1321369Sdim//===- CoverageMapping.h - Code coverage mapping support --------*- C++ -*-===//
2303231Sdim//
3353358Sdim// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4353358Sdim// See https://llvm.org/LICENSE.txt for license information.
5353358Sdim// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6303231Sdim//
7303231Sdim//===----------------------------------------------------------------------===//
8303231Sdim//
9303231Sdim// Code coverage mapping data is generated by clang and read by
10303231Sdim// llvm-cov to show code coverage statistics for a file.
11303231Sdim//
12303231Sdim//===----------------------------------------------------------------------===//
13303231Sdim
14321369Sdim#ifndef LLVM_PROFILEDATA_COVERAGE_COVERAGEMAPPING_H
15321369Sdim#define LLVM_PROFILEDATA_COVERAGE_COVERAGEMAPPING_H
16303231Sdim
17303231Sdim#include "llvm/ADT/ArrayRef.h"
18303231Sdim#include "llvm/ADT/DenseMap.h"
19341825Sdim#include "llvm/ADT/DenseSet.h"
20303231Sdim#include "llvm/ADT/Hashing.h"
21321369Sdim#include "llvm/ADT/None.h"
22321369Sdim#include "llvm/ADT/StringRef.h"
23314564Sdim#include "llvm/ADT/StringSet.h"
24303231Sdim#include "llvm/ADT/iterator.h"
25321369Sdim#include "llvm/ADT/iterator_range.h"
26303231Sdim#include "llvm/ProfileData/InstrProf.h"
27321369Sdim#include "llvm/Support/Compiler.h"
28303231Sdim#include "llvm/Support/Debug.h"
29303231Sdim#include "llvm/Support/Endian.h"
30321369Sdim#include "llvm/Support/Error.h"
31303231Sdim#include "llvm/Support/raw_ostream.h"
32321369Sdim#include <cassert>
33321369Sdim#include <cstdint>
34321369Sdim#include <iterator>
35321369Sdim#include <memory>
36321369Sdim#include <string>
37303231Sdim#include <system_error>
38303231Sdim#include <tuple>
39321369Sdim#include <utility>
40321369Sdim#include <vector>
41303231Sdim
42303231Sdimnamespace llvm {
43321369Sdim
44321369Sdimclass IndexedInstrProfReader;
45321369Sdim
46303231Sdimnamespace coverage {
47303231Sdim
48321369Sdimclass CoverageMappingReader;
49321369Sdimstruct CoverageMappingRecord;
50321369Sdim
51303231Sdimenum class coveragemap_error {
52303231Sdim  success = 0,
53303231Sdim  eof,
54303231Sdim  no_data_found,
55303231Sdim  unsupported_version,
56303231Sdim  truncated,
57303231Sdim  malformed
58303231Sdim};
59303231Sdim
60303231Sdimconst std::error_category &coveragemap_category();
61303231Sdim
62303231Sdiminline std::error_code make_error_code(coveragemap_error E) {
63303231Sdim  return std::error_code(static_cast<int>(E), coveragemap_category());
64303231Sdim}
65303231Sdim
66303231Sdimclass CoverageMapError : public ErrorInfo<CoverageMapError> {
67303231Sdimpublic:
68303231Sdim  CoverageMapError(coveragemap_error Err) : Err(Err) {
69303231Sdim    assert(Err != coveragemap_error::success && "Not an error");
70303231Sdim  }
71303231Sdim
72303231Sdim  std::string message() const override;
73303231Sdim
74303231Sdim  void log(raw_ostream &OS) const override { OS << message(); }
75303231Sdim
76303231Sdim  std::error_code convertToErrorCode() const override {
77303231Sdim    return make_error_code(Err);
78303231Sdim  }
79303231Sdim
80303231Sdim  coveragemap_error get() const { return Err; }
81303231Sdim
82303231Sdim  static char ID;
83303231Sdim
84303231Sdimprivate:
85303231Sdim  coveragemap_error Err;
86303231Sdim};
87303231Sdim
88327952Sdim/// A Counter is an abstract value that describes how to compute the
89303231Sdim/// execution count for a region of code using the collected profile count data.
90303231Sdimstruct Counter {
91303231Sdim  enum CounterKind { Zero, CounterValueReference, Expression };
92303231Sdim  static const unsigned EncodingTagBits = 2;
93303231Sdim  static const unsigned EncodingTagMask = 0x3;
94303231Sdim  static const unsigned EncodingCounterTagAndExpansionRegionTagBits =
95303231Sdim      EncodingTagBits + 1;
96303231Sdim
97303231Sdimprivate:
98321369Sdim  CounterKind Kind = Zero;
99321369Sdim  unsigned ID = 0;
100303231Sdim
101303231Sdim  Counter(CounterKind Kind, unsigned ID) : Kind(Kind), ID(ID) {}
102303231Sdim
103303231Sdimpublic:
104321369Sdim  Counter() = default;
105303231Sdim
106303231Sdim  CounterKind getKind() const { return Kind; }
107303231Sdim
108303231Sdim  bool isZero() const { return Kind == Zero; }
109303231Sdim
110303231Sdim  bool isExpression() const { return Kind == Expression; }
111303231Sdim
112303231Sdim  unsigned getCounterID() const { return ID; }
113303231Sdim
114303231Sdim  unsigned getExpressionID() const { return ID; }
115303231Sdim
116303231Sdim  friend bool operator==(const Counter &LHS, const Counter &RHS) {
117303231Sdim    return LHS.Kind == RHS.Kind && LHS.ID == RHS.ID;
118303231Sdim  }
119303231Sdim
120303231Sdim  friend bool operator!=(const Counter &LHS, const Counter &RHS) {
121303231Sdim    return !(LHS == RHS);
122303231Sdim  }
123303231Sdim
124303231Sdim  friend bool operator<(const Counter &LHS, const Counter &RHS) {
125303231Sdim    return std::tie(LHS.Kind, LHS.ID) < std::tie(RHS.Kind, RHS.ID);
126303231Sdim  }
127303231Sdim
128327952Sdim  /// Return the counter that represents the number zero.
129303231Sdim  static Counter getZero() { return Counter(); }
130303231Sdim
131327952Sdim  /// Return the counter that corresponds to a specific profile counter.
132303231Sdim  static Counter getCounter(unsigned CounterId) {
133303231Sdim    return Counter(CounterValueReference, CounterId);
134303231Sdim  }
135303231Sdim
136327952Sdim  /// Return the counter that corresponds to a specific addition counter
137327952Sdim  /// expression.
138303231Sdim  static Counter getExpression(unsigned ExpressionId) {
139303231Sdim    return Counter(Expression, ExpressionId);
140303231Sdim  }
141303231Sdim};
142303231Sdim
143327952Sdim/// A Counter expression is a value that represents an arithmetic operation
144327952Sdim/// with two counters.
145303231Sdimstruct CounterExpression {
146303231Sdim  enum ExprKind { Subtract, Add };
147303231Sdim  ExprKind Kind;
148303231Sdim  Counter LHS, RHS;
149303231Sdim
150303231Sdim  CounterExpression(ExprKind Kind, Counter LHS, Counter RHS)
151303231Sdim      : Kind(Kind), LHS(LHS), RHS(RHS) {}
152303231Sdim};
153303231Sdim
154327952Sdim/// A Counter expression builder is used to construct the counter expressions.
155327952Sdim/// It avoids unnecessary duplication and simplifies algebraic expressions.
156303231Sdimclass CounterExpressionBuilder {
157327952Sdim  /// A list of all the counter expressions
158303231Sdim  std::vector<CounterExpression> Expressions;
159321369Sdim
160327952Sdim  /// A lookup table for the index of a given expression.
161321369Sdim  DenseMap<CounterExpression, unsigned> ExpressionIndices;
162303231Sdim
163327952Sdim  /// Return the counter which corresponds to the given expression.
164303231Sdim  ///
165303231Sdim  /// If the given expression is already stored in the builder, a counter
166303231Sdim  /// that references that expression is returned. Otherwise, the given
167303231Sdim  /// expression is added to the builder's collection of expressions.
168303231Sdim  Counter get(const CounterExpression &E);
169303231Sdim
170321369Sdim  /// Represents a term in a counter expression tree.
171321369Sdim  struct Term {
172321369Sdim    unsigned CounterID;
173321369Sdim    int Factor;
174321369Sdim
175321369Sdim    Term(unsigned CounterID, int Factor)
176321369Sdim        : CounterID(CounterID), Factor(Factor) {}
177321369Sdim  };
178321369Sdim
179327952Sdim  /// Gather the terms of the expression tree for processing.
180303231Sdim  ///
181303231Sdim  /// This collects each addition and subtraction referenced by the counter into
182303231Sdim  /// a sequence that can be sorted and combined to build a simplified counter
183303231Sdim  /// expression.
184321369Sdim  void extractTerms(Counter C, int Sign, SmallVectorImpl<Term> &Terms);
185303231Sdim
186327952Sdim  /// Simplifies the given expression tree
187303231Sdim  /// by getting rid of algebraically redundant operations.
188303231Sdim  Counter simplify(Counter ExpressionTree);
189303231Sdim
190303231Sdimpublic:
191303231Sdim  ArrayRef<CounterExpression> getExpressions() const { return Expressions; }
192303231Sdim
193327952Sdim  /// Return a counter that represents the expression that adds LHS and RHS.
194303231Sdim  Counter add(Counter LHS, Counter RHS);
195303231Sdim
196327952Sdim  /// Return a counter that represents the expression that subtracts RHS from
197327952Sdim  /// LHS.
198303231Sdim  Counter subtract(Counter LHS, Counter RHS);
199303231Sdim};
200303231Sdim
201327952Sdimusing LineColPair = std::pair<unsigned, unsigned>;
202327952Sdim
203327952Sdim/// A Counter mapping region associates a source range with a specific counter.
204303231Sdimstruct CounterMappingRegion {
205303231Sdim  enum RegionKind {
206327952Sdim    /// A CodeRegion associates some code with a counter
207303231Sdim    CodeRegion,
208303231Sdim
209341825Sdim    /// An ExpansionRegion represents a file expansion region that associates
210327952Sdim    /// a source range with the expansion of a virtual source file, such as
211327952Sdim    /// for a macro instantiation or #include file.
212303231Sdim    ExpansionRegion,
213303231Sdim
214327952Sdim    /// A SkippedRegion represents a source range with code that was skipped
215327952Sdim    /// by a preprocessor or similar means.
216327952Sdim    SkippedRegion,
217327952Sdim
218327952Sdim    /// A GapRegion is like a CodeRegion, but its count is only set as the
219327952Sdim    /// line execution count when its the only region in the line.
220327952Sdim    GapRegion
221303231Sdim  };
222303231Sdim
223303231Sdim  Counter Count;
224303231Sdim  unsigned FileID, ExpandedFileID;
225303231Sdim  unsigned LineStart, ColumnStart, LineEnd, ColumnEnd;
226303231Sdim  RegionKind Kind;
227303231Sdim
228303231Sdim  CounterMappingRegion(Counter Count, unsigned FileID, unsigned ExpandedFileID,
229303231Sdim                       unsigned LineStart, unsigned ColumnStart,
230303231Sdim                       unsigned LineEnd, unsigned ColumnEnd, RegionKind Kind)
231303231Sdim      : Count(Count), FileID(FileID), ExpandedFileID(ExpandedFileID),
232303231Sdim        LineStart(LineStart), ColumnStart(ColumnStart), LineEnd(LineEnd),
233303231Sdim        ColumnEnd(ColumnEnd), Kind(Kind) {}
234303231Sdim
235303231Sdim  static CounterMappingRegion
236303231Sdim  makeRegion(Counter Count, unsigned FileID, unsigned LineStart,
237303231Sdim             unsigned ColumnStart, unsigned LineEnd, unsigned ColumnEnd) {
238303231Sdim    return CounterMappingRegion(Count, FileID, 0, LineStart, ColumnStart,
239303231Sdim                                LineEnd, ColumnEnd, CodeRegion);
240303231Sdim  }
241303231Sdim
242303231Sdim  static CounterMappingRegion
243303231Sdim  makeExpansion(unsigned FileID, unsigned ExpandedFileID, unsigned LineStart,
244303231Sdim                unsigned ColumnStart, unsigned LineEnd, unsigned ColumnEnd) {
245303231Sdim    return CounterMappingRegion(Counter(), FileID, ExpandedFileID, LineStart,
246303231Sdim                                ColumnStart, LineEnd, ColumnEnd,
247303231Sdim                                ExpansionRegion);
248303231Sdim  }
249303231Sdim
250303231Sdim  static CounterMappingRegion
251303231Sdim  makeSkipped(unsigned FileID, unsigned LineStart, unsigned ColumnStart,
252303231Sdim              unsigned LineEnd, unsigned ColumnEnd) {
253303231Sdim    return CounterMappingRegion(Counter(), FileID, 0, LineStart, ColumnStart,
254303231Sdim                                LineEnd, ColumnEnd, SkippedRegion);
255303231Sdim  }
256303231Sdim
257327952Sdim  static CounterMappingRegion
258327952Sdim  makeGapRegion(Counter Count, unsigned FileID, unsigned LineStart,
259327952Sdim                unsigned ColumnStart, unsigned LineEnd, unsigned ColumnEnd) {
260327952Sdim    return CounterMappingRegion(Count, FileID, 0, LineStart, ColumnStart,
261327952Sdim                                LineEnd, (1U << 31) | ColumnEnd, GapRegion);
262303231Sdim  }
263303231Sdim
264327952Sdim  inline LineColPair startLoc() const {
265327952Sdim    return LineColPair(LineStart, ColumnStart);
266303231Sdim  }
267327952Sdim
268327952Sdim  inline LineColPair endLoc() const { return LineColPair(LineEnd, ColumnEnd); }
269303231Sdim};
270303231Sdim
271327952Sdim/// Associates a source range with an execution count.
272303231Sdimstruct CountedRegion : public CounterMappingRegion {
273303231Sdim  uint64_t ExecutionCount;
274303231Sdim
275303231Sdim  CountedRegion(const CounterMappingRegion &R, uint64_t ExecutionCount)
276303231Sdim      : CounterMappingRegion(R), ExecutionCount(ExecutionCount) {}
277303231Sdim};
278303231Sdim
279327952Sdim/// A Counter mapping context is used to connect the counters, expressions
280327952Sdim/// and the obtained counter values.
281303231Sdimclass CounterMappingContext {
282303231Sdim  ArrayRef<CounterExpression> Expressions;
283303231Sdim  ArrayRef<uint64_t> CounterValues;
284303231Sdim
285303231Sdimpublic:
286303231Sdim  CounterMappingContext(ArrayRef<CounterExpression> Expressions,
287303231Sdim                        ArrayRef<uint64_t> CounterValues = None)
288303231Sdim      : Expressions(Expressions), CounterValues(CounterValues) {}
289303231Sdim
290303231Sdim  void setCounts(ArrayRef<uint64_t> Counts) { CounterValues = Counts; }
291303231Sdim
292321369Sdim  void dump(const Counter &C, raw_ostream &OS) const;
293303231Sdim  void dump(const Counter &C) const { dump(C, dbgs()); }
294303231Sdim
295327952Sdim  /// Return the number of times that a region of code associated with this
296327952Sdim  /// counter was executed.
297303231Sdim  Expected<int64_t> evaluate(const Counter &C) const;
298303231Sdim};
299303231Sdim
300327952Sdim/// Code coverage information for a single function.
301303231Sdimstruct FunctionRecord {
302327952Sdim  /// Raw function name.
303303231Sdim  std::string Name;
304360784Sdim  /// Mapping from FileID (i.e. vector index) to filename. Used to support
305360784Sdim  /// macro expansions within a function in which the macro and function are
306360784Sdim  /// defined in separate files.
307360784Sdim  ///
308360784Sdim  /// TODO: Uniquing filenames across all function records may be a performance
309360784Sdim  /// optimization.
310303231Sdim  std::vector<std::string> Filenames;
311327952Sdim  /// Regions in the function along with their counts.
312303231Sdim  std::vector<CountedRegion> CountedRegions;
313327952Sdim  /// The number of times this function was executed.
314360784Sdim  uint64_t ExecutionCount = 0;
315303231Sdim
316303231Sdim  FunctionRecord(StringRef Name, ArrayRef<StringRef> Filenames)
317303231Sdim      : Name(Name), Filenames(Filenames.begin(), Filenames.end()) {}
318303231Sdim
319314564Sdim  FunctionRecord(FunctionRecord &&FR) = default;
320314564Sdim  FunctionRecord &operator=(FunctionRecord &&) = default;
321314564Sdim
322303231Sdim  void pushRegion(CounterMappingRegion Region, uint64_t Count) {
323303231Sdim    if (CountedRegions.empty())
324303231Sdim      ExecutionCount = Count;
325303231Sdim    CountedRegions.emplace_back(Region, Count);
326303231Sdim  }
327303231Sdim};
328303231Sdim
329327952Sdim/// Iterator over Functions, optionally filtered to a single file.
330303231Sdimclass FunctionRecordIterator
331303231Sdim    : public iterator_facade_base<FunctionRecordIterator,
332303231Sdim                                  std::forward_iterator_tag, FunctionRecord> {
333303231Sdim  ArrayRef<FunctionRecord> Records;
334303231Sdim  ArrayRef<FunctionRecord>::iterator Current;
335303231Sdim  StringRef Filename;
336303231Sdim
337327952Sdim  /// Skip records whose primary file is not \c Filename.
338303231Sdim  void skipOtherFiles();
339303231Sdim
340303231Sdimpublic:
341303231Sdim  FunctionRecordIterator(ArrayRef<FunctionRecord> Records_,
342303231Sdim                         StringRef Filename = "")
343303231Sdim      : Records(Records_), Current(Records.begin()), Filename(Filename) {
344303231Sdim    skipOtherFiles();
345303231Sdim  }
346303231Sdim
347303231Sdim  FunctionRecordIterator() : Current(Records.begin()) {}
348303231Sdim
349303231Sdim  bool operator==(const FunctionRecordIterator &RHS) const {
350303231Sdim    return Current == RHS.Current && Filename == RHS.Filename;
351303231Sdim  }
352303231Sdim
353303231Sdim  const FunctionRecord &operator*() const { return *Current; }
354303231Sdim
355303231Sdim  FunctionRecordIterator &operator++() {
356303231Sdim    assert(Current != Records.end() && "incremented past end");
357303231Sdim    ++Current;
358303231Sdim    skipOtherFiles();
359303231Sdim    return *this;
360303231Sdim  }
361303231Sdim};
362303231Sdim
363327952Sdim/// Coverage information for a macro expansion or #included file.
364303231Sdim///
365303231Sdim/// When covered code has pieces that can be expanded for more detail, such as a
366303231Sdim/// preprocessor macro use and its definition, these are represented as
367303231Sdim/// expansions whose coverage can be looked up independently.
368303231Sdimstruct ExpansionRecord {
369327952Sdim  /// The abstract file this expansion covers.
370303231Sdim  unsigned FileID;
371327952Sdim  /// The region that expands to this record.
372303231Sdim  const CountedRegion &Region;
373327952Sdim  /// Coverage for the expansion.
374303231Sdim  const FunctionRecord &Function;
375303231Sdim
376303231Sdim  ExpansionRecord(const CountedRegion &Region,
377303231Sdim                  const FunctionRecord &Function)
378303231Sdim      : FileID(Region.ExpandedFileID), Region(Region), Function(Function) {}
379303231Sdim};
380303231Sdim
381327952Sdim/// The execution count information starting at a point in a file.
382303231Sdim///
383303231Sdim/// A sequence of CoverageSegments gives execution counts for a file in format
384303231Sdim/// that's simple to iterate through for processing.
385303231Sdimstruct CoverageSegment {
386327952Sdim  /// The line where this segment begins.
387303231Sdim  unsigned Line;
388327952Sdim  /// The column where this segment begins.
389303231Sdim  unsigned Col;
390327952Sdim  /// The execution count, or zero if no count was recorded.
391303231Sdim  uint64_t Count;
392327952Sdim  /// When false, the segment was uninstrumented or skipped.
393303231Sdim  bool HasCount;
394327952Sdim  /// Whether this enters a new region or returns to a previous count.
395303231Sdim  bool IsRegionEntry;
396327952Sdim  /// Whether this enters a gap region.
397327952Sdim  bool IsGapRegion;
398303231Sdim
399303231Sdim  CoverageSegment(unsigned Line, unsigned Col, bool IsRegionEntry)
400303231Sdim      : Line(Line), Col(Col), Count(0), HasCount(false),
401327952Sdim        IsRegionEntry(IsRegionEntry), IsGapRegion(false) {}
402303231Sdim
403303231Sdim  CoverageSegment(unsigned Line, unsigned Col, uint64_t Count,
404327952Sdim                  bool IsRegionEntry, bool IsGapRegion = false)
405303231Sdim      : Line(Line), Col(Col), Count(Count), HasCount(true),
406327952Sdim        IsRegionEntry(IsRegionEntry), IsGapRegion(IsGapRegion) {}
407303231Sdim
408303231Sdim  friend bool operator==(const CoverageSegment &L, const CoverageSegment &R) {
409327952Sdim    return std::tie(L.Line, L.Col, L.Count, L.HasCount, L.IsRegionEntry,
410327952Sdim                    L.IsGapRegion) == std::tie(R.Line, R.Col, R.Count,
411327952Sdim                                               R.HasCount, R.IsRegionEntry,
412327952Sdim                                               R.IsGapRegion);
413303231Sdim  }
414303231Sdim};
415303231Sdim
416327952Sdim/// An instantiation group contains a \c FunctionRecord list, such that each
417327952Sdim/// record corresponds to a distinct instantiation of the same function.
418303231Sdim///
419327952Sdim/// Note that it's possible for a function to have more than one instantiation
420327952Sdim/// (consider C++ template specializations or static inline functions).
421327952Sdimclass InstantiationGroup {
422327952Sdim  friend class CoverageMapping;
423327952Sdim
424327952Sdim  unsigned Line;
425327952Sdim  unsigned Col;
426327952Sdim  std::vector<const FunctionRecord *> Instantiations;
427327952Sdim
428327952Sdim  InstantiationGroup(unsigned Line, unsigned Col,
429327952Sdim                     std::vector<const FunctionRecord *> Instantiations)
430327952Sdim      : Line(Line), Col(Col), Instantiations(std::move(Instantiations)) {}
431327952Sdim
432327952Sdimpublic:
433327952Sdim  InstantiationGroup(const InstantiationGroup &) = delete;
434327952Sdim  InstantiationGroup(InstantiationGroup &&) = default;
435327952Sdim
436327952Sdim  /// Get the number of instantiations in this group.
437327952Sdim  size_t size() const { return Instantiations.size(); }
438327952Sdim
439327952Sdim  /// Get the line where the common function was defined.
440327952Sdim  unsigned getLine() const { return Line; }
441327952Sdim
442327952Sdim  /// Get the column where the common function was defined.
443327952Sdim  unsigned getColumn() const { return Col; }
444327952Sdim
445327952Sdim  /// Check if the instantiations in this group have a common mangled name.
446327952Sdim  bool hasName() const {
447327952Sdim    for (unsigned I = 1, E = Instantiations.size(); I < E; ++I)
448327952Sdim      if (Instantiations[I]->Name != Instantiations[0]->Name)
449327952Sdim        return false;
450327952Sdim    return true;
451327952Sdim  }
452327952Sdim
453327952Sdim  /// Get the common mangled name for instantiations in this group.
454327952Sdim  StringRef getName() const {
455327952Sdim    assert(hasName() && "Instantiations don't have a shared name");
456327952Sdim    return Instantiations[0]->Name;
457327952Sdim  }
458327952Sdim
459327952Sdim  /// Get the total execution count of all instantiations in this group.
460327952Sdim  uint64_t getTotalExecutionCount() const {
461327952Sdim    uint64_t Count = 0;
462327952Sdim    for (const FunctionRecord *F : Instantiations)
463327952Sdim      Count += F->ExecutionCount;
464327952Sdim    return Count;
465327952Sdim  }
466327952Sdim
467327952Sdim  /// Get the instantiations in this group.
468327952Sdim  ArrayRef<const FunctionRecord *> getInstantiations() const {
469327952Sdim    return Instantiations;
470327952Sdim  }
471327952Sdim};
472327952Sdim
473327952Sdim/// Coverage information to be processed or displayed.
474327952Sdim///
475303231Sdim/// This represents the coverage of an entire file, expansion, or function. It
476303231Sdim/// provides a sequence of CoverageSegments to iterate through, as well as the
477303231Sdim/// list of expansions that can be further processed.
478303231Sdimclass CoverageData {
479321369Sdim  friend class CoverageMapping;
480321369Sdim
481303231Sdim  std::string Filename;
482303231Sdim  std::vector<CoverageSegment> Segments;
483303231Sdim  std::vector<ExpansionRecord> Expansions;
484303231Sdim
485303231Sdimpublic:
486321369Sdim  CoverageData() = default;
487303231Sdim
488303231Sdim  CoverageData(StringRef Filename) : Filename(Filename) {}
489303231Sdim
490327952Sdim  /// Get the name of the file this data covers.
491303231Sdim  StringRef getFilename() const { return Filename; }
492303231Sdim
493327952Sdim  /// Get an iterator over the coverage segments for this object. The segments
494327952Sdim  /// are guaranteed to be uniqued and sorted by location.
495314564Sdim  std::vector<CoverageSegment>::const_iterator begin() const {
496314564Sdim    return Segments.begin();
497314564Sdim  }
498321369Sdim
499314564Sdim  std::vector<CoverageSegment>::const_iterator end() const {
500314564Sdim    return Segments.end();
501314564Sdim  }
502321369Sdim
503314564Sdim  bool empty() const { return Segments.empty(); }
504303231Sdim
505327952Sdim  /// Expansions that can be further processed.
506314564Sdim  ArrayRef<ExpansionRecord> getExpansions() const { return Expansions; }
507303231Sdim};
508303231Sdim
509327952Sdim/// The mapping of profile information to coverage data.
510303231Sdim///
511303231Sdim/// This is the main interface to get coverage information, using a profile to
512303231Sdim/// fill out execution counts.
513303231Sdimclass CoverageMapping {
514341825Sdim  DenseMap<size_t, DenseSet<size_t>> RecordProvenance;
515303231Sdim  std::vector<FunctionRecord> Functions;
516360784Sdim  DenseMap<size_t, SmallVector<unsigned, 0>> FilenameHash2RecordIndices;
517327952Sdim  std::vector<std::pair<std::string, uint64_t>> FuncHashMismatches;
518303231Sdim
519321369Sdim  CoverageMapping() = default;
520303231Sdim
521327952Sdim  /// Add a function record corresponding to \p Record.
522314564Sdim  Error loadFunctionRecord(const CoverageMappingRecord &Record,
523314564Sdim                           IndexedInstrProfReader &ProfileReader);
524314564Sdim
525360784Sdim  /// Look up the indices for function records which are at least partially
526360784Sdim  /// defined in the specified file. This is guaranteed to return a superset of
527360784Sdim  /// such records: extra records not in the file may be included if there is
528360784Sdim  /// a hash collision on the filename. Clients must be robust to collisions.
529360784Sdim  ArrayRef<unsigned>
530360784Sdim  getImpreciseRecordIndicesForFilename(StringRef Filename) const;
531360784Sdim
532303231Sdimpublic:
533321369Sdim  CoverageMapping(const CoverageMapping &) = delete;
534321369Sdim  CoverageMapping &operator=(const CoverageMapping &) = delete;
535321369Sdim
536327952Sdim  /// Load the coverage mapping using the given readers.
537303231Sdim  static Expected<std::unique_ptr<CoverageMapping>>
538314564Sdim  load(ArrayRef<std::unique_ptr<CoverageMappingReader>> CoverageReaders,
539314564Sdim       IndexedInstrProfReader &ProfileReader);
540314564Sdim
541327952Sdim  /// Load the coverage mapping from the given object files and profile. If
542327952Sdim  /// \p Arches is non-empty, it must specify an architecture for each object.
543360784Sdim  /// Ignores non-instrumented object files unless all are not instrumented.
544303231Sdim  static Expected<std::unique_ptr<CoverageMapping>>
545314564Sdim  load(ArrayRef<StringRef> ObjectFilenames, StringRef ProfileFilename,
546327952Sdim       ArrayRef<StringRef> Arches = None);
547303231Sdim
548327952Sdim  /// The number of functions that couldn't have their profiles mapped.
549303231Sdim  ///
550303231Sdim  /// This is a count of functions whose profile is out of date or otherwise
551303231Sdim  /// can't be associated with any coverage information.
552344779Sdim  unsigned getMismatchedCount() const { return FuncHashMismatches.size(); }
553303231Sdim
554327952Sdim  /// A hash mismatch occurs when a profile record for a symbol does not have
555327952Sdim  /// the same hash as a coverage mapping record for the same symbol. This
556327952Sdim  /// returns a list of hash mismatches, where each mismatch is a pair of the
557327952Sdim  /// symbol name and its coverage mapping hash.
558327952Sdim  ArrayRef<std::pair<std::string, uint64_t>> getHashMismatches() const {
559327952Sdim    return FuncHashMismatches;
560327952Sdim  }
561327952Sdim
562327952Sdim  /// Returns a lexicographically sorted, unique list of files that are
563314564Sdim  /// covered.
564303231Sdim  std::vector<StringRef> getUniqueSourceFiles() const;
565303231Sdim
566327952Sdim  /// Get the coverage for a particular file.
567303231Sdim  ///
568303231Sdim  /// The given filename must be the name as recorded in the coverage
569303231Sdim  /// information. That is, only names returned from getUniqueSourceFiles will
570303231Sdim  /// yield a result.
571303231Sdim  CoverageData getCoverageForFile(StringRef Filename) const;
572303231Sdim
573327952Sdim  /// Get the coverage for a particular function.
574327952Sdim  CoverageData getCoverageForFunction(const FunctionRecord &Function) const;
575327952Sdim
576327952Sdim  /// Get the coverage for an expansion within a coverage set.
577327952Sdim  CoverageData getCoverageForExpansion(const ExpansionRecord &Expansion) const;
578327952Sdim
579327952Sdim  /// Gets all of the functions covered by this profile.
580303231Sdim  iterator_range<FunctionRecordIterator> getCoveredFunctions() const {
581303231Sdim    return make_range(FunctionRecordIterator(Functions),
582303231Sdim                      FunctionRecordIterator());
583303231Sdim  }
584303231Sdim
585327952Sdim  /// Gets all of the functions in a particular file.
586303231Sdim  iterator_range<FunctionRecordIterator>
587303231Sdim  getCoveredFunctions(StringRef Filename) const {
588303231Sdim    return make_range(FunctionRecordIterator(Functions, Filename),
589303231Sdim                      FunctionRecordIterator());
590303231Sdim  }
591303231Sdim
592327952Sdim  /// Get the list of function instantiation groups in a particular file.
593303231Sdim  ///
594327952Sdim  /// Every instantiation group in a program is attributed to exactly one file:
595327952Sdim  /// the file in which the definition for the common function begins.
596327952Sdim  std::vector<InstantiationGroup>
597327952Sdim  getInstantiationGroups(StringRef Filename) const;
598327952Sdim};
599303231Sdim
600327952Sdim/// Coverage statistics for a single line.
601327952Sdimclass LineCoverageStats {
602327952Sdim  uint64_t ExecutionCount;
603327952Sdim  bool HasMultipleRegions;
604327952Sdim  bool Mapped;
605327952Sdim  unsigned Line;
606327952Sdim  ArrayRef<const CoverageSegment *> LineSegments;
607327952Sdim  const CoverageSegment *WrappedSegment;
608303231Sdim
609327952Sdim  friend class LineCoverageIterator;
610327952Sdim  LineCoverageStats() = default;
611327952Sdim
612327952Sdimpublic:
613327952Sdim  LineCoverageStats(ArrayRef<const CoverageSegment *> LineSegments,
614327952Sdim                    const CoverageSegment *WrappedSegment, unsigned Line);
615327952Sdim
616327952Sdim  uint64_t getExecutionCount() const { return ExecutionCount; }
617327952Sdim
618327952Sdim  bool hasMultipleRegions() const { return HasMultipleRegions; }
619327952Sdim
620327952Sdim  bool isMapped() const { return Mapped; }
621327952Sdim
622327952Sdim  unsigned getLine() const { return Line; }
623327952Sdim
624327952Sdim  ArrayRef<const CoverageSegment *> getLineSegments() const {
625327952Sdim    return LineSegments;
626327952Sdim  }
627327952Sdim
628327952Sdim  const CoverageSegment *getWrappedSegment() const { return WrappedSegment; }
629303231Sdim};
630303231Sdim
631327952Sdim/// An iterator over the \c LineCoverageStats objects for lines described by
632327952Sdim/// a \c CoverageData instance.
633327952Sdimclass LineCoverageIterator
634327952Sdim    : public iterator_facade_base<
635327952Sdim          LineCoverageIterator, std::forward_iterator_tag, LineCoverageStats> {
636327952Sdimpublic:
637327952Sdim  LineCoverageIterator(const CoverageData &CD)
638327952Sdim      : LineCoverageIterator(CD, CD.begin()->Line) {}
639327952Sdim
640327952Sdim  LineCoverageIterator(const CoverageData &CD, unsigned Line)
641327952Sdim      : CD(CD), WrappedSegment(nullptr), Next(CD.begin()), Ended(false),
642327952Sdim        Line(Line), Segments(), Stats() {
643327952Sdim    this->operator++();
644327952Sdim  }
645327952Sdim
646327952Sdim  bool operator==(const LineCoverageIterator &R) const {
647327952Sdim    return &CD == &R.CD && Next == R.Next && Ended == R.Ended;
648327952Sdim  }
649327952Sdim
650327952Sdim  const LineCoverageStats &operator*() const { return Stats; }
651327952Sdim
652327952Sdim  LineCoverageStats &operator*() { return Stats; }
653327952Sdim
654327952Sdim  LineCoverageIterator &operator++();
655327952Sdim
656327952Sdim  LineCoverageIterator getEnd() const {
657327952Sdim    auto EndIt = *this;
658327952Sdim    EndIt.Next = CD.end();
659327952Sdim    EndIt.Ended = true;
660327952Sdim    return EndIt;
661327952Sdim  }
662327952Sdim
663327952Sdimprivate:
664327952Sdim  const CoverageData &CD;
665327952Sdim  const CoverageSegment *WrappedSegment;
666327952Sdim  std::vector<CoverageSegment>::const_iterator Next;
667327952Sdim  bool Ended;
668327952Sdim  unsigned Line;
669327952Sdim  SmallVector<const CoverageSegment *, 4> Segments;
670327952Sdim  LineCoverageStats Stats;
671327952Sdim};
672327952Sdim
673327952Sdim/// Get a \c LineCoverageIterator range for the lines described by \p CD.
674327952Sdimstatic inline iterator_range<LineCoverageIterator>
675327952SdimgetLineCoverageStats(const coverage::CoverageData &CD) {
676327952Sdim  auto Begin = LineCoverageIterator(CD);
677327952Sdim  auto End = Begin.getEnd();
678327952Sdim  return make_range(Begin, End);
679327952Sdim}
680327952Sdim
681303231Sdim// Profile coverage map has the following layout:
682303231Sdim// [CoverageMapFileHeader]
683303231Sdim// [ArrayStart]
684303231Sdim//  [CovMapFunctionRecord]
685303231Sdim//  [CovMapFunctionRecord]
686303231Sdim//  ...
687303231Sdim// [ArrayEnd]
688303231Sdim// [Encoded Region Mapping Data]
689303231SdimLLVM_PACKED_START
690303231Sdimtemplate <class IntPtrT> struct CovMapFunctionRecordV1 {
691303231Sdim#define COVMAP_V1
692303231Sdim#define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) Type Name;
693303231Sdim#include "llvm/ProfileData/InstrProfData.inc"
694303231Sdim#undef COVMAP_V1
695303231Sdim
696303231Sdim  // Return the structural hash associated with the function.
697303231Sdim  template <support::endianness Endian> uint64_t getFuncHash() const {
698303231Sdim    return support::endian::byte_swap<uint64_t, Endian>(FuncHash);
699303231Sdim  }
700321369Sdim
701303231Sdim  // Return the coverage map data size for the funciton.
702303231Sdim  template <support::endianness Endian> uint32_t getDataSize() const {
703303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(DataSize);
704303231Sdim  }
705321369Sdim
706303231Sdim  // Return function lookup key. The value is consider opaque.
707303231Sdim  template <support::endianness Endian> IntPtrT getFuncNameRef() const {
708303231Sdim    return support::endian::byte_swap<IntPtrT, Endian>(NamePtr);
709303231Sdim  }
710321369Sdim
711303231Sdim  // Return the PGO name of the function */
712303231Sdim  template <support::endianness Endian>
713303231Sdim  Error getFuncName(InstrProfSymtab &ProfileNames, StringRef &FuncName) const {
714303231Sdim    IntPtrT NameRef = getFuncNameRef<Endian>();
715303231Sdim    uint32_t NameS = support::endian::byte_swap<uint32_t, Endian>(NameSize);
716303231Sdim    FuncName = ProfileNames.getFuncName(NameRef, NameS);
717303231Sdim    if (NameS && FuncName.empty())
718303231Sdim      return make_error<CoverageMapError>(coveragemap_error::malformed);
719303231Sdim    return Error::success();
720303231Sdim  }
721303231Sdim};
722303231Sdim
723303231Sdimstruct CovMapFunctionRecord {
724303231Sdim#define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) Type Name;
725303231Sdim#include "llvm/ProfileData/InstrProfData.inc"
726303231Sdim
727303231Sdim  // Return the structural hash associated with the function.
728303231Sdim  template <support::endianness Endian> uint64_t getFuncHash() const {
729303231Sdim    return support::endian::byte_swap<uint64_t, Endian>(FuncHash);
730303231Sdim  }
731321369Sdim
732303231Sdim  // Return the coverage map data size for the funciton.
733303231Sdim  template <support::endianness Endian> uint32_t getDataSize() const {
734303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(DataSize);
735303231Sdim  }
736321369Sdim
737303231Sdim  // Return function lookup key. The value is consider opaque.
738303231Sdim  template <support::endianness Endian> uint64_t getFuncNameRef() const {
739303231Sdim    return support::endian::byte_swap<uint64_t, Endian>(NameRef);
740303231Sdim  }
741321369Sdim
742303231Sdim  // Return the PGO name of the function */
743303231Sdim  template <support::endianness Endian>
744303231Sdim  Error getFuncName(InstrProfSymtab &ProfileNames, StringRef &FuncName) const {
745303231Sdim    uint64_t NameRef = getFuncNameRef<Endian>();
746303231Sdim    FuncName = ProfileNames.getFuncName(NameRef);
747303231Sdim    return Error::success();
748303231Sdim  }
749303231Sdim};
750303231Sdim
751303231Sdim// Per module coverage mapping data header, i.e. CoverageMapFileHeader
752303231Sdim// documented above.
753303231Sdimstruct CovMapHeader {
754303231Sdim#define COVMAP_HEADER(Type, LLVMType, Name, Init) Type Name;
755303231Sdim#include "llvm/ProfileData/InstrProfData.inc"
756303231Sdim  template <support::endianness Endian> uint32_t getNRecords() const {
757303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(NRecords);
758303231Sdim  }
759321369Sdim
760303231Sdim  template <support::endianness Endian> uint32_t getFilenamesSize() const {
761303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(FilenamesSize);
762303231Sdim  }
763321369Sdim
764303231Sdim  template <support::endianness Endian> uint32_t getCoverageSize() const {
765303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(CoverageSize);
766303231Sdim  }
767321369Sdim
768303231Sdim  template <support::endianness Endian> uint32_t getVersion() const {
769303231Sdim    return support::endian::byte_swap<uint32_t, Endian>(Version);
770303231Sdim  }
771303231Sdim};
772303231Sdim
773303231SdimLLVM_PACKED_END
774303231Sdim
775303231Sdimenum CovMapVersion {
776303231Sdim  Version1 = 0,
777303231Sdim  // Function's name reference from CovMapFuncRecord is changed from raw
778303231Sdim  // name string pointer to MD5 to support name section compression. Name
779303231Sdim  // section is also compressed.
780303231Sdim  Version2 = 1,
781327952Sdim  // A new interpretation of the columnEnd field is added in order to mark
782327952Sdim  // regions as gap areas.
783327952Sdim  Version3 = 2,
784327952Sdim  // The current version is Version3
785303231Sdim  CurrentVersion = INSTR_PROF_COVMAP_VERSION
786303231Sdim};
787303231Sdim
788303231Sdimtemplate <int CovMapVersion, class IntPtrT> struct CovMapTraits {
789321369Sdim  using CovMapFuncRecordType = CovMapFunctionRecord;
790321369Sdim  using NameRefType = uint64_t;
791303231Sdim};
792303231Sdim
793303231Sdimtemplate <class IntPtrT> struct CovMapTraits<CovMapVersion::Version1, IntPtrT> {
794321369Sdim  using CovMapFuncRecordType = CovMapFunctionRecordV1<IntPtrT>;
795321369Sdim  using NameRefType = IntPtrT;
796303231Sdim};
797303231Sdim
798303231Sdim} // end namespace coverage
799303231Sdim
800327952Sdim/// Provide DenseMapInfo for CounterExpression
801303231Sdimtemplate<> struct DenseMapInfo<coverage::CounterExpression> {
802303231Sdim  static inline coverage::CounterExpression getEmptyKey() {
803303231Sdim    using namespace coverage;
804321369Sdim
805303231Sdim    return CounterExpression(CounterExpression::ExprKind::Subtract,
806303231Sdim                             Counter::getCounter(~0U),
807303231Sdim                             Counter::getCounter(~0U));
808303231Sdim  }
809303231Sdim
810303231Sdim  static inline coverage::CounterExpression getTombstoneKey() {
811303231Sdim    using namespace coverage;
812321369Sdim
813303231Sdim    return CounterExpression(CounterExpression::ExprKind::Add,
814303231Sdim                             Counter::getCounter(~0U),
815303231Sdim                             Counter::getCounter(~0U));
816303231Sdim  }
817303231Sdim
818303231Sdim  static unsigned getHashValue(const coverage::CounterExpression &V) {
819303231Sdim    return static_cast<unsigned>(
820303231Sdim        hash_combine(V.Kind, V.LHS.getKind(), V.LHS.getCounterID(),
821303231Sdim                     V.RHS.getKind(), V.RHS.getCounterID()));
822303231Sdim  }
823303231Sdim
824303231Sdim  static bool isEqual(const coverage::CounterExpression &LHS,
825303231Sdim                      const coverage::CounterExpression &RHS) {
826303231Sdim    return LHS.Kind == RHS.Kind && LHS.LHS == RHS.LHS && LHS.RHS == RHS.RHS;
827303231Sdim  }
828303231Sdim};
829303231Sdim
830303231Sdim} // end namespace llvm
831303231Sdim
832321369Sdim#endif // LLVM_PROFILEDATA_COVERAGE_COVERAGEMAPPING_H
833