet()) return true; 399 if (isEmptySet() || Other.isFullSet()) return false; 400 401 if (!isUpperWrapped()) { 402 if (Other.isUpperWrapped()) 403 return false; 404 405 return Lower.ule(Other.getLower()) && Other.getUpper().ule(Upper); 406 } 407 408 if (!Other.isUpperWrapped()) 409 return Other.getUpper().ule(Upper) || 410 Lower.ule(Other.getLower()); 411 412 return Other.getUpper().ule(Upper) && Lower.ule(Other.getLower()); 413 } 414 415 ConstantRange ConstantRange::subtract(const APInt &Val) const { 416 assert(Val.getBitWidth() == getBitWidth() && "Wrong bit width"); 417 // If the set is empty or full, don't modify the endpoints. 418 if (Lower == Upper) 419 return *this; 420 return ConstantRange(Lower - Val, Upper - Val); 421 } 422 423 ConstantRange ConstantRange::difference(const ConstantRange &CR) const { 424 return intersectWith(CR.inverse()); 425 } 426 427 static ConstantRange getPreferredRange( 428 const ConstantRange &CR1, const ConstantRange &CR2, 429 ConstantRange::PreferredRangeType Type) { 430 if (Type == ConstantRange::Unsigned) { 431 if (!CR1.isWrappedSet() && CR2.isWrappedSet()) 432 return CR1; 433 if (CR1.isWrappedSet() && !CR2.isWrappedSet()) 434 return CR2; 435 } else if (Type == ConstantRange::Signed) { 436 if (!CR1.isSignWrappedSet() && CR2.isSignWrappedSet()) 437 return CR1; 438 if (CR1.isSignWrappedSet() && !CR2.isSignWrappedSet()) 439 return CR2; 440 } 441 442 if (CR1.isSizeStrictlySmallerThan(CR2)) 443 return CR1; 444 return CR2; 445 } 446 447 ConstantRange ConstantRange::intersectWith(const ConstantRange &CR, 448 PreferredRangeType Type) const { 449 assert(getBitWidth() == CR.getBitWidth() && 450 "ConstantRange types don't agree!"); 451 452 // Handle common cases. 453 if ( isEmptySet() || CR.isFullSet()) return *this; 454 if (CR.isEmptySet() || isFullSet()) return CR; 455 456 if (!isUpperWrapped() && CR.isUpperWrapped()) 457 return CR.intersectWith(*this, Type); 458 459 if (!isUpperWrapped() && !CR.isUpperWrapped()) { 460 if (Lower.ult(CR.Lower)) { 461 // L---U : this 462 // L---U : CR 463 if (Upper.ule(CR.Lower)) 464 return getEmpty(); 465 466 // L---U : this 467 // L---U : CR 468 if (Upper.ult(CR.Upper)) 469 return ConstantRange(CR.Lower, Upper); 470 471 // L-------U : this 472 // L---U : CR 473 return CR; 474 } 475 // L---U : this 476 // L-------U : CR 477 if (Upper.ult(CR.Upper)) 478 return *this; 479 480 // L-----U : this 481 // L-----U : CR 482 if (Lower.ult(CR.Upper)) 483 return ConstantRange(Lower, CR.Upper); 484 485 // L---U : this 486 // L---U : CR 487 return getEmpty(); 488 } 489 490 if (isUpperWrapped() && !CR.isUpperWrapped()) { 491 if (CR.Lower.ult(Upper)) { 492 // ------U L--- : this 493 // L--U : CR 494 if (CR.Upper.ult(Upper)) 495 return CR; 496 497 // ------U L--- : this 498 // L------U : CR 499 if (CR.Upper.ule(Lower)) 500 return ConstantRange(CR.Lower, Upper); 501 502 // ------U L--- : this 503 // L----------U : CR 504 return getPreferredRange(*this, CR, Type); 505 } 506 if (CR.Lower.ult(Lower)) { 507 // --U L---- : this 508 // L--U : CR 509 if (CR.Upper.ule(Lower)) 510 return getEmpty(); 511 512 // --U L---- : this 513 // L------U : CR 514 return ConstantRange(Lower, CR.Upper); 515 } 516 517 // --U L------ : this 518 // L--U : CR 519 return CR; 520 } 521 522 if (CR.Upper.ult(Upper)) { 523 // ------U L-- : this 524 // --U L------ : CR 525 if (CR.Lower.ult(Upper)) 526 return getPreferredRange(*this, CR, Type); 527 528 // ----U L-- : this 529 // --U L---- : CR 530 if (CR.Lower.ult(Lower)) 531 return ConstantRange(Lower, CR.Upper); 532 533 // ----U L---- : this 534 // --U L-- : CR 535 return CR; 536 } 537 if (CR.Upper.ule(Lower)) { 538 // --U L-- : this 539 // ----U L---- : CR 540 if (CR.Lower.ult(Lower)) 541 return *this; 542 543 // --U L---- : this 544 // ----U L-- : CR 545 return ConstantRange(CR.Lower, Upper); 546 } 547 548 // --U L------ : this 549 // ------U L-- : CR 550 return getPreferredRange(*this, CR, Type); 551 } 552 553 ConstantRange ConstantRange::unionWith(const ConstantRange &CR, 554 PreferredRangeType Type) const { 555 assert(getBitWidth() == CR.getBitWidth() && 556 "ConstantRange types don't agree!"); 557 558 if ( isFullSet() || CR.isEmptySet()) return *this; 559 if (CR.isFullSet() || isEmptySet()) return CR; 560 561 if (!isUpperWrapped() && CR.isUpperWrapped()) 562 return CR.unionWith(*this, Type); 563 564 if (!isUpperWrapped() && !CR.isUpperWrapped()) { 565 // L---U and L---U : this 566 // L---U L---U : CR 567 // result in one of 568 // L---------U 569 // -----U L----- 570 if (CR.Upper.ult(Lower) || Upper.ult(CR.Lower)) 571 return getPreferredRange( 572 ConstantRange(Lower, CR.Upper), ConstantRange(CR.Lower, Upper), Type); 573 574 APInt L = CR.Lower.ult(Lower) ? CR.Lower : Lower; 575 APInt U = (CR.Upper - 1).ugt(Upper - 1) ? CR.Upper : Upper; 576 577 if (L.isNullValue() && U.isNullValue()) 578 return getFull(); 579 580 return ConstantRange(std::move(L), std::move(U)); 581 } 582 583 if (!CR.isUpperWrapped()) { 584 // ------U L----- and ------U L----- : this 585 // L--U L--U : CR 586 if (CR.Upper.ule(Upper) || CR.Lower.uge(Lower)) 587 return *this; 588 589 // ------U L----- : this 590 // L---------U : CR 591 if (CR.Lower.ule(Upper) && Lower.ule(CR.Upper)) 592 return getFull(); 593 594 // ----U L---- : this 595 // L---U : CR 596 // results in one of 597 // ----------U L---- 598 // ----U L---------- 599 if (Upper.ult(CR.Lower) && CR.Upper.ult(Lower)) 600 return getPreferredRange( 601 ConstantRange(Lower, CR.Upper), ConstantRange(CR.Lower, Upper), Type); 602 603 // ----U L----- : this 604 // L----U : CR 605 if (Upper.ult(CR.Lower) && Lower.ule(CR.Upper)) 606 return ConstantRange(CR.Lower, Upper); 607 608 // ------U L---- : this 609 // L-----U : CR 610 assert(CR.Lower.ule(Upper) && CR.Upper.ult(Lower) && 611 "ConstantRange::unionWith missed a case with one range wrapped"); 612 return ConstantRange(Lower, CR.Upper); 613 } 614 615 // ------U L---- and ------U L---- : this 616 // -U L----------- and ------------U L : CR 617 if (CR.Lower.ule(Upper) || Lower.ule(CR.Upper)) 618 return getFull(); 619 620 APInt L = CR.Lower.ult(Lower) ? CR.Lower : Lower; 621 APInt U = CR.Upper.ugt(Upper) ? CR.Upper : Upper; 622 623 return ConstantRange(std::move(L), std::move(U)); 624 } 625 626 ConstantRange ConstantRange::castOp(Instruction::CastOps CastOp, 627 uint32_t ResultBitWidth) const { 628 switch (CastOp) { 629 default: 630 llvm_unreachable("unsupported cast type"); 631 case Instruction::Trunc: 632 return truncate(ResultBitWidth); 633 case Instruction::SExt: 634 return signExtend(ResultBitWidth); 635 case Instruction::ZExt: 636 return zeroExtend(ResultBitWidth); 637 case Instruction::BitCast: 638 return *this; 639 case Instruction::FPToUI: 640 case Instruction::FPToSI: 641 if (getBitWidth() == ResultBitWidth) 642 return *this; 643 else 644 return getFull(ResultBitWidth); 645 case Instruction::UIToFP: { 646 // TODO: use input range if available 647 auto BW = getBitWidth(); 648 APInt Min = APInt::getMinValue(BW).zextOrSelf(ResultBitWidth); 649 APInt Max = APInt::getMaxValue(BW).zextOrSelf(ResultBitWidth); 650 return ConstantRange(std::move(Min), std::move(Max)); 651 } 652 case Instruction::SIToFP: { 653 // TODO: use input range if available 654 auto BW = getBitWidth(); 655 APInt SMin = APInt::getSignedMinValue(BW).sextOrSelf(ResultBitWidth); 656 APInt SMax = APInt::getSignedMaxValue(BW).sextOrSelf(ResultBitWidth); 657 return ConstantRange(std::move(SMin), std::move(SMax)); 658 } 659 case Instruction::FPTrunc: 660 case Instruction::FPExt: 661 case Instruction::IntToPtr: 662 case Instruction::PtrToInt: 663 case Instruction::AddrSpaceCast: 664 // Conservatively return getFull set. 665 return getFull(ResultBitWidth); 666 }; 667 } 668 669 ConstantRange ConstantRange::zeroExtend(uint32_t DstTySize) const { 670 if (isEmptySet()) return getEmpty(DstTySize); 671 672 unsigned SrcTySize = getBitWidth(); 673 assert(SrcTySize < DstTySize && "Not a value extension"); 674 if (isFullSet() || isUpperWrapped()) { 675 // Change into [0, 1 << src bit width) 676 APInt LowerExt(DstTySize, 0); 677 if (!Upper) // special case: [X, 0) -- not really wrapping around 678 LowerExt = Lower.zext(DstTySize); 679 return ConstantRange(std::move(LowerExt), 680 APInt::getOneBitSet(DstTySize, SrcTySize)); 681 } 682 683 return ConstantRange(Lower.zext(DstTySize), Upper.zext(DstTySize)); 684 } 685 686 ConstantRange ConstantRange::signExtend(uint32_t DstTySize) const { 687 if (isEmptySet()) return getEmpty(DstTySize); 688 689 unsigned SrcTySize = getBitWidth(); 690 assert(SrcTySize < DstTySize && "Not a value extension"); 691 692 // special case: [X, INT_MIN) -- not really wrapping around 693 if (Upper.isMinSignedValue()) 694 return ConstantRange(Lower.sext(DstTySize), Upper.zext(DstTySize)); 695 696 if (isFullSet() || isSignWrappedSet()) { 697 return ConstantRange(APInt::getHighBitsSet(DstTySize,DstTySize-SrcTySize+1), 698 APInt::getLowBitsSet(DstTySize, SrcTySize-1) + 1); 699 } 700 701 return ConstantRange(Lower.sext(DstTySize), Upper.sext(DstTySize)); 702 } 703 704 ConstantRange ConstantRange::truncate(uint32_t DstTySize) const { 705 assert(getBitWidth() > DstTySize && "Not a value truncation"); 706 if (isEmptySet()) 707 return getEmpty(DstTySize); 708 if (isFullSet()) 709 return getFull(DstTySize); 710 711 APInt LowerDiv(Lower), UpperDiv(Upper); 712 ConstantRange Union(DstTySize, /*isFullSet=*/false); 713 714 // Analyze wrapped sets in their two parts: [0, Upper) \/ [Lower, MaxValue] 715 // We use the non-wrapped set code to analyze the [Lower, MaxValue) part, and 716 // then we do the union with [MaxValue, Upper) 717 if (isUpperWrapped()) { 718 // If Upper is greater than or equal to MaxValue(DstTy), it covers the whole 719 // truncated range. 720 if (Upper.getActiveBits() > DstTySize || 721 Upper.countTrailingOnes() == DstTySize) 722 return getFull(DstTySize); 723 724 Union = ConstantRange(APInt::getMaxValue(DstTySize),Upper.trunc(DstTySize)); 725 UpperDiv.setAllBits(); 726 727 // Union covers the MaxValue case, so return if the remaining range is just 728 // MaxValue(DstTy). 729 if (LowerDiv == UpperDiv) 730 return Union; 731 } 732 733 // Chop off the most significant bits that are past the destination bitwidth. 734 if (LowerDiv.getActiveBits() > DstTySize) { 735 // Mask to just the signficant bits and subtract from LowerDiv/UpperDiv. 736 APInt Adjust = LowerDiv & APInt::getBitsSetFrom(getBitWidth(), DstTySize); 737 LowerDiv -= Adjust; 738 UpperDiv -= Adjust; 739 } 740 741 unsigned UpperDivWidth = UpperDiv.getActiveBits(); 742 if (UpperDivWidth <= DstTySize) 743 return ConstantRange(LowerDiv.trunc(DstTySize), 744 UpperDiv.trunc(DstTySize)).unionWith(Union); 745 746 // The truncated value wraps around. Check if we can do better than fullset. 747 if (UpperDivWidth == DstTySize + 1) { 748 // Clear the MSB so that UpperDiv wraps around. 749 UpperDiv.clearBit(DstTySize); 750 if (UpperDiv.ult(LowerDiv)) 751 return ConstantRange(LowerDiv.trunc(DstTySize), 752 UpperDiv.trunc(DstTySize)).unionWith(Union); 753 } 754 755 return getFull(DstTySize); 756 } 757 758 ConstantRange ConstantRange::zextOrTrunc(uint32_t DstTySize) const { 759 unsigned SrcTySize = getBitWidth(); 760 if (SrcTySize > DstTySize) 761 return truncate(DstTySize); 762 if (SrcTySize < DstTySize) 763 return zeroExtend(DstTySize); 764 return *this; 765 } 766 767 ConstantRange ConstantRange::sextOrTrunc(uint32_t DstTySize) const { 768 unsigned SrcTySize = getBitWidth(); 769 if (SrcTySize > DstTySize) 770 return truncate(DstTySize); 771 if (SrcTySize < DstTySize) 772 return signExtend(DstTySize); 773 return *this; 774 } 775 776 ConstantRange ConstantRange::binaryOp(Instruction::BinaryOps BinOp, 777 const ConstantRange &Other) const { 778 assert(Instruction::isBinaryOp(BinOp) && "Binary operators only!"); 779 780 switch (BinOp) { 781 case Instruction::Add: 782 return add(Other); 783 case Instruction::Sub: 784 return sub(Other); 785 case Instruction::Mul: 786 return multiply(Other); 787 case Instruction::UDiv: 788 return udiv(Other); 789 case Instruction::SDiv: 790 return sdiv(Other); 791 case Instruction::URem: 792 return urem(Other); 793 case Instruction::SRem: 794 return srem(Other); 795 case Instruction::Shl: 796 return shl(Other); 797 case Instruction::LShr: 798 return lshr(Other); 799 case Instruction::AShr: 800 return ashr(Other); 801 case Instruction::And: 802 return binaryAnd(Other); 803 case Instruction::Or: 804 return binaryOr(Other); 805 // Note: floating point operations applied to abstract ranges are just 806 // ideal integer operations with a lossy representation 807 case Instruction::FAdd: 808 return add(Other); 809 case Instruction::FSub: 810 return sub(Other); 811 case Instruction::FMul: 812 return multiply(Other); 813 default: 814 // Conservatively return getFull set. 815 return getFull(); 816 } 817 } 818 819 ConstantRange ConstantRange::overflowingBinaryOp(Instruction::BinaryOps BinOp, 820 const ConstantRange &Other, 821 unsigned NoWrapKind) const { 822 assert(Instruction::isBinaryOp(BinOp) && "Binary operators only!"); 823 824 switch (BinOp) { 825 case Instruction::Add: 826 return addWithNoWrap(Other, NoWrapKind); 827 case Instruction::Sub: 828 return subWithNoWrap(Other, NoWrapKind); 829 default: 830 // Don't know about this Overflowing Binary Operation. 831 // Conservatively fallback to plain binop handling. 832 return binaryOp(BinOp, Other); 833 } 834 } 835 836 ConstantRange 837 ConstantRange::add(const ConstantRange &Other) const { 838 if (isEmptySet() || Other.isEmptySet()) 839 return getEmpty(); 840 if (isFullSet() || Other.isFullSet()) 841 return getFull(); 842 843 APInt NewLower = getLower() + Other.getLower(); 844 APInt NewUpper = getUpper() + Other.getUpper() - 1; 845 if (NewLower == NewUpper) 846 return getFull(); 847 848 ConstantRange X = ConstantRange(std::move(NewLower), std::move(NewUpper)); 849 if (X.isSizeStrictlySmallerThan(*this) || 850 X.isSizeStrictlySmallerThan(Other)) 851 // We've wrapped, therefore, full set. 852 return getFull(); 853 return X; 854 } 855 856 ConstantRange ConstantRange::addWithNoWrap(const ConstantRange &Other, 857 unsigned NoWrapKind, 858 PreferredRangeType RangeType) const { 859 // Calculate the range for "X + Y" which is guaranteed not to wrap(overflow). 860 // (X is from this, and Y is from Other) 861 if (isEmptySet() || Other.isEmptySet()) 862 return getEmpty(); 863 if (isFullSet() && Other.isFullSet()) 864 return getFull(); 865 866 using OBO = OverflowingBinaryOperator; 867 ConstantRange Result = add(Other); 868 869 // If an overflow happens for every value pair in these two constant ranges, 870 // we must return Empty set. In this case, we get that for free, because we 871 // get lucky that intersection of add() with uadd_sat()/sadd_sat() results 872 // in an empty set. 873 874 if (NoWrapKind & OBO::NoSignedWrap) 875 Result = Result.intersectWith(sadd_sat(Other), RangeType); 876 877 if (NoWrapKind & OBO::NoUnsignedWrap) 878 Result = Result.intersectWith(uadd_sat(Other), RangeType); 879 880 return Result; 881 } 882 883 ConstantRange 884 ConstantRange::sub(const ConstantRange &Other) const { 885 if (isEmptySet() || Other.isEmptySet()) 886 return getEmpty(); 887 if (isFullSet() || Other.isFullSet()) 888 return getFull(); 889 890 APInt NewLower = getLower() - Other.getUpper() + 1; 891 APInt NewUpper = getUpper() - Other.getLower(); 892 if (NewLower == NewUpper) 893 return getFull(); 894 895 ConstantRange X = ConstantRange(std::move(NewLower), std::move(NewUpper)); 896 if (X.isSizeStrictlySmallerThan(*this) || 897 X.isSizeStrictlySmallerThan(Other)) 898 // We've wrapped, therefore, full set. 899 return getFull(); 900 return X; 901 } 902 903 ConstantRange ConstantRange::subWithNoWrap(const ConstantRange &Other, 904 unsigned NoWrapKind, 905 PreferredRangeType RangeType) const { 906 // Calculate the range for "X - Y" which is guaranteed not to wrap(overflow). 907 // (X is from this, and Y is from Other) 908 if (isEmptySet() || Other.isEmptySet()) 909 return getEmpty(); 910 if (isFullSet() && Other.isFullSet()) 911 return getFull(); 912 913 using OBO = OverflowingBinaryOperator; 914 ConstantRange Result = sub(Other); 915 916 // If an overflow happens for every value pair in these two constant ranges, 917 // we must return Empty set. In signed case, we get that for free, because we 918 // get lucky that intersection of sub() with ssub_sat() results in an 919 // empty set. But for unsigned we must perform the overflow check manually. 920 921 if (NoWrapKind & OBO::NoSignedWrap) 922 Result = Result.intersectWith(ssub_sat(Other), RangeType); 923 924 if (NoWrapKind & OBO::NoUnsignedWrap) { 925 if (getUnsignedMax().ult(Other.getUnsignedMin())) 926 return getEmpty(); // Always overflows. 927 Result = Result.intersectWith(usub_sat(Other), RangeType); 928 } 929 930 return Result; 931 } 932 933 ConstantRange 934 ConstantRange::multiply(const ConstantRange &Other) const { 935 // TODO: If either operand is a single element and the multiply is known to 936 // be non-wrapping, round the result min and max value to the appropriate 937 // multiple of that element. If wrapping is possible, at least adjust the 938 // range according to the greatest power-of-two factor of the single element. 939 940 if (isEmptySet() || Other.isEmptySet()) 941 return getEmpty(); 942 943 // Multiplication is signedness-independent. However different ranges can be 944 // obtained depending on how the input ranges are treated. These different 945 // ranges are all conservatively correct, but one might be better than the 946 // other. We calculate two ranges; one treating the inputs as unsigned 947 // and the other signed, then return the smallest of these ranges. 948 949 // Unsigned range first. 950 APInt this_min = getUnsignedMin().zext(getBitWidth() * 2); 951 APInt this_max = getUnsignedMax().zext(getBitWidth() * 2); 952 APInt Other_min = Other.getUnsignedMin().zext(getBitWidth() * 2); 953 APInt Other_max = Other.getUnsignedMax().zext(getBitWidth() * 2); 954 955 ConstantRange Result_zext = ConstantRange(this_min * Other_min, 956 this_max * Other_max + 1); 957 ConstantRange UR = Result_zext.truncate(getBitWidth()); 958 959 // If the unsigned range doesn't wrap, and isn't negative then it's a range 960 // from one positive number to another which is as good as we can generate. 961 // In this case, skip the extra work of generating signed ranges which aren't 962 // going to be better than this range. 963 if (!UR.isUpperWrapped() && 964 (UR.getUpper().isNonNegative() || UR.getUpper().isMinSignedValue())) 965 return UR; 966 967 // Now the signed range. Because we could be dealing with negative numbers 968 // here, the lower bound is the smallest of the cartesian product of the 969 // lower and upper ranges; for example: 970 // [-1,4) * [-2,3) = min(-1*-2, -1*2, 3*-2, 3*2) = -6. 971 // Similarly for the upper bound, swapping min for max. 972 973 this_min = getSignedMin().sext(getBitWidth() * 2); 974 this_max = getSignedMax().sext(getBitWidth() * 2); 975 Other_min = Other.getSignedMin().sext(getBitWidth() * 2); 976 Other_max = Other.getSignedMax().sext(getBitWidth() * 2); 977 978 auto L = {this_min * Other_min, this_min * Other_max, 979 this_max * Other_min, this_max * Other_max}; 980 auto Compare = [](const APInt &A, const APInt &B) { return A.slt(B); }; 981 ConstantRange Result_sext(std::min(L, Compare), std::max(L, Compare) + 1); 982 ConstantRange SR = Result_sext.truncate(getBitWidth()); 983 984 return UR.isSizeStrictlySmallerThan(SR) ? UR : SR; 985 } 986 987 ConstantRange 988 ConstantRange::smax(const ConstantRange &Other) const { 989 // X smax Y is: range(smax(X_smin, Y_smin), 990 // smax(X_smax, Y_smax)) 991 if (isEmptySet() || Other.isEmptySet()) 992 return getEmpty(); 993 APInt NewL = APIntOps::smax(getSignedMin(), Other.getSignedMin()); 994 APInt NewU = APIntOps::smax(getSignedMax(), Other.getSignedMax()) + 1; 995 return getNonEmpty(std::move(NewL), std::move(NewU)); 996 } 997 998 ConstantRange 999 ConstantRange::umax(const ConstantRange &Other) const { 1000 // X umax Y is: range(umax(X_umin, Y_umin), 1001 // umax(X_umax, Y_umax)) 1002 if (isEmptySet() || Other.isEmptySet()) 1003 return getEmpty(); 1004 APInt NewL = APIntOps::umax(getUnsignedMin(), Other.getUnsignedMin()); 1005 APInt NewU = APIntOps::umax(getUnsignedMax(), Other.getUnsignedMax()) + 1; 1006 return getNonEmpty(std::move(NewL), std::move(NewU)); 1007 } 1008 1009 ConstantRange 1010 ConstantRange::smin(const ConstantRange &Other) const { 1011 // X smin Y is: range(smin(X_smin, Y_smin), 1012 // smin(X_smax, Y_smax)) 1013 if (isEmptySet() || Other.isEmptySet()) 1014 return getEmpty(); 1015 APInt NewL = APIntOps::smin(getSignedMin(), Other.getSignedMin()); 1016 APInt NewU = APIntOps::smin(getSignedMax(), Other.getSignedMax()) + 1; 1017 return getNonEmpty(std::move(NewL), std::move(NewU)); 1018 } 1019 1020 ConstantRange 1021 ConstantRange::umin(const ConstantRange &Other) const { 1022 // X umin Y is: range(umin(X_umin, Y_umin), 1023 // umin(X_umax, Y_umax)) 1024 if (isEmptySet() || Other.isEmptySet()) 1025 return getEmpty(); 1026 APInt NewL = APIntOps::umin(getUnsignedMin(), Other.getUnsignedMin()); 1027 APInt NewU = APIntOps::umin(getUnsignedMax(), Other.getUnsignedMax()) + 1; 1028 return getNonEmpty(std::move(NewL), std::move(NewU)); 1029 } 1030 1031 ConstantRange 1032 ConstantRange::udiv(const ConstantRange &RHS) const { 1033 if (isEmptySet() || RHS.isEmptySet() || RHS.getUnsignedMax().isNullValue()) 1034 return getEmpty(); 1035 1036 APInt Lower = getUnsignedMin().udiv(RHS.getUnsignedMax()); 1037 1038 APInt RHS_umin = RHS.getUnsignedMin(); 1039 if (RHS_umin.isNullValue()) { 1040 // We want the lowest value in RHS excluding zero. Usually that would be 1 1041 // except for a range in the form of [X, 1) in which case it would be X. 1042 if (RHS.getUpper() == 1) 1043 RHS_umin = RHS.getLower(); 1044 else 1045 RHS_umin = 1; 1046 } 1047 1048 APInt Upper = getUnsignedMax().udiv(RHS_umin) + 1; 1049 return getNonEmpty(std::move(Lower), std::move(Upper)); 1050 } 1051 1052 ConstantRange ConstantRange::sdiv(const ConstantRange &RHS) const { 1053 // We split up the LHS and RHS into positive and negative components 1054 // and then also compute the positive and negative components of the result 1055 // separately by combining division results with the appropriate signs. 1056 APInt Zero = APInt::getNullValue(getBitWidth()); 1057 APInt SignedMin = APInt::getSignedMinValue(getBitWidth()); 1058 ConstantRange PosFilter(APInt(getBitWidth(), 1), SignedMin); 1059 ConstantRange NegFilter(SignedMin, Zero); 1060 ConstantRange PosL = intersectWith(PosFilter); 1061 ConstantRange NegL = intersectWith(NegFilter); 1062 ConstantRange PosR = RHS.intersectWith(PosFilter); 1063 ConstantRange NegR = RHS.intersectWith(NegFilter); 1064 1065 ConstantRange PosRes = getEmpty(); 1066 if (!PosL.isEmptySet() && !PosR.isEmptySet()) 1067 // pos / pos = pos. 1068 PosRes = ConstantRange(PosL.Lower.sdiv(PosR.Upper - 1), 1069 (PosL.Upper - 1).sdiv(PosR.Lower) + 1); 1070 1071 if (!NegL.isEmptySet() && !NegR.isEmptySet()) { 1072 // neg / neg = pos. 1073 // 1074 // We need to deal with one tricky case here: SignedMin / -1 is UB on the 1075 // IR level, so we'll want to exclude this case when calculating bounds. 1076 // (For APInts the operation is well-defined and yields SignedMin.) We 1077 // handle this by dropping either SignedMin from the LHS or -1 from the RHS. 1078 APInt Lo = (NegL.Upper - 1).sdiv(NegR.Lower); 1079 if (NegL.Lower.isMinSignedValue() && NegR.Upper.isNullValue()) { 1080 // Remove -1 from the LHS. Skip if it's the only element, as this would 1081 // leave us with an empty set. 1082 if (!NegR.Lower.isAllOnesValue()) { 1083 APInt AdjNegRUpper; 1084 if (RHS.Lower.isAllOnesValue()) 1085 // Negative part of [-1, X] without -1 is [SignedMin, X]. 1086 AdjNegRUpper = RHS.Upper; 1087 else 1088 // [X, -1] without -1 is [X, -2]. 1089 AdjNegRUpper = NegR.Upper - 1; 1090 1091 PosRes = PosRes.unionWith( 1092 ConstantRange(Lo, NegL.Lower.sdiv(AdjNegRUpper - 1) + 1)); 1093 } 1094 1095 // Remove SignedMin from the RHS. Skip if it's the only element, as this 1096 // would leave us with an empty set. 1097 if (NegL.Upper != SignedMin + 1) { 1098 APInt AdjNegLLower; 1099 if (Upper == SignedMin + 1) 1100 // Negative part of [X, SignedMin] without SignedMin is [X, -1]. 1101 AdjNegLLower = Lower; 1102 else 1103 // [SignedMin, X] without SignedMin is [SignedMin + 1, X]. 1104 AdjNegLLower = NegL.Lower + 1; 1105 1106 PosRes = PosRes.unionWith( 1107 ConstantRange(std::move(Lo), 1108 AdjNegLLower.sdiv(NegR.Upper - 1) + 1)); 1109 } 1110 } else { 1111 PosRes = PosRes.unionWith( 1112 ConstantRange(std::move(Lo), NegL.Lower.sdiv(NegR.Upper - 1) + 1)); 1113 } 1114 } 1115 1116 ConstantRange NegRes = getEmpty(); 1117 if (!PosL.isEmptySet() && !NegR.isEmptySet()) 1118 // pos / neg = neg. 1119 NegRes = ConstantRange((PosL.Upper - 1).sdiv(NegR.Upper - 1), 1120 PosL.Lower.sdiv(NegR.Lower) + 1); 1121 1122 if (!NegL.isEmptySet() && !PosR.isEmptySet()) 1123 // neg / pos = neg. 1124 NegRes = NegRes.unionWith( 1125 ConstantRange(NegL.Lower.sdiv(PosR.Lower), 1126 (NegL.Upper - 1).sdiv(PosR.Upper - 1) + 1)); 1127 1128 // Prefer a non-wrapping signed range here. 1129 ConstantRange Res = NegRes.unionWith(PosRes, PreferredRangeType::Signed); 1130 1131 // Preserve the zero that we dropped when splitting the LHS by sign. 1132 if (contains(Zero) && (!PosR.isEmptySet() || !NegR.isEmptySet())) 1133 Res = Res.unionWith(ConstantRange(Zero)); 1134 return Res; 1135 } 1136 1137 ConstantRange ConstantRange::urem(const ConstantRange &RHS) const { 1138 if (isEmptySet() || RHS.isEmptySet() || RHS.getUnsignedMax().isNullValue()) 1139 return getEmpty(); 1140 1141 // L % R for L < R is L. 1142 if (getUnsignedMax().ult(RHS.getUnsignedMin())) 1143 return *this; 1144 1145 // L % R is <= L and < R. 1146 APInt Upper = APIntOps::umin(getUnsignedMax(), RHS.getUnsignedMax() - 1) + 1; 1147 return getNonEmpty(APInt::getNullValue(getBitWidth()), std::move(Upper)); 1148 } 1149 1150 ConstantRange ConstantRange::srem(const ConstantRange &RHS) const { 1151 if (isEmptySet() || RHS.isEmptySet()) 1152 return getEmpty(); 1153 1154 ConstantRange AbsRHS = RHS.abs(); 1155 APInt MinAbsRHS = AbsRHS.getUnsignedMin(); 1156 APInt MaxAbsRHS = AbsRHS.getUnsignedMax(); 1157 1158 // Modulus by zero is UB. 1159 if (MaxAbsRHS.isNullValue()) 1160 return getEmpty(); 1161 1162 if (MinAbsRHS.isNullValue()) 1163 ++MinAbsRHS; 1164 1165 APInt MinLHS = getSignedMin(), MaxLHS = getSignedMax(); 1166 1167 if (MinLHS.isNonNegative()) { 1168 // L % R for L < R is L. 1169 if (MaxLHS.ult(MinAbsRHS)) 1170 return *this; 1171 1172 // L % R is <= L and < R. 1173 APInt Upper = APIntOps::umin(MaxLHS, MaxAbsRHS - 1) + 1; 1174 return ConstantRange(APInt::getNullValue(getBitWidth()), std::move(Upper)); 1175 } 1176 1177 // Same basic logic as above, but the result is negative. 1178 if (MaxLHS.isNegative()) { 1179 if (MinLHS.ugt(-MinAbsRHS)) 1180 return *this; 1181 1182 APInt Lower = APIntOps::umax(MinLHS, -MaxAbsRHS + 1); 1183 return ConstantRange(std::move(Lower), APInt(getBitWidth(), 1)); 1184 } 1185 1186 // LHS range crosses zero. 1187 APInt Lower = APIntOps::umax(MinLHS, -MaxAbsRHS + 1); 1188 APInt Upper = APIntOps::umin(MaxLHS, MaxAbsRHS - 1) + 1; 1189 return ConstantRange(std::move(Lower), std::move(Upper)); 1190 } 1191 1192 ConstantRange 1193 ConstantRange::binaryAnd(const ConstantRange &Other) const { 1194 if (isEmptySet() || Other.isEmptySet()) 1195 return getEmpty(); 1196 1197 // TODO: replace this with something less conservative 1198 1199 APInt umin = APIntOps::umin(Other.getUnsignedMax(), getUnsignedMax()); 1200 return getNonEmpty(APInt::getNullValue(getBitWidth()), std::move(umin) + 1); 1201 } 1202 1203 ConstantRange 1204 ConstantRange::binaryOr(const ConstantRange &Other) const { 1205 if (isEmptySet() || Other.isEmptySet()) 1206 return getEmpty(); 1207 1208 // TODO: replace this with something less conservative 1209 1210 APInt umax = APIntOps::umax(getUnsignedMin(), Other.getUnsignedMin()); 1211 return getNonEmpty(std::move(umax), APInt::getNullValue(getBitWidth())); 1212 } 1213 1214 ConstantRange 1215 ConstantRange::shl(const ConstantRange &Other) const { 1216 if (isEmptySet() || Other.isEmptySet()) 1217 return getEmpty(); 1218 1219 APInt max = getUnsignedMax(); 1220 APInt Other_umax = Other.getUnsignedMax(); 1221 1222 // If we are shifting by maximum amount of 1223 // zero return return the original range. 1224 if (Other_umax.isNullValue()) 1225 return *this; 1226 // there's overflow! 1227 if (Other_umax.ugt(max.countLeadingZeros())) 1228 return getFull(); 1229 1230 // FIXME: implement the other tricky cases 1231 1232 APInt min = getUnsignedMin(); 1233 min <<= Other.getUnsignedMin(); 1234 max <<= Other_umax; 1235 1236 return ConstantRange(std::move(min), std::move(max) + 1); 1237 } 1238 1239 ConstantRange 1240 ConstantRange::lshr(const ConstantRange &Other) const { 1241 if (isEmptySet() || Other.isEmptySet()) 1242 return getEmpty(); 1243 1244 APInt max = getUnsignedMax().lshr(Other.getUnsignedMin()) + 1; 1245 APInt min = getUnsignedMin().lshr(Other.getUnsignedMax()); 1246 return getNonEmpty(std::move(min), std::move(max)); 1247 } 1248 1249 ConstantRange 1250 ConstantRange::ashr(const ConstantRange &Other) const { 1251 if (isEmptySet() || Other.isEmptySet()) 1252 return getEmpty(); 1253 1254 // May straddle zero, so handle both positive and negative cases. 1255 // 'PosMax' is the upper bound of the result of the ashr 1256 // operation, when Upper of the LHS of ashr is a non-negative. 1257 // number. Since ashr of a non-negative number will result in a 1258 // smaller number, the Upper value of LHS is shifted right with 1259 // the minimum value of 'Other' instead of the maximum value. 1260 APInt PosMax = getSignedMax().ashr(Other.getUnsignedMin()) + 1; 1261 1262 // 'PosMin' is the lower bound of the result of the ashr 1263 // operation, when Lower of the LHS is a non-negative number. 1264 // Since ashr of a non-negative number will result in a smaller 1265 // number, the Lower value of LHS is shifted right with the 1266 // maximum value of 'Other'. 1267 APInt PosMin = getSignedMin().ashr(Other.getUnsignedMax()); 1268 1269 // 'NegMax' is the upper bound of the result of the ashr 1270 // operation, when Upper of the LHS of ashr is a negative number. 1271 // Since 'ashr' of a negative number will result in a bigger 1272 // number, the Upper value of LHS is shifted right with the 1273 // maximum value of 'Other'. 1274 APInt NegMax = getSignedMax().ashr(Other.getUnsignedMax()) + 1; 1275 1276 // 'NegMin' is the lower bound of the result of the ashr 1277 // operation, when Lower of the LHS of ashr is a negative number. 1278 // Since 'ashr' of a negative number will result in a bigger 1279 // number, the Lower value of LHS is shifted right with the 1280 // minimum value of 'Other'. 1281 APInt NegMin = getSignedMin().ashr(Other.getUnsignedMin()); 1282 1283 APInt max, min; 1284 if (getSignedMin().isNonNegative()) { 1285 // Upper and Lower of LHS are non-negative. 1286 min = PosMin; 1287 max = PosMax; 1288 } else if (getSignedMax().isNegative()) { 1289 // Upper and Lower of LHS are negative. 1290 min = NegMin; 1291 max = NegMax; 1292 } else { 1293 // Upper is non-negative and Lower is negative. 1294 min = NegMin; 1295 max = PosMax; 1296 } 1297 return getNonEmpty(std::move(min), std::move(max)); 1298 } 1299 1300 ConstantRange ConstantRange::uadd_sat(const ConstantRange &Other) const { 1301 if (isEmptySet() || Other.isEmptySet()) 1302 return getEmpty(); 1303 1304 APInt NewL = getUnsignedMin().uadd_sat(Other.getUnsignedMin()); 1305 APInt NewU = getUnsignedMax().uadd_sat(Other.getUnsignedMax()) + 1; 1306 return getNonEmpty(std::move(NewL), std::move(NewU)); 1307 } 1308 1309 ConstantRange ConstantRange::sadd_sat(const ConstantRange &Other) const { 1310 if (isEmptySet() || Other.isEmptySet()) 1311 return getEmpty(); 1312 1313 APInt NewL = getSignedMin().sadd_sat(Other.getSignedMin()); 1314 APInt NewU = getSignedMax().sadd_sat(Other.getSignedMax()) + 1; 1315 return getNonEmpty(std::move(NewL), std::move(NewU)); 1316 } 1317 1318 ConstantRange ConstantRange::usub_sat(const ConstantRange &Other) const { 1319 if (isEmptySet() || Other.isEmptySet()) 1320 return getEmpty(); 1321 1322 APInt NewL = getUnsignedMin().usub_sat(Other.getUnsignedMax()); 1323 APInt NewU = getUnsignedMax().usub_sat(Other.getUnsignedMin()) + 1; 1324 return getNonEmpty(std::move(NewL), std::move(NewU)); 1325 } 1326 1327 ConstantRange ConstantRange::ssub_sat(const ConstantRange &Other) const { 1328 if (isEmptySet() || Other.isEmptySet()) 1329 return getEmpty(); 1330 1331 APInt NewL = getSignedMin().ssub_sat(Other.getSignedMax()); 1332 APInt NewU = getSignedMax().ssub_sat(Other.getSignedMin()) + 1; 1333 return getNonEmpty(std::move(NewL), std::move(NewU)); 1334 } 1335 1336 ConstantRange ConstantRange::umul_sat(const ConstantRange &Other) const { 1337 if (isEmptySet() || Other.isEmptySet()) 1338 return getEmpty(); 1339 1340 APInt NewL = getUnsignedMin().umul_sat(Other.getUnsignedMin()); 1341 APInt NewU = getUnsignedMax().umul_sat(Other.getUnsignedMax()) + 1; 1342 return getNonEmpty(std::move(NewL), std::move(NewU)); 1343 } 1344 1345 ConstantRange ConstantRange::smul_sat(const ConstantRange &Other) const { 1346 if (isEmptySet() || Other.isEmptySet()) 1347 return getEmpty(); 1348 1349 // Because we could be dealing with negative numbers here, the lower bound is 1350 // the smallest of the cartesian product of the lower and upper ranges; 1351 // for example: 1352 // [-1,4) * [-2,3) = min(-1*-2, -1*2, 3*-2, 3*2) = -6. 1353 // Similarly for the upper bound, swapping min for max. 1354 1355 APInt this_min = getSignedMin().sext(getBitWidth() * 2); 1356 APInt this_max = getSignedMax().sext(getBitWidth() * 2); 1357 APInt Other_min = Other.getSignedMin().sext(getBitWidth() * 2); 1358 APInt Other_max = Other.getSignedMax().sext(getBitWidth() * 2); 1359 1360 auto L = {this_min * Other_min, this_min * Other_max, this_max * Other_min, 1361 this_max * Other_max}; 1362 auto Compare = [](const APInt &A, const APInt &B) { return A.slt(B); }; 1363 1364 // Note that we wanted to perform signed saturating multiplication, 1365 // so since we performed plain multiplication in twice the bitwidth, 1366 // we need to perform signed saturating truncation. 1367 return getNonEmpty(std::min(L, Compare).truncSSat(getBitWidth()), 1368 std::max(L, Compare).truncSSat(getBitWidth()) + 1); 1369 } 1370 1371 ConstantRange ConstantRange::ushl_sat(const ConstantRange &Other) const { 1372 if (isEmptySet() || Other.isEmptySet()) 1373 return getEmpty(); 1374 1375 APInt NewL = getUnsignedMin().ushl_sat(Other.getUnsignedMin()); 1376 APInt NewU = getUnsignedMax().ushl_sat(Other.getUnsignedMax()) + 1; 1377 return getNonEmpty(std::move(NewL), std::move(NewU)); 1378 } 1379 1380 ConstantRange ConstantRange::sshl_sat(const ConstantRange &Other) const { 1381 if (isEmptySet() || Other.isEmptySet()) 1382 return getEmpty(); 1383 1384 APInt Min = getSignedMin(), Max = getSignedMax(); 1385 APInt ShAmtMin = Other.getUnsignedMin(), ShAmtMax = Other.getUnsignedMax(); 1386 APInt NewL = Min.sshl_sat(Min.isNonNegative() ? ShAmtMin : ShAmtMax); 1387 APInt NewU = Max.sshl_sat(Max.isNegative() ? ShAmtMin : ShAmtMax) + 1; 1388 return getNonEmpty(std::move(NewL), std::move(NewU)); 1389 } 1390 1391 ConstantRange ConstantRange::inverse() const { 1392 if (isFullSet()) 1393 return getEmpty(); 1394 if (isEmptySet()) 1395 return getFull(); 1396 return ConstantRange(Upper, Lower); 1397 } 1398 1399 ConstantRange ConstantRange::abs() const { 1400 if (isEmptySet()) 1401 return getEmpty(); 1402 1403 if (isSignWrappedSet()) { 1404 APInt Lo; 1405 // Check whether the range crosses zero. 1406 if (Upper.isStrictlyPositive() || !Lower.isStrictlyPositive()) 1407 Lo = APInt::getNullValue(getBitWidth()); 1408 else 1409 Lo = APIntOps::umin(Lower, -Upper + 1); 1410 1411 // SignedMin is included in the result range. 1412 return ConstantRange(Lo, APInt::getSignedMinValue(getBitWidth()) + 1); 1413 } 1414 1415 APInt SMin = getSignedMin(), SMax = getSignedMax(); 1416 1417 // All non-negative. 1418 if (SMin.isNonNegative()) 1419 return *this; 1420 1421 // All negative. 1422 if (SMax.isNegative()) 1423 return ConstantRange(-SMax, -SMin + 1); 1424 1425 // Range crosses zero. 1426 return ConstantRange(APInt::getNullValue(getBitWidth()), 1427 APIntOps::umax(-SMin, SMax) + 1); 1428 } 1429 1430 ConstantRange::OverflowResult ConstantRange::unsignedAddMayOverflow( 1431 const ConstantRange &Other) const { 1432 if (isEmptySet() || Other.isEmptySet()) 1433 return OverflowResult::MayOverflow; 1434 1435 APInt Min = getUnsignedMin(), Max = getUnsignedMax(); 1436 APInt OtherMin = Other.getUnsignedMin(), OtherMax = Other.getUnsignedMax(); 1437 1438 // a u+ b overflows high iff a u> ~b. 1439 if (Min.ugt(~OtherMin)) 1440 return OverflowResult::AlwaysOverflowsHigh; 1441 if (Max.ugt(~OtherMax)) 1442 return OverflowResult::MayOverflow; 1443 return OverflowResult::NeverOverflows; 1444 } 1445 1446 ConstantRange::OverflowResult ConstantRange::signedAddMayOverflow( 1447 const ConstantRange &Other) const { 1448 if (isEmptySet() || Other.isEmptySet()) 1449 return OverflowResult::MayOverflow; 1450 1451 APInt Min = getSignedMin(), Max = getSignedMax(); 1452 APInt OtherMin = Other.getSignedMin(), OtherMax = Other.getSignedMax(); 1453 1454 APInt SignedMin = APInt::getSignedMinValue(getBitWidth()); 1455 APInt SignedMax = APInt::getSignedMaxValue(getBitWidth()); 1456 1457 // a s+ b overflows high iff a s>=0 && b s>= 0 && a s> smax - b. 1458 // a s+ b overflows low iff a s< 0 && b s< 0 && a s< smin - b. 1459 if (Min.isNonNegative() && OtherMin.isNonNegative() && 1460 Min.sgt(SignedMax - OtherMin)) 1461 return OverflowResult::AlwaysOverflowsHigh; 1462 if (Max.isNegative() && OtherMax.isNegative() && 1463 Max.slt(SignedMin - OtherMax)) 1464 return OverflowResult::AlwaysOverflowsLow; 1465 1466 if (Max.isNonNegative() && OtherMax.isNonNegative() && 1467 Max.sgt(SignedMax - OtherMax)) 1468 return OverflowResult::MayOverflow; 1469 if (Min.isNegative() && OtherMin.isNegative() && 1470 Min.slt(SignedMin - OtherMin)) 1471 return OverflowResult::MayOverflow; 1472 1473 return OverflowResult::NeverOverflows; 1474 } 1475 1476 ConstantRange::OverflowResult ConstantRange::unsignedSubMayOverflow( 1477 const ConstantRange &Other) const { 1478 if (isEmptySet() || Other.isEmptySet()) 1479 return OverflowResult::MayOverflow; 1480 1481 APInt Min = getUnsignedMin(), Max = getUnsignedMax(); 1482 APInt OtherMin = Other.getUnsignedMin(), OtherMax = Other.getUnsignedMax(); 1483 1484 // a u- b overflows low iff a u< b. 1485 if (Max.ult(OtherMin)) 1486 return OverflowResult::AlwaysOverflowsLow; 1487 if (Min.ult(OtherMax)) 1488 return OverflowResult::MayOverflow; 1489 return OverflowResult::NeverOverflows; 1490 } 1491 1492 ConstantRange::OverflowResult ConstantRange::signedSubMayOverflow( 1493 const ConstantRange &Other) const { 1494 if (isEmptySet() || Other.isEmptySet()) 1495 return OverflowResult::MayOverflow; 1496 1497 APInt Min = getSignedMin(), Max = getSignedMax(); 1498 APInt OtherMin = Other.getSignedMin(), OtherMax = Other.getSignedMax(); 1499 1500 APInt SignedMin = APInt::getSignedMinValue(getBitWidth()); 1501 APInt SignedMax = APInt::getSignedMaxValue(getBitWidth()); 1502 1503 // a s- b overflows high iff a s>=0 && b s< 0 && a s> smax + b. 1504 // a s- b overflows low iff a s< 0 && b s>= 0 && a s< smin + b. 1505 if (Min.isNonNegative() && OtherMax.isNegative() && 1506 Min.sgt(SignedMax + OtherMax)) 1507 return OverflowResult::AlwaysOverflowsHigh; 1508 if (Max.isNegative() && OtherMin.isNonNegative() && 1509 Max.slt(SignedMin + OtherMin)) 1510 return OverflowResult::AlwaysOverflowsLow; 1511 1512 if (Max.isNonNegative() && OtherMin.isNegative() && 1513 Max.sgt(SignedMax + OtherMin)) 1514 return OverflowResult::MayOverflow; 1515 if (Min.isNegative() && OtherMax.isNonNegative() && 1516 Min.slt(SignedMin + OtherMax)) 1517 return OverflowResult::MayOverflow; 1518 1519 return OverflowResult::NeverOverflows; 1520 } 1521 1522 ConstantRange::OverflowResult ConstantRange::unsignedMulMayOverflow( 1523 const ConstantRange &Other) const { 1524 if (isEmptySet() || Other.isEmptySet()) 1525 return OverflowResult::MayOverflow; 1526 1527 APInt Min = getUnsignedMin(), Max = getUnsignedMax(); 1528 APInt OtherMin = Other.getUnsignedMin(), OtherMax = Other.getUnsignedMax(); 1529 bool Overflow; 1530 1531 (void) Min.umul_ov(OtherMin, Overflow); 1532 if (Overflow) 1533 return OverflowResult::AlwaysOverflowsHigh; 1534 1535 (void) Max.umul_ov(OtherMax, Overflow); 1536 if (Overflow) 1537 return OverflowResult::MayOverflow; 1538 1539 return OverflowResult::NeverOverflows; 1540 } 1541 1542 void ConstantRange::print(raw_ostream &OS) const { 1543 if (isFullSet()) 1544 OS << "full-set"; 1545 else if (isEmptySet()) 1546 OS << "empty-set"; 1547 else 1548 OS << "[" << Lower << "," << Upper << ")"; 1549 } 1550 1551 #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP) 1552 LLVM_DUMP_METHOD void ConstantRange::dump() const { 1553 print(dbgs()); 1554 } 1555 #endif 1556 1557 ConstantRange llvm::getConstantRangeFromMetadata(const MDNode &Ranges) { 1558 const unsigned NumRanges = Ranges.getNumOperands() / 2; 1559 assert(NumRanges >= 1 && "Must have at least one range!"); 1560 assert(Ranges.getNumOperands() % 2 == 0 && "Must be a sequence of pairs"); 1561 1562 auto *FirstLow = mdconst::extract<ConstantInt>(Ranges.getOperand(0)); 1563 auto *FirstHigh = mdconst::extract<ConstantInt>(Ranges.getOperand(1)); 1564 1565 ConstantRange CR(FirstLow->getValue(), FirstHigh->getValue()); 1566 1567 for (unsigned i = 1; i < NumRanges; ++i) { 1568 auto *Low = mdconst::extract<ConstantInt>(Ranges.getOperand(2 * i + 0)); 1569 auto *High = mdconst::extract<ConstantInt>(Ranges.getOperand(2 * i + 1)); 1570 1571 // Note: unionWith will potentially create a range that contains values not 1572 // contained in any of the original N ranges. 1573 CR = CR.unionWith(ConstantRange(Low->getValue(), High->getValue())); 1574 } 1575 1576 return CR; 1577 } 1578