LLVM  8.0.0svn
ConstantRange.h
Go to the documentation of this file.
1 //===- ConstantRange.h - Represent a range ----------------------*- C++ -*-===//
2 //
3 // The LLVM Compiler Infrastructure
4 //
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
7 //
8 //===----------------------------------------------------------------------===//
9 //
10 // Represent a range of possible values that may occur when the program is run
11 // for an integral value. This keeps track of a lower and upper bound for the
12 // constant, which MAY wrap around the end of the numeric range. To do this, it
13 // keeps track of a [lower, upper) bound, which specifies an interval just like
14 // STL iterators. When used with boolean values, the following are important
15 // ranges: :
16 //
17 // [F, F) = {} = Empty set
18 // [T, F) = {T}
19 // [F, T) = {F}
20 // [T, T) = {F, T} = Full set
21 //
22 // The other integral ranges use min/max values for special range values. For
23 // example, for 8-bit types, it uses:
24 // [0, 0) = {} = Empty set
25 // [255, 255) = {0..255} = Full Set
26 //
27 // Note that ConstantRange can be used to represent either signed or
28 // unsigned ranges.
29 //
30 //===----------------------------------------------------------------------===//
31 
32 #ifndef LLVM_IR_CONSTANTRANGE_H
33 #define LLVM_IR_CONSTANTRANGE_H
34 
35 #include "llvm/ADT/APInt.h"
36 #include "llvm/IR/InstrTypes.h"
37 #include "llvm/IR/Instruction.h"
38 #include "llvm/Support/Compiler.h"
39 #include <cstdint>
40 
41 namespace llvm {
42 
43 class MDNode;
44 class raw_ostream;
45 
46 /// This class represents a range of values.
48  APInt Lower, Upper;
49 
50 public:
51  /// Initialize a full (the default) or empty set for the specified bit width.
52  explicit ConstantRange(uint32_t BitWidth, bool isFullSet = true);
53 
54  /// Initialize a range to hold the single specified value.
56 
57  /// Initialize a range of values explicitly. This will assert out if
58  /// Lower==Upper and Lower != Min or Max value for its type. It will also
59  /// assert out if the two APInt's are not the same bit width.
60  ConstantRange(APInt Lower, APInt Upper);
61 
62  /// Produce the smallest range such that all values that may satisfy the given
63  /// predicate with any value contained within Other is contained in the
64  /// returned range. Formally, this returns a superset of
65  /// 'union over all y in Other . { x : icmp op x y is true }'. If the exact
66  /// answer is not representable as a ConstantRange, the return value will be a
67  /// proper superset of the above.
68  ///
69  /// Example: Pred = ult and Other = i8 [2, 5) returns Result = [0, 4)
70  static ConstantRange makeAllowedICmpRegion(CmpInst::Predicate Pred,
71  const ConstantRange &Other);
72 
73  /// Produce the largest range such that all values in the returned range
74  /// satisfy the given predicate with all values contained within Other.
75  /// Formally, this returns a subset of
76  /// 'intersection over all y in Other . { x : icmp op x y is true }'. If the
77  /// exact answer is not representable as a ConstantRange, the return value
78  /// will be a proper subset of the above.
79  ///
80  /// Example: Pred = ult and Other = i8 [2, 5) returns [0, 2)
81  static ConstantRange makeSatisfyingICmpRegion(CmpInst::Predicate Pred,
82  const ConstantRange &Other);
83 
84  /// Produce the exact range such that all values in the returned range satisfy
85  /// the given predicate with any value contained within Other. Formally, this
86  /// returns the exact answer when the superset of 'union over all y in Other
87  /// is exactly same as the subset of intersection over all y in Other.
88  /// { x : icmp op x y is true}'.
89  ///
90  /// Example: Pred = ult and Other = i8 3 returns [0, 3)
91  static ConstantRange makeExactICmpRegion(CmpInst::Predicate Pred,
92  const APInt &Other);
93 
94  /// Return the largest range containing all X such that "X BinOpC Y" is
95  /// guaranteed not to wrap (overflow) for all Y in Other.
96  ///
97  /// NB! The returned set does *not* contain **all** possible values of X for
98  /// which "X BinOpC Y" does not wrap -- some viable values of X may be
99  /// missing, so you cannot use this to constrain X's range. E.g. in the
100  /// fourth example, "(-2) + 1" is both nsw and nuw (so the "X" could be -2),
101  /// but (-2) is not in the set returned.
102  ///
103  /// Examples:
104  /// typedef OverflowingBinaryOperator OBO;
105  /// #define MGNR makeGuaranteedNoWrapRegion
106  /// MGNR(Add, [i8 1, 2), OBO::NoSignedWrap) == [-128, 127)
107  /// MGNR(Add, [i8 1, 2), OBO::NoUnsignedWrap) == [0, -1)
108  /// MGNR(Add, [i8 0, 1), OBO::NoUnsignedWrap) == Full Set
109  /// MGNR(Add, [i8 1, 2), OBO::NoUnsignedWrap | OBO::NoSignedWrap)
110  /// == [0,INT_MAX)
111  /// MGNR(Add, [i8 -1, 6), OBO::NoSignedWrap) == [INT_MIN+1, INT_MAX-4)
112  /// MGNR(Sub, [i8 1, 2), OBO::NoSignedWrap) == [-127, 128)
113  /// MGNR(Sub, [i8 1, 2), OBO::NoUnsignedWrap) == [1, 0)
114  /// MGNR(Sub, [i8 1, 2), OBO::NoUnsignedWrap | OBO::NoSignedWrap)
115  /// == [1,INT_MAX)
116  static ConstantRange makeGuaranteedNoWrapRegion(Instruction::BinaryOps BinOp,
117  const ConstantRange &Other,
118  unsigned NoWrapKind);
119 
120  /// Set up \p Pred and \p RHS such that
121  /// ConstantRange::makeExactICmpRegion(Pred, RHS) == *this. Return true if
122  /// successful.
123  bool getEquivalentICmp(CmpInst::Predicate &Pred, APInt &RHS) const;
124 
125  /// Return the lower value for this range.
126  const APInt &getLower() const { return Lower; }
127 
128  /// Return the upper value for this range.
129  const APInt &getUpper() const { return Upper; }
130 
131  /// Get the bit width of this ConstantRange.
132  uint32_t getBitWidth() const { return Lower.getBitWidth(); }
133 
134  /// Return true if this set contains all of the elements possible
135  /// for this data-type.
136  bool isFullSet() const;
137 
138  /// Return true if this set contains no members.
139  bool isEmptySet() const;
140 
141  /// Return true if this set wraps around the top of the range.
142  /// For example: [100, 8).
143  bool isWrappedSet() const;
144 
145  /// Return true if this set wraps around the INT_MIN of
146  /// its bitwidth. For example: i8 [120, 140).
147  bool isSignWrappedSet() const;
148 
149  /// Return true if the specified value is in the set.
150  bool contains(const APInt &Val) const;
151 
152  /// Return true if the other range is a subset of this one.
153  bool contains(const ConstantRange &CR) const;
154 
155  /// If this set contains a single element, return it, otherwise return null.
156  const APInt *getSingleElement() const {
157  if (Upper == Lower + 1)
158  return &Lower;
159  return nullptr;
160  }
161 
162  /// If this set contains all but a single element, return it, otherwise return
163  /// null.
165  if (Lower == Upper + 1)
166  return &Upper;
167  return nullptr;
168  }
169 
170  /// Return true if this set contains exactly one member.
171  bool isSingleElement() const { return getSingleElement() != nullptr; }
172 
173  /// Return the number of elements in this set.
174  APInt getSetSize() const;
175 
176  /// Compare set size of this range with the range CR.
177  bool isSizeStrictlySmallerThan(const ConstantRange &CR) const;
178 
179  // Compare set size of this range with Value.
180  bool isSizeLargerThan(uint64_t MaxSize) const;
181 
182  /// Return the largest unsigned value contained in the ConstantRange.
183  APInt getUnsignedMax() const;
184 
185  /// Return the smallest unsigned value contained in the ConstantRange.
186  APInt getUnsignedMin() const;
187 
188  /// Return the largest signed value contained in the ConstantRange.
189  APInt getSignedMax() const;
190 
191  /// Return the smallest signed value contained in the ConstantRange.
192  APInt getSignedMin() const;
193 
194  /// Return true if this range is equal to another range.
195  bool operator==(const ConstantRange &CR) const {
196  return Lower == CR.Lower && Upper == CR.Upper;
197  }
198  bool operator!=(const ConstantRange &CR) const {
199  return !operator==(CR);
200  }
201 
202  /// Subtract the specified constant from the endpoints of this constant range.
203  ConstantRange subtract(const APInt &CI) const;
204 
205  /// Subtract the specified range from this range (aka relative complement of
206  /// the sets).
207  ConstantRange difference(const ConstantRange &CR) const;
208 
209  /// Return the range that results from the intersection of
210  /// this range with another range. The resultant range is guaranteed to
211  /// include all elements contained in both input ranges, and to have the
212  /// smallest possible set size that does so. Because there may be two
213  /// intersections with the same set size, A.intersectWith(B) might not
214  /// be equal to B.intersectWith(A).
215  ConstantRange intersectWith(const ConstantRange &CR) const;
216 
217  /// Return the range that results from the union of this range
218  /// with another range. The resultant range is guaranteed to include the
219  /// elements of both sets, but may contain more. For example, [3, 9) union
220  /// [12,15) is [3, 15), which includes 9, 10, and 11, which were not included
221  /// in either set before.
222  ConstantRange unionWith(const ConstantRange &CR) const;
223 
224  /// Return a new range representing the possible values resulting
225  /// from an application of the specified cast operator to this range. \p
226  /// BitWidth is the target bitwidth of the cast. For casts which don't
227  /// change bitwidth, it must be the same as the source bitwidth. For casts
228  /// which do change bitwidth, the bitwidth must be consistent with the
229  /// requested cast and source bitwidth.
230  ConstantRange castOp(Instruction::CastOps CastOp,
231  uint32_t BitWidth) const;
232 
233  /// Return a new range in the specified integer type, which must
234  /// be strictly larger than the current type. The returned range will
235  /// correspond to the possible range of values if the source range had been
236  /// zero extended to BitWidth.
237  ConstantRange zeroExtend(uint32_t BitWidth) const;
238 
239  /// Return a new range in the specified integer type, which must
240  /// be strictly larger than the current type. The returned range will
241  /// correspond to the possible range of values if the source range had been
242  /// sign extended to BitWidth.
243  ConstantRange signExtend(uint32_t BitWidth) const;
244 
245  /// Return a new range in the specified integer type, which must be
246  /// strictly smaller than the current type. The returned range will
247  /// correspond to the possible range of values if the source range had been
248  /// truncated to the specified type.
249  ConstantRange truncate(uint32_t BitWidth) const;
250 
251  /// Make this range have the bit width given by \p BitWidth. The
252  /// value is zero extended, truncated, or left alone to make it that width.
253  ConstantRange zextOrTrunc(uint32_t BitWidth) const;
254 
255  /// Make this range have the bit width given by \p BitWidth. The
256  /// value is sign extended, truncated, or left alone to make it that width.
257  ConstantRange sextOrTrunc(uint32_t BitWidth) const;
258 
259  /// Return a new range representing the possible values resulting
260  /// from an application of the specified binary operator to an left hand side
261  /// of this range and a right hand side of \p Other.
262  ConstantRange binaryOp(Instruction::BinaryOps BinOp,
263  const ConstantRange &Other) const;
264 
265  /// Return a new range representing the possible values resulting
266  /// from an addition of a value in this range and a value in \p Other.
267  ConstantRange add(const ConstantRange &Other) const;
268 
269  /// Return a new range representing the possible values resulting from a
270  /// known NSW addition of a value in this range and \p Other constant.
271  ConstantRange addWithNoSignedWrap(const APInt &Other) const;
272 
273  /// Return a new range representing the possible values resulting
274  /// from a subtraction of a value in this range and a value in \p Other.
275  ConstantRange sub(const ConstantRange &Other) const;
276 
277  /// Return a new range representing the possible values resulting
278  /// from a multiplication of a value in this range and a value in \p Other,
279  /// treating both this and \p Other as unsigned ranges.
280  ConstantRange multiply(const ConstantRange &Other) const;
281 
282  /// Return a new range representing the possible values resulting
283  /// from a signed maximum of a value in this range and a value in \p Other.
284  ConstantRange smax(const ConstantRange &Other) const;
285 
286  /// Return a new range representing the possible values resulting
287  /// from an unsigned maximum of a value in this range and a value in \p Other.
288  ConstantRange umax(const ConstantRange &Other) const;
289 
290  /// Return a new range representing the possible values resulting
291  /// from a signed minimum of a value in this range and a value in \p Other.
292  ConstantRange smin(const ConstantRange &Other) const;
293 
294  /// Return a new range representing the possible values resulting
295  /// from an unsigned minimum of a value in this range and a value in \p Other.
296  ConstantRange umin(const ConstantRange &Other) const;
297 
298  /// Return a new range representing the possible values resulting
299  /// from an unsigned division of a value in this range and a value in
300  /// \p Other.
301  ConstantRange udiv(const ConstantRange &Other) const;
302 
303  /// Return a new range representing the possible values resulting
304  /// from a binary-and of a value in this range by a value in \p Other.
305  ConstantRange binaryAnd(const ConstantRange &Other) const;
306 
307  /// Return a new range representing the possible values resulting
308  /// from a binary-or of a value in this range by a value in \p Other.
309  ConstantRange binaryOr(const ConstantRange &Other) const;
310 
311  /// Return a new range representing the possible values resulting
312  /// from a left shift of a value in this range by a value in \p Other.
313  /// TODO: This isn't fully implemented yet.
314  ConstantRange shl(const ConstantRange &Other) const;
315 
316  /// Return a new range representing the possible values resulting from a
317  /// logical right shift of a value in this range and a value in \p Other.
318  ConstantRange lshr(const ConstantRange &Other) const;
319 
320  /// Return a new range representing the possible values resulting from a
321  /// arithmetic right shift of a value in this range and a value in \p Other.
322  ConstantRange ashr(const ConstantRange &Other) const;
323 
324  /// Return a new range that is the logical not of the current set.
325  ConstantRange inverse() const;
326 
327  /// Print out the bounds to a stream.
328  void print(raw_ostream &OS) const;
329 
330  /// Allow printing from a debugger easily.
331  void dump() const;
332 };
333 
335  CR.print(OS);
336  return OS;
337 }
338 
339 /// Parse out a conservative ConstantRange from !range metadata.
340 ///
341 /// E.g. if RangeMD is !{i32 0, i32 10, i32 15, i32 20} then return [0, 20).
343 
344 } // end namespace llvm
345 
346 #endif // LLVM_IR_CONSTANTRANGE_H
This class represents lattice values for constants.
Definition: AllocatorList.h:24
const APInt & getUpper() const
Return the upper value for this range.
const APInt * getSingleElement() const
If this set contains a single element, return it, otherwise return null.
Metadata node.
Definition: Metadata.h:864
void print(raw_ostream &OS) const
Print out the bounds to a stream.
unsigned getBitWidth() const
Return the number of bits in the APInt.
Definition: APInt.h:1509
return AArch64::GPR64RegClass contains(Reg)
#define LLVM_NODISCARD
LLVM_NODISCARD - Warn if a type or return value is discarded.
Definition: Compiler.h:129
uint32_t getBitWidth() const
Get the bit width of this ConstantRange.
ELFYAML::ELF_STO Other
Definition: ELFYAML.cpp:783
This file implements a class to represent arbitrary precision integral constant values and operations...
const APInt & smax(const APInt &A, const APInt &B)
Determine the larger of two APInts considered to be signed.
Definition: APInt.h:2110
const APInt & smin(const APInt &A, const APInt &B)
Determine the smaller of two APInts considered to be signed.
Definition: APInt.h:2105
void dump(const SparseBitVector< ElementSize > &LHS, raw_ostream &out)
ConstantRange getConstantRangeFromMetadata(const MDNode &RangeMD)
Parse out a conservative ConstantRange from !range metadata.
Predicate
This enumeration lists the possible predicates for CmpInst subclasses.
Definition: InstrTypes.h:642
bool operator!=(const ConstantRange &CR) const
static void print(raw_ostream &Out, object::Archive::Kind Kind, T Val)
const APInt & umin(const APInt &A, const APInt &B)
Determine the smaller of two APInts considered to be signed.
Definition: APInt.h:2115
This class represents a range of values.
Definition: ConstantRange.h:47
Class for arbitrary precision integers.
Definition: APInt.h:70
const APInt & umax(const APInt &A, const APInt &B)
Determine the larger of two APInts considered to be unsigned.
Definition: APInt.h:2120
const APInt * getSingleMissingElement() const
If this set contains all but a single element, return it, otherwise return null.
const APInt & getLower() const
Return the lower value for this range.
raw_ostream & operator<<(raw_ostream &OS, const APInt &I)
Definition: APInt.h:2039
bool isSingleElement() const
Return true if this set contains exactly one member.
LLVM Value Representation.
Definition: Value.h:73
bool operator==(const ConstantRange &CR) const
Return true if this range is equal to another range.
This class implements an extremely fast bulk output stream that can only output to a stream...
Definition: raw_ostream.h:46
bool operator==(uint64_t V1, const APInt &V2)
Definition: APInt.h:1967