1 //===--- CoverageMappingGen.cpp - Coverage mapping generation ---*- 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 // Instrumentation-based code coverage mapping generator
11 //
12 //===----------------------------------------------------------------------===//
13 
14 #include "CoverageMappingGen.h"
15 #include "CodeGenFunction.h"
16 #include "clang/AST/StmtVisitor.h"
17 #include "clang/Lex/Lexer.h"
18 #include "llvm/ADT/SmallSet.h"
19 #include "llvm/ADT/StringExtras.h"
20 #include "llvm/ADT/Optional.h"
21 #include "llvm/ProfileData/Coverage/CoverageMapping.h"
22 #include "llvm/ProfileData/Coverage/CoverageMappingReader.h"
23 #include "llvm/ProfileData/Coverage/CoverageMappingWriter.h"
24 #include "llvm/ProfileData/InstrProfReader.h"
25 #include "llvm/Support/FileSystem.h"
26 #include "llvm/Support/Path.h"
27 
28 using namespace clang;
29 using namespace CodeGen;
30 using namespace llvm::coverage;
31 
32 void CoverageSourceInfo::SourceRangeSkipped(SourceRange Range) {
33   SkippedRanges.push_back(Range);
34 }
35 
36 namespace {
37 
38 /// \brief A region of source code that can be mapped to a counter.
39 class SourceMappingRegion {
40   Counter Count;
41 
42   /// \brief The region's starting location.
43   Optional<SourceLocation> LocStart;
44 
45   /// \brief The region's ending location.
46   Optional<SourceLocation> LocEnd;
47 
48 public:
49   SourceMappingRegion(Counter Count, Optional<SourceLocation> LocStart,
50                       Optional<SourceLocation> LocEnd)
51       : Count(Count), LocStart(LocStart), LocEnd(LocEnd) {}
52 
53   const Counter &getCounter() const { return Count; }
54 
55   void setCounter(Counter C) { Count = C; }
56 
57   bool hasStartLoc() const { return LocStart.hasValue(); }
58 
59   void setStartLoc(SourceLocation Loc) { LocStart = Loc; }
60 
61   SourceLocation getStartLoc() const {
62     assert(LocStart && "Region has no start location");
63     return *LocStart;
64   }
65 
66   bool hasEndLoc() const { return LocEnd.hasValue(); }
67 
68   void setEndLoc(SourceLocation Loc) { LocEnd = Loc; }
69 
70   SourceLocation getEndLoc() const {
71     assert(LocEnd && "Region has no end location");
72     return *LocEnd;
73   }
74 };
75 
76 /// \brief Provides the common functionality for the different
77 /// coverage mapping region builders.
78 class CoverageMappingBuilder {
79 public:
80   CoverageMappingModuleGen &CVM;
81   SourceManager &SM;
82   const LangOptions &LangOpts;
83 
84 private:
85   /// \brief Map of clang's FileIDs to IDs used for coverage mapping.
86   llvm::SmallDenseMap<FileID, std::pair<unsigned, SourceLocation>, 8>
87       FileIDMapping;
88 
89 public:
90   /// \brief The coverage mapping regions for this function
91   llvm::SmallVector<CounterMappingRegion, 32> MappingRegions;
92   /// \brief The source mapping regions for this function.
93   std::vector<SourceMappingRegion> SourceRegions;
94 
95   CoverageMappingBuilder(CoverageMappingModuleGen &CVM, SourceManager &SM,
96                          const LangOptions &LangOpts)
97       : CVM(CVM), SM(SM), LangOpts(LangOpts) {}
98 
99   /// \brief Return the precise end location for the given token.
100   SourceLocation getPreciseTokenLocEnd(SourceLocation Loc) {
101     // We avoid getLocForEndOfToken here, because it doesn't do what we want for
102     // macro locations, which we just treat as expanded files.
103     unsigned TokLen =
104         Lexer::MeasureTokenLength(SM.getSpellingLoc(Loc), SM, LangOpts);
105     return Loc.getLocWithOffset(TokLen);
106   }
107 
108   /// \brief Return the start location of an included file or expanded macro.
109   SourceLocation getStartOfFileOrMacro(SourceLocation Loc) {
110     if (Loc.isMacroID())
111       return Loc.getLocWithOffset(-SM.getFileOffset(Loc));
112     return SM.getLocForStartOfFile(SM.getFileID(Loc));
113   }
114 
115   /// \brief Return the end location of an included file or expanded macro.
116   SourceLocation getEndOfFileOrMacro(SourceLocation Loc) {
117     if (Loc.isMacroID())
118       return Loc.getLocWithOffset(SM.getFileIDSize(SM.getFileID(Loc)) -
119                                   SM.getFileOffset(Loc));
120     return SM.getLocForEndOfFile(SM.getFileID(Loc));
121   }
122 
123   /// \brief Find out where the current file is included or macro is expanded.
124   SourceLocation getIncludeOrExpansionLoc(SourceLocation Loc) {
125     return Loc.isMacroID() ? SM.getImmediateExpansionRange(Loc).first
126                            : SM.getIncludeLoc(SM.getFileID(Loc));
127   }
128 
129   /// \brief Return true if \c Loc is a location in a built-in macro.
130   bool isInBuiltin(SourceLocation Loc) {
131     return strcmp(SM.getBufferName(SM.getSpellingLoc(Loc)), "<built-in>") == 0;
132   }
133 
134   /// \brief Check whether \c Loc is included or expanded from \c Parent.
135   bool isNestedIn(SourceLocation Loc, FileID Parent) {
136     do {
137       Loc = getIncludeOrExpansionLoc(Loc);
138       if (Loc.isInvalid())
139         return false;
140     } while (!SM.isInFileID(Loc, Parent));
141     return true;
142   }
143 
144   /// \brief Get the start of \c S ignoring macro arguments and builtin macros.
145   SourceLocation getStart(const Stmt *S) {
146     SourceLocation Loc = S->getLocStart();
147     while (SM.isMacroArgExpansion(Loc) || isInBuiltin(Loc))
148       Loc = SM.getImmediateExpansionRange(Loc).first;
149     return Loc;
150   }
151 
152   /// \brief Get the end of \c S ignoring macro arguments and builtin macros.
153   SourceLocation getEnd(const Stmt *S) {
154     SourceLocation Loc = S->getLocEnd();
155     while (SM.isMacroArgExpansion(Loc) || isInBuiltin(Loc))
156       Loc = SM.getImmediateExpansionRange(Loc).first;
157     return getPreciseTokenLocEnd(Loc);
158   }
159 
160   /// \brief Find the set of files we have regions for and assign IDs
161   ///
162   /// Fills \c Mapping with the virtual file mapping needed to write out
163   /// coverage and collects the necessary file information to emit source and
164   /// expansion regions.
165   void gatherFileIDs(SmallVectorImpl<unsigned> &Mapping) {
166     FileIDMapping.clear();
167 
168     llvm::SmallSet<FileID, 8> Visited;
169     SmallVector<std::pair<SourceLocation, unsigned>, 8> FileLocs;
170     for (const auto &Region : SourceRegions) {
171       SourceLocation Loc = Region.getStartLoc();
172       FileID File = SM.getFileID(Loc);
173       if (!Visited.insert(File).second)
174         continue;
175 
176       // Do not map FileID's associated with system headers.
177       if (SM.isInSystemHeader(SM.getSpellingLoc(Loc)))
178         continue;
179 
180       unsigned Depth = 0;
181       for (SourceLocation Parent = getIncludeOrExpansionLoc(Loc);
182            Parent.isValid(); Parent = getIncludeOrExpansionLoc(Parent))
183         ++Depth;
184       FileLocs.push_back(std::make_pair(Loc, Depth));
185     }
186     std::stable_sort(FileLocs.begin(), FileLocs.end(), llvm::less_second());
187 
188     for (const auto &FL : FileLocs) {
189       SourceLocation Loc = FL.first;
190       FileID SpellingFile = SM.getDecomposedSpellingLoc(Loc).first;
191       auto Entry = SM.getFileEntryForID(SpellingFile);
192       if (!Entry)
193         continue;
194 
195       FileIDMapping[SM.getFileID(Loc)] = std::make_pair(Mapping.size(), Loc);
196       Mapping.push_back(CVM.getFileID(Entry));
197     }
198   }
199 
200   /// \brief Get the coverage mapping file ID for \c Loc.
201   ///
202   /// If such file id doesn't exist, return None.
203   Optional<unsigned> getCoverageFileID(SourceLocation Loc) {
204     auto Mapping = FileIDMapping.find(SM.getFileID(Loc));
205     if (Mapping != FileIDMapping.end())
206       return Mapping->second.first;
207     return None;
208   }
209 
210   /// \brief Gather all the regions that were skipped by the preprocessor
211   /// using the constructs like #if.
212   void gatherSkippedRegions() {
213     /// An array of the minimum lineStarts and the maximum lineEnds
214     /// for mapping regions from the appropriate source files.
215     llvm::SmallVector<std::pair<unsigned, unsigned>, 8> FileLineRanges;
216     FileLineRanges.resize(
217         FileIDMapping.size(),
218         std::make_pair(std::numeric_limits<unsigned>::max(), 0));
219     for (const auto &R : MappingRegions) {
220       FileLineRanges[R.FileID].first =
221           std::min(FileLineRanges[R.FileID].first, R.LineStart);
222       FileLineRanges[R.FileID].second =
223           std::max(FileLineRanges[R.FileID].second, R.LineEnd);
224     }
225 
226     auto SkippedRanges = CVM.getSourceInfo().getSkippedRanges();
227     for (const auto &I : SkippedRanges) {
228       auto LocStart = I.getBegin();
229       auto LocEnd = I.getEnd();
230       assert(SM.isWrittenInSameFile(LocStart, LocEnd) &&
231              "region spans multiple files");
232 
233       auto CovFileID = getCoverageFileID(LocStart);
234       if (!CovFileID)
235         continue;
236       unsigned LineStart = SM.getSpellingLineNumber(LocStart);
237       unsigned ColumnStart = SM.getSpellingColumnNumber(LocStart);
238       unsigned LineEnd = SM.getSpellingLineNumber(LocEnd);
239       unsigned ColumnEnd = SM.getSpellingColumnNumber(LocEnd);
240       auto Region = CounterMappingRegion::makeSkipped(
241           *CovFileID, LineStart, ColumnStart, LineEnd, ColumnEnd);
242       // Make sure that we only collect the regions that are inside
243       // the souce code of this function.
244       if (Region.LineStart >= FileLineRanges[*CovFileID].first &&
245           Region.LineEnd <= FileLineRanges[*CovFileID].second)
246         MappingRegions.push_back(Region);
247     }
248   }
249 
250   /// \brief Generate the coverage counter mapping regions from collected
251   /// source regions.
252   void emitSourceRegions() {
253     for (const auto &Region : SourceRegions) {
254       assert(Region.hasEndLoc() && "incomplete region");
255 
256       SourceLocation LocStart = Region.getStartLoc();
257       assert(SM.getFileID(LocStart).isValid() && "region in invalid file");
258 
259       // Ignore regions from system headers.
260       if (SM.isInSystemHeader(SM.getSpellingLoc(LocStart)))
261         continue;
262 
263       auto CovFileID = getCoverageFileID(LocStart);
264       // Ignore regions that don't have a file, such as builtin macros.
265       if (!CovFileID)
266         continue;
267 
268       SourceLocation LocEnd = Region.getEndLoc();
269       assert(SM.isWrittenInSameFile(LocStart, LocEnd) &&
270              "region spans multiple files");
271 
272       // Find the spilling locations for the mapping region.
273       unsigned LineStart = SM.getSpellingLineNumber(LocStart);
274       unsigned ColumnStart = SM.getSpellingColumnNumber(LocStart);
275       unsigned LineEnd = SM.getSpellingLineNumber(LocEnd);
276       unsigned ColumnEnd = SM.getSpellingColumnNumber(LocEnd);
277 
278       assert(LineStart <= LineEnd && "region start and end out of order");
279       MappingRegions.push_back(CounterMappingRegion::makeRegion(
280           Region.getCounter(), *CovFileID, LineStart, ColumnStart, LineEnd,
281           ColumnEnd));
282     }
283   }
284 
285   /// \brief Generate expansion regions for each virtual file we've seen.
286   void emitExpansionRegions() {
287     for (const auto &FM : FileIDMapping) {
288       SourceLocation ExpandedLoc = FM.second.second;
289       SourceLocation ParentLoc = getIncludeOrExpansionLoc(ExpandedLoc);
290       if (ParentLoc.isInvalid())
291         continue;
292 
293       auto ParentFileID = getCoverageFileID(ParentLoc);
294       if (!ParentFileID)
295         continue;
296       auto ExpandedFileID = getCoverageFileID(ExpandedLoc);
297       assert(ExpandedFileID && "expansion in uncovered file");
298 
299       SourceLocation LocEnd = getPreciseTokenLocEnd(ParentLoc);
300       assert(SM.isWrittenInSameFile(ParentLoc, LocEnd) &&
301              "region spans multiple files");
302 
303       unsigned LineStart = SM.getSpellingLineNumber(ParentLoc);
304       unsigned ColumnStart = SM.getSpellingColumnNumber(ParentLoc);
305       unsigned LineEnd = SM.getSpellingLineNumber(LocEnd);
306       unsigned ColumnEnd = SM.getSpellingColumnNumber(LocEnd);
307 
308       MappingRegions.push_back(CounterMappingRegion::makeExpansion(
309           *ParentFileID, *ExpandedFileID, LineStart, ColumnStart, LineEnd,
310           ColumnEnd));
311     }
312   }
313 };
314 
315 /// \brief Creates unreachable coverage regions for the functions that
316 /// are not emitted.
317 struct EmptyCoverageMappingBuilder : public CoverageMappingBuilder {
318   EmptyCoverageMappingBuilder(CoverageMappingModuleGen &CVM, SourceManager &SM,
319                               const LangOptions &LangOpts)
320       : CoverageMappingBuilder(CVM, SM, LangOpts) {}
321 
322   void VisitDecl(const Decl *D) {
323     if (!D->hasBody())
324       return;
325     auto Body = D->getBody();
326     SourceLocation Start = getStart(Body);
327     SourceLocation End = getEnd(Body);
328     if (!SM.isWrittenInSameFile(Start, End)) {
329       // Walk up to find the common ancestor.
330       // Correct the locations accordingly.
331       FileID StartFileID = SM.getFileID(Start);
332       FileID EndFileID = SM.getFileID(End);
333       while (StartFileID != EndFileID && !isNestedIn(End, StartFileID)) {
334         Start = getIncludeOrExpansionLoc(Start);
335         assert(Start.isValid() &&
336                "Declaration start location not nested within a known region");
337         StartFileID = SM.getFileID(Start);
338       }
339       while (StartFileID != EndFileID) {
340         End = getPreciseTokenLocEnd(getIncludeOrExpansionLoc(End));
341         assert(End.isValid() &&
342                "Declaration end location not nested within a known region");
343         EndFileID = SM.getFileID(End);
344       }
345     }
346     SourceRegions.emplace_back(Counter(), Start, End);
347   }
348 
349   /// \brief Write the mapping data to the output stream
350   void write(llvm::raw_ostream &OS) {
351     SmallVector<unsigned, 16> FileIDMapping;
352     gatherFileIDs(FileIDMapping);
353     emitSourceRegions();
354 
355     if (MappingRegions.empty())
356       return;
357 
358     CoverageMappingWriter Writer(FileIDMapping, None, MappingRegions);
359     Writer.write(OS);
360   }
361 };
362 
363 /// \brief A StmtVisitor that creates coverage mapping regions which map
364 /// from the source code locations to the PGO counters.
365 struct CounterCoverageMappingBuilder
366     : public CoverageMappingBuilder,
367       public ConstStmtVisitor<CounterCoverageMappingBuilder> {
368   /// \brief The map of statements to count values.
369   llvm::DenseMap<const Stmt *, unsigned> &CounterMap;
370 
371   /// \brief A stack of currently live regions.
372   std::vector<SourceMappingRegion> RegionStack;
373 
374   CounterExpressionBuilder Builder;
375 
376   /// \brief A location in the most recently visited file or macro.
377   ///
378   /// This is used to adjust the active source regions appropriately when
379   /// expressions cross file or macro boundaries.
380   SourceLocation MostRecentLocation;
381 
382   /// \brief Return a counter for the subtraction of \c RHS from \c LHS
383   Counter subtractCounters(Counter LHS, Counter RHS) {
384     return Builder.subtract(LHS, RHS);
385   }
386 
387   /// \brief Return a counter for the sum of \c LHS and \c RHS.
388   Counter addCounters(Counter LHS, Counter RHS) {
389     return Builder.add(LHS, RHS);
390   }
391 
392   Counter addCounters(Counter C1, Counter C2, Counter C3) {
393     return addCounters(addCounters(C1, C2), C3);
394   }
395 
396   /// \brief Return the region counter for the given statement.
397   ///
398   /// This should only be called on statements that have a dedicated counter.
399   Counter getRegionCounter(const Stmt *S) {
400     return Counter::getCounter(CounterMap[S]);
401   }
402 
403   /// \brief Push a region onto the stack.
404   ///
405   /// Returns the index on the stack where the region was pushed. This can be
406   /// used with popRegions to exit a "scope", ending the region that was pushed.
407   size_t pushRegion(Counter Count, Optional<SourceLocation> StartLoc = None,
408                     Optional<SourceLocation> EndLoc = None) {
409     if (StartLoc)
410       MostRecentLocation = *StartLoc;
411     RegionStack.emplace_back(Count, StartLoc, EndLoc);
412 
413     return RegionStack.size() - 1;
414   }
415 
416   /// \brief Pop regions from the stack into the function's list of regions.
417   ///
418   /// Adds all regions from \c ParentIndex to the top of the stack to the
419   /// function's \c SourceRegions.
420   void popRegions(size_t ParentIndex) {
421     assert(RegionStack.size() >= ParentIndex && "parent not in stack");
422     while (RegionStack.size() > ParentIndex) {
423       SourceMappingRegion &Region = RegionStack.back();
424       if (Region.hasStartLoc()) {
425         SourceLocation StartLoc = Region.getStartLoc();
426         SourceLocation EndLoc = Region.hasEndLoc()
427                                     ? Region.getEndLoc()
428                                     : RegionStack[ParentIndex].getEndLoc();
429         while (!SM.isWrittenInSameFile(StartLoc, EndLoc)) {
430           // The region ends in a nested file or macro expansion. Create a
431           // separate region for each expansion.
432           SourceLocation NestedLoc = getStartOfFileOrMacro(EndLoc);
433           assert(SM.isWrittenInSameFile(NestedLoc, EndLoc));
434 
435           SourceRegions.emplace_back(Region.getCounter(), NestedLoc, EndLoc);
436 
437           EndLoc = getPreciseTokenLocEnd(getIncludeOrExpansionLoc(EndLoc));
438           if (EndLoc.isInvalid())
439             llvm::report_fatal_error("File exit not handled before popRegions");
440         }
441         Region.setEndLoc(EndLoc);
442 
443         MostRecentLocation = EndLoc;
444         // If this region happens to span an entire expansion, we need to make
445         // sure we don't overlap the parent region with it.
446         if (StartLoc == getStartOfFileOrMacro(StartLoc) &&
447             EndLoc == getEndOfFileOrMacro(EndLoc))
448           MostRecentLocation = getIncludeOrExpansionLoc(EndLoc);
449 
450         assert(SM.isWrittenInSameFile(Region.getStartLoc(), EndLoc));
451         SourceRegions.push_back(Region);
452       }
453       RegionStack.pop_back();
454     }
455   }
456 
457   /// \brief Return the currently active region.
458   SourceMappingRegion &getRegion() {
459     assert(!RegionStack.empty() && "statement has no region");
460     return RegionStack.back();
461   }
462 
463   /// \brief Propagate counts through the children of \c S.
464   Counter propagateCounts(Counter TopCount, const Stmt *S) {
465     size_t Index = pushRegion(TopCount, getStart(S), getEnd(S));
466     Visit(S);
467     Counter ExitCount = getRegion().getCounter();
468     popRegions(Index);
469 
470     // The statement may be spanned by an expansion. Make sure we handle a file
471     // exit out of this expansion before moving to the next statement.
472     if (SM.isBeforeInTranslationUnit(getStart(S), S->getLocStart()))
473       MostRecentLocation = getEnd(S);
474 
475     return ExitCount;
476   }
477 
478   /// \brief Check whether a region with bounds \c StartLoc and \c EndLoc
479   /// is already added to \c SourceRegions.
480   bool isRegionAlreadyAdded(SourceLocation StartLoc, SourceLocation EndLoc) {
481     return SourceRegions.rend() !=
482            std::find_if(SourceRegions.rbegin(), SourceRegions.rend(),
483                         [&](const SourceMappingRegion &Region) {
484                           return Region.getStartLoc() == StartLoc &&
485                                  Region.getEndLoc() == EndLoc;
486                         });
487   }
488 
489   /// \brief Adjust the most recently visited location to \c EndLoc.
490   ///
491   /// This should be used after visiting any statements in non-source order.
492   void adjustForOutOfOrderTraversal(SourceLocation EndLoc) {
493     MostRecentLocation = EndLoc;
494     // The code region for a whole macro is created in handleFileExit() when
495     // it detects exiting of the virtual file of that macro. If we visited
496     // statements in non-source order, we might already have such a region
497     // added, for example, if a body of a loop is divided among multiple
498     // macros. Avoid adding duplicate regions in such case.
499     if (getRegion().hasEndLoc() &&
500         MostRecentLocation == getEndOfFileOrMacro(MostRecentLocation) &&
501         isRegionAlreadyAdded(getStartOfFileOrMacro(MostRecentLocation),
502                              MostRecentLocation))
503       MostRecentLocation = getIncludeOrExpansionLoc(MostRecentLocation);
504   }
505 
506   /// \brief Adjust regions and state when \c NewLoc exits a file.
507   ///
508   /// If moving from our most recently tracked location to \c NewLoc exits any
509   /// files, this adjusts our current region stack and creates the file regions
510   /// for the exited file.
511   void handleFileExit(SourceLocation NewLoc) {
512     if (NewLoc.isInvalid() ||
513         SM.isWrittenInSameFile(MostRecentLocation, NewLoc))
514       return;
515 
516     // If NewLoc is not in a file that contains MostRecentLocation, walk up to
517     // find the common ancestor.
518     SourceLocation LCA = NewLoc;
519     FileID ParentFile = SM.getFileID(LCA);
520     while (!isNestedIn(MostRecentLocation, ParentFile)) {
521       LCA = getIncludeOrExpansionLoc(LCA);
522       if (LCA.isInvalid() || SM.isWrittenInSameFile(LCA, MostRecentLocation)) {
523         // Since there isn't a common ancestor, no file was exited. We just need
524         // to adjust our location to the new file.
525         MostRecentLocation = NewLoc;
526         return;
527       }
528       ParentFile = SM.getFileID(LCA);
529     }
530 
531     llvm::SmallSet<SourceLocation, 8> StartLocs;
532     Optional<Counter> ParentCounter;
533     for (SourceMappingRegion &I : llvm::reverse(RegionStack)) {
534       if (!I.hasStartLoc())
535         continue;
536       SourceLocation Loc = I.getStartLoc();
537       if (!isNestedIn(Loc, ParentFile)) {
538         ParentCounter = I.getCounter();
539         break;
540       }
541 
542       while (!SM.isInFileID(Loc, ParentFile)) {
543         // The most nested region for each start location is the one with the
544         // correct count. We avoid creating redundant regions by stopping once
545         // we've seen this region.
546         if (StartLocs.insert(Loc).second)
547           SourceRegions.emplace_back(I.getCounter(), Loc,
548                                      getEndOfFileOrMacro(Loc));
549         Loc = getIncludeOrExpansionLoc(Loc);
550       }
551       I.setStartLoc(getPreciseTokenLocEnd(Loc));
552     }
553 
554     if (ParentCounter) {
555       // If the file is contained completely by another region and doesn't
556       // immediately start its own region, the whole file gets a region
557       // corresponding to the parent.
558       SourceLocation Loc = MostRecentLocation;
559       while (isNestedIn(Loc, ParentFile)) {
560         SourceLocation FileStart = getStartOfFileOrMacro(Loc);
561         if (StartLocs.insert(FileStart).second)
562           SourceRegions.emplace_back(*ParentCounter, FileStart,
563                                      getEndOfFileOrMacro(Loc));
564         Loc = getIncludeOrExpansionLoc(Loc);
565       }
566     }
567 
568     MostRecentLocation = NewLoc;
569   }
570 
571   /// \brief Ensure that \c S is included in the current region.
572   void extendRegion(const Stmt *S) {
573     SourceMappingRegion &Region = getRegion();
574     SourceLocation StartLoc = getStart(S);
575 
576     handleFileExit(StartLoc);
577     if (!Region.hasStartLoc())
578       Region.setStartLoc(StartLoc);
579   }
580 
581   /// \brief Mark \c S as a terminator, starting a zero region.
582   void terminateRegion(const Stmt *S) {
583     extendRegion(S);
584     SourceMappingRegion &Region = getRegion();
585     if (!Region.hasEndLoc())
586       Region.setEndLoc(getEnd(S));
587     pushRegion(Counter::getZero());
588   }
589 
590   /// \brief Keep counts of breaks and continues inside loops.
591   struct BreakContinue {
592     Counter BreakCount;
593     Counter ContinueCount;
594   };
595   SmallVector<BreakContinue, 8> BreakContinueStack;
596 
597   CounterCoverageMappingBuilder(
598       CoverageMappingModuleGen &CVM,
599       llvm::DenseMap<const Stmt *, unsigned> &CounterMap, SourceManager &SM,
600       const LangOptions &LangOpts)
601       : CoverageMappingBuilder(CVM, SM, LangOpts), CounterMap(CounterMap) {}
602 
603   /// \brief Write the mapping data to the output stream
604   void write(llvm::raw_ostream &OS) {
605     llvm::SmallVector<unsigned, 8> VirtualFileMapping;
606     gatherFileIDs(VirtualFileMapping);
607     emitSourceRegions();
608     emitExpansionRegions();
609     gatherSkippedRegions();
610 
611     if (MappingRegions.empty())
612       return;
613 
614     CoverageMappingWriter Writer(VirtualFileMapping, Builder.getExpressions(),
615                                  MappingRegions);
616     Writer.write(OS);
617   }
618 
619   void VisitStmt(const Stmt *S) {
620     if (S->getLocStart().isValid())
621       extendRegion(S);
622     for (const Stmt *Child : S->children())
623       if (Child)
624         this->Visit(Child);
625     handleFileExit(getEnd(S));
626   }
627 
628   void VisitDecl(const Decl *D) {
629     Stmt *Body = D->getBody();
630 
631     // Do not propagate region counts into system headers.
632     if (Body && SM.isInSystemHeader(SM.getSpellingLoc(getStart(Body))))
633       return;
634 
635     propagateCounts(getRegionCounter(Body), Body);
636   }
637 
638   void VisitReturnStmt(const ReturnStmt *S) {
639     extendRegion(S);
640     if (S->getRetValue())
641       Visit(S->getRetValue());
642     terminateRegion(S);
643   }
644 
645   void VisitCXXThrowExpr(const CXXThrowExpr *E) {
646     extendRegion(E);
647     if (E->getSubExpr())
648       Visit(E->getSubExpr());
649     terminateRegion(E);
650   }
651 
652   void VisitGotoStmt(const GotoStmt *S) { terminateRegion(S); }
653 
654   void VisitLabelStmt(const LabelStmt *S) {
655     SourceLocation Start = getStart(S);
656     // We can't extendRegion here or we risk overlapping with our new region.
657     handleFileExit(Start);
658     pushRegion(getRegionCounter(S), Start);
659     Visit(S->getSubStmt());
660   }
661 
662   void VisitBreakStmt(const BreakStmt *S) {
663     assert(!BreakContinueStack.empty() && "break not in a loop or switch!");
664     BreakContinueStack.back().BreakCount = addCounters(
665         BreakContinueStack.back().BreakCount, getRegion().getCounter());
666     terminateRegion(S);
667   }
668 
669   void VisitContinueStmt(const ContinueStmt *S) {
670     assert(!BreakContinueStack.empty() && "continue stmt not in a loop!");
671     BreakContinueStack.back().ContinueCount = addCounters(
672         BreakContinueStack.back().ContinueCount, getRegion().getCounter());
673     terminateRegion(S);
674   }
675 
676   void VisitWhileStmt(const WhileStmt *S) {
677     extendRegion(S);
678 
679     Counter ParentCount = getRegion().getCounter();
680     Counter BodyCount = getRegionCounter(S);
681 
682     // Handle the body first so that we can get the backedge count.
683     BreakContinueStack.push_back(BreakContinue());
684     extendRegion(S->getBody());
685     Counter BackedgeCount = propagateCounts(BodyCount, S->getBody());
686     BreakContinue BC = BreakContinueStack.pop_back_val();
687 
688     // Go back to handle the condition.
689     Counter CondCount =
690         addCounters(ParentCount, BackedgeCount, BC.ContinueCount);
691     propagateCounts(CondCount, S->getCond());
692     adjustForOutOfOrderTraversal(getEnd(S));
693 
694     Counter OutCount =
695         addCounters(BC.BreakCount, subtractCounters(CondCount, BodyCount));
696     if (OutCount != ParentCount)
697       pushRegion(OutCount);
698   }
699 
700   void VisitDoStmt(const DoStmt *S) {
701     extendRegion(S);
702 
703     Counter ParentCount = getRegion().getCounter();
704     Counter BodyCount = getRegionCounter(S);
705 
706     BreakContinueStack.push_back(BreakContinue());
707     extendRegion(S->getBody());
708     Counter BackedgeCount =
709         propagateCounts(addCounters(ParentCount, BodyCount), S->getBody());
710     BreakContinue BC = BreakContinueStack.pop_back_val();
711 
712     Counter CondCount = addCounters(BackedgeCount, BC.ContinueCount);
713     propagateCounts(CondCount, S->getCond());
714 
715     Counter OutCount =
716         addCounters(BC.BreakCount, subtractCounters(CondCount, BodyCount));
717     if (OutCount != ParentCount)
718       pushRegion(OutCount);
719   }
720 
721   void VisitForStmt(const ForStmt *S) {
722     extendRegion(S);
723     if (S->getInit())
724       Visit(S->getInit());
725 
726     Counter ParentCount = getRegion().getCounter();
727     Counter BodyCount = getRegionCounter(S);
728 
729     // Handle the body first so that we can get the backedge count.
730     BreakContinueStack.push_back(BreakContinue());
731     extendRegion(S->getBody());
732     Counter BackedgeCount = propagateCounts(BodyCount, S->getBody());
733     BreakContinue BC = BreakContinueStack.pop_back_val();
734 
735     // The increment is essentially part of the body but it needs to include
736     // the count for all the continue statements.
737     if (const Stmt *Inc = S->getInc())
738       propagateCounts(addCounters(BackedgeCount, BC.ContinueCount), Inc);
739 
740     // Go back to handle the condition.
741     Counter CondCount =
742         addCounters(ParentCount, BackedgeCount, BC.ContinueCount);
743     if (const Expr *Cond = S->getCond()) {
744       propagateCounts(CondCount, Cond);
745       adjustForOutOfOrderTraversal(getEnd(S));
746     }
747 
748     Counter OutCount =
749         addCounters(BC.BreakCount, subtractCounters(CondCount, BodyCount));
750     if (OutCount != ParentCount)
751       pushRegion(OutCount);
752   }
753 
754   void VisitCXXForRangeStmt(const CXXForRangeStmt *S) {
755     extendRegion(S);
756     Visit(S->getLoopVarStmt());
757     Visit(S->getRangeStmt());
758 
759     Counter ParentCount = getRegion().getCounter();
760     Counter BodyCount = getRegionCounter(S);
761 
762     BreakContinueStack.push_back(BreakContinue());
763     extendRegion(S->getBody());
764     Counter BackedgeCount = propagateCounts(BodyCount, S->getBody());
765     BreakContinue BC = BreakContinueStack.pop_back_val();
766 
767     Counter LoopCount =
768         addCounters(ParentCount, BackedgeCount, BC.ContinueCount);
769     Counter OutCount =
770         addCounters(BC.BreakCount, subtractCounters(LoopCount, BodyCount));
771     if (OutCount != ParentCount)
772       pushRegion(OutCount);
773   }
774 
775   void VisitObjCForCollectionStmt(const ObjCForCollectionStmt *S) {
776     extendRegion(S);
777     Visit(S->getElement());
778 
779     Counter ParentCount = getRegion().getCounter();
780     Counter BodyCount = getRegionCounter(S);
781 
782     BreakContinueStack.push_back(BreakContinue());
783     extendRegion(S->getBody());
784     Counter BackedgeCount = propagateCounts(BodyCount, S->getBody());
785     BreakContinue BC = BreakContinueStack.pop_back_val();
786 
787     Counter LoopCount =
788         addCounters(ParentCount, BackedgeCount, BC.ContinueCount);
789     Counter OutCount =
790         addCounters(BC.BreakCount, subtractCounters(LoopCount, BodyCount));
791     if (OutCount != ParentCount)
792       pushRegion(OutCount);
793   }
794 
795   void VisitSwitchStmt(const SwitchStmt *S) {
796     extendRegion(S);
797     Visit(S->getCond());
798 
799     BreakContinueStack.push_back(BreakContinue());
800 
801     const Stmt *Body = S->getBody();
802     extendRegion(Body);
803     if (const auto *CS = dyn_cast<CompoundStmt>(Body)) {
804       if (!CS->body_empty()) {
805         // The body of the switch needs a zero region so that fallthrough counts
806         // behave correctly, but it would be misleading to include the braces of
807         // the compound statement in the zeroed area, so we need to handle this
808         // specially.
809         size_t Index =
810             pushRegion(Counter::getZero(), getStart(CS->body_front()),
811                        getEnd(CS->body_back()));
812         for (const auto *Child : CS->children())
813           Visit(Child);
814         popRegions(Index);
815       }
816     } else
817       propagateCounts(Counter::getZero(), Body);
818     BreakContinue BC = BreakContinueStack.pop_back_val();
819 
820     if (!BreakContinueStack.empty())
821       BreakContinueStack.back().ContinueCount = addCounters(
822           BreakContinueStack.back().ContinueCount, BC.ContinueCount);
823 
824     Counter ExitCount = getRegionCounter(S);
825     SourceLocation ExitLoc = getEnd(S);
826     pushRegion(ExitCount, getStart(S), ExitLoc);
827     handleFileExit(ExitLoc);
828   }
829 
830   void VisitSwitchCase(const SwitchCase *S) {
831     extendRegion(S);
832 
833     SourceMappingRegion &Parent = getRegion();
834 
835     Counter Count = addCounters(Parent.getCounter(), getRegionCounter(S));
836     // Reuse the existing region if it starts at our label. This is typical of
837     // the first case in a switch.
838     if (Parent.hasStartLoc() && Parent.getStartLoc() == getStart(S))
839       Parent.setCounter(Count);
840     else
841       pushRegion(Count, getStart(S));
842 
843     if (const auto *CS = dyn_cast<CaseStmt>(S)) {
844       Visit(CS->getLHS());
845       if (const Expr *RHS = CS->getRHS())
846         Visit(RHS);
847     }
848     Visit(S->getSubStmt());
849   }
850 
851   void VisitIfStmt(const IfStmt *S) {
852     extendRegion(S);
853     // Extend into the condition before we propagate through it below - this is
854     // needed to handle macros that generate the "if" but not the condition.
855     extendRegion(S->getCond());
856 
857     Counter ParentCount = getRegion().getCounter();
858     Counter ThenCount = getRegionCounter(S);
859 
860     // Emitting a counter for the condition makes it easier to interpret the
861     // counter for the body when looking at the coverage.
862     propagateCounts(ParentCount, S->getCond());
863 
864     extendRegion(S->getThen());
865     Counter OutCount = propagateCounts(ThenCount, S->getThen());
866 
867     Counter ElseCount = subtractCounters(ParentCount, ThenCount);
868     if (const Stmt *Else = S->getElse()) {
869       extendRegion(S->getElse());
870       OutCount = addCounters(OutCount, propagateCounts(ElseCount, Else));
871     } else
872       OutCount = addCounters(OutCount, ElseCount);
873 
874     if (OutCount != ParentCount)
875       pushRegion(OutCount);
876   }
877 
878   void VisitCXXTryStmt(const CXXTryStmt *S) {
879     extendRegion(S);
880     // Handle macros that generate the "try" but not the rest.
881     extendRegion(S->getTryBlock());
882 
883     Counter ParentCount = getRegion().getCounter();
884     propagateCounts(ParentCount, S->getTryBlock());
885 
886     for (unsigned I = 0, E = S->getNumHandlers(); I < E; ++I)
887       Visit(S->getHandler(I));
888 
889     Counter ExitCount = getRegionCounter(S);
890     pushRegion(ExitCount);
891   }
892 
893   void VisitCXXCatchStmt(const CXXCatchStmt *S) {
894     propagateCounts(getRegionCounter(S), S->getHandlerBlock());
895   }
896 
897   void VisitAbstractConditionalOperator(const AbstractConditionalOperator *E) {
898     extendRegion(E);
899 
900     Counter ParentCount = getRegion().getCounter();
901     Counter TrueCount = getRegionCounter(E);
902 
903     Visit(E->getCond());
904 
905     if (!isa<BinaryConditionalOperator>(E)) {
906       extendRegion(E->getTrueExpr());
907       propagateCounts(TrueCount, E->getTrueExpr());
908     }
909     extendRegion(E->getFalseExpr());
910     propagateCounts(subtractCounters(ParentCount, TrueCount),
911                     E->getFalseExpr());
912   }
913 
914   void VisitBinLAnd(const BinaryOperator *E) {
915     extendRegion(E);
916     Visit(E->getLHS());
917 
918     extendRegion(E->getRHS());
919     propagateCounts(getRegionCounter(E), E->getRHS());
920   }
921 
922   void VisitBinLOr(const BinaryOperator *E) {
923     extendRegion(E);
924     Visit(E->getLHS());
925 
926     extendRegion(E->getRHS());
927     propagateCounts(getRegionCounter(E), E->getRHS());
928   }
929 
930   void VisitLambdaExpr(const LambdaExpr *LE) {
931     // Lambdas are treated as their own functions for now, so we shouldn't
932     // propagate counts into them.
933   }
934 };
935 
936 bool isMachO(const CodeGenModule &CGM) {
937   return CGM.getTarget().getTriple().isOSBinFormatMachO();
938 }
939 
940 StringRef getCoverageSection(const CodeGenModule &CGM) {
941   return llvm::getInstrProfCoverageSectionName(isMachO(CGM));
942 }
943 
944 std::string normalizeFilename(StringRef Filename) {
945   llvm::SmallString<256> Path(Filename);
946   llvm::sys::fs::make_absolute(Path);
947   llvm::sys::path::remove_dots(Path, /*remove_dot_dots=*/true);
948   return Path.str().str();
949 }
950 
951 } // end anonymous namespace
952 
953 static void dump(llvm::raw_ostream &OS, StringRef FunctionName,
954                  ArrayRef<CounterExpression> Expressions,
955                  ArrayRef<CounterMappingRegion> Regions) {
956   OS << FunctionName << ":\n";
957   CounterMappingContext Ctx(Expressions);
958   for (const auto &R : Regions) {
959     OS.indent(2);
960     switch (R.Kind) {
961     case CounterMappingRegion::CodeRegion:
962       break;
963     case CounterMappingRegion::ExpansionRegion:
964       OS << "Expansion,";
965       break;
966     case CounterMappingRegion::SkippedRegion:
967       OS << "Skipped,";
968       break;
969     }
970 
971     OS << "File " << R.FileID << ", " << R.LineStart << ":" << R.ColumnStart
972        << " -> " << R.LineEnd << ":" << R.ColumnEnd << " = ";
973     Ctx.dump(R.Count, OS);
974     if (R.Kind == CounterMappingRegion::ExpansionRegion)
975       OS << " (Expanded file = " << R.ExpandedFileID << ")";
976     OS << "\n";
977   }
978 }
979 
980 void CoverageMappingModuleGen::addFunctionMappingRecord(
981     llvm::GlobalVariable *NamePtr, StringRef NameValue, uint64_t FuncHash,
982     const std::string &CoverageMapping, bool IsUsed) {
983   llvm::LLVMContext &Ctx = CGM.getLLVMContext();
984   if (!FunctionRecordTy) {
985 #define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) LLVMType,
986     llvm::Type *FunctionRecordTypes[] = {
987       #include "llvm/ProfileData/InstrProfData.inc"
988     };
989     FunctionRecordTy =
990         llvm::StructType::get(Ctx, makeArrayRef(FunctionRecordTypes),
991                               /*isPacked=*/true);
992   }
993 
994   #define COVMAP_FUNC_RECORD(Type, LLVMType, Name, Init) Init,
995   llvm::Constant *FunctionRecordVals[] = {
996       #include "llvm/ProfileData/InstrProfData.inc"
997   };
998   FunctionRecords.push_back(llvm::ConstantStruct::get(
999       FunctionRecordTy, makeArrayRef(FunctionRecordVals)));
1000   if (!IsUsed)
1001     FunctionNames.push_back(
1002         llvm::ConstantExpr::getBitCast(NamePtr, llvm::Type::getInt8PtrTy(Ctx)));
1003   CoverageMappings.push_back(CoverageMapping);
1004 
1005   if (CGM.getCodeGenOpts().DumpCoverageMapping) {
1006     // Dump the coverage mapping data for this function by decoding the
1007     // encoded data. This allows us to dump the mapping regions which were
1008     // also processed by the CoverageMappingWriter which performs
1009     // additional minimization operations such as reducing the number of
1010     // expressions.
1011     std::vector<StringRef> Filenames;
1012     std::vector<CounterExpression> Expressions;
1013     std::vector<CounterMappingRegion> Regions;
1014     llvm::SmallVector<StringRef, 16> FilenameRefs;
1015     FilenameRefs.resize(FileEntries.size());
1016     for (const auto &Entry : FileEntries)
1017       FilenameRefs[Entry.second] = normalizeFilename(Entry.first->getName());
1018     RawCoverageMappingReader Reader(CoverageMapping, FilenameRefs, Filenames,
1019                                     Expressions, Regions);
1020     if (Reader.read())
1021       return;
1022     dump(llvm::outs(), NameValue, Expressions, Regions);
1023   }
1024 }
1025 
1026 void CoverageMappingModuleGen::emit() {
1027   if (FunctionRecords.empty())
1028     return;
1029   llvm::LLVMContext &Ctx = CGM.getLLVMContext();
1030   auto *Int32Ty = llvm::Type::getInt32Ty(Ctx);
1031 
1032   // Create the filenames and merge them with coverage mappings
1033   llvm::SmallVector<std::string, 16> FilenameStrs;
1034   llvm::SmallVector<StringRef, 16> FilenameRefs;
1035   FilenameStrs.resize(FileEntries.size());
1036   FilenameRefs.resize(FileEntries.size());
1037   for (const auto &Entry : FileEntries) {
1038     auto I = Entry.second;
1039     FilenameStrs[I] = normalizeFilename(Entry.first->getName());
1040     FilenameRefs[I] = FilenameStrs[I];
1041   }
1042 
1043   std::string FilenamesAndCoverageMappings;
1044   llvm::raw_string_ostream OS(FilenamesAndCoverageMappings);
1045   CoverageFilenamesSectionWriter(FilenameRefs).write(OS);
1046   std::string RawCoverageMappings =
1047       llvm::join(CoverageMappings.begin(), CoverageMappings.end(), "");
1048   OS << RawCoverageMappings;
1049   size_t CoverageMappingSize = RawCoverageMappings.size();
1050   size_t FilenamesSize = OS.str().size() - CoverageMappingSize;
1051   // Append extra zeroes if necessary to ensure that the size of the filenames
1052   // and coverage mappings is a multiple of 8.
1053   if (size_t Rem = OS.str().size() % 8) {
1054     CoverageMappingSize += 8 - Rem;
1055     for (size_t I = 0, S = 8 - Rem; I < S; ++I)
1056       OS << '\0';
1057   }
1058   auto *FilenamesAndMappingsVal =
1059       llvm::ConstantDataArray::getString(Ctx, OS.str(), false);
1060 
1061   // Create the deferred function records array
1062   auto RecordsTy =
1063       llvm::ArrayType::get(FunctionRecordTy, FunctionRecords.size());
1064   auto RecordsVal = llvm::ConstantArray::get(RecordsTy, FunctionRecords);
1065 
1066   llvm::Type *CovDataHeaderTypes[] = {
1067 #define COVMAP_HEADER(Type, LLVMType, Name, Init) LLVMType,
1068 #include "llvm/ProfileData/InstrProfData.inc"
1069   };
1070   auto CovDataHeaderTy =
1071       llvm::StructType::get(Ctx, makeArrayRef(CovDataHeaderTypes));
1072   llvm::Constant *CovDataHeaderVals[] = {
1073 #define COVMAP_HEADER(Type, LLVMType, Name, Init) Init,
1074 #include "llvm/ProfileData/InstrProfData.inc"
1075   };
1076   auto CovDataHeaderVal = llvm::ConstantStruct::get(
1077       CovDataHeaderTy, makeArrayRef(CovDataHeaderVals));
1078 
1079   // Create the coverage data record
1080   llvm::Type *CovDataTypes[] = {CovDataHeaderTy, RecordsTy,
1081                                 FilenamesAndMappingsVal->getType()};
1082   auto CovDataTy = llvm::StructType::get(Ctx, makeArrayRef(CovDataTypes));
1083   llvm::Constant *TUDataVals[] = {CovDataHeaderVal, RecordsVal,
1084                                   FilenamesAndMappingsVal};
1085   auto CovDataVal =
1086       llvm::ConstantStruct::get(CovDataTy, makeArrayRef(TUDataVals));
1087   auto CovData = new llvm::GlobalVariable(
1088       CGM.getModule(), CovDataTy, true, llvm::GlobalValue::InternalLinkage,
1089       CovDataVal, llvm::getCoverageMappingVarName());
1090 
1091   CovData->setSection(getCoverageSection(CGM));
1092   CovData->setAlignment(8);
1093 
1094   // Make sure the data doesn't get deleted.
1095   CGM.addUsedGlobal(CovData);
1096   // Create the deferred function records array
1097   if (!FunctionNames.empty()) {
1098     auto NamesArrTy = llvm::ArrayType::get(llvm::Type::getInt8PtrTy(Ctx),
1099                                            FunctionNames.size());
1100     auto NamesArrVal = llvm::ConstantArray::get(NamesArrTy, FunctionNames);
1101     // This variable will *NOT* be emitted to the object file. It is used
1102     // to pass the list of names referenced to codegen.
1103     new llvm::GlobalVariable(CGM.getModule(), NamesArrTy, true,
1104                              llvm::GlobalValue::InternalLinkage, NamesArrVal,
1105                              llvm::getCoverageUnusedNamesVarName());
1106   }
1107 }
1108 
1109 unsigned CoverageMappingModuleGen::getFileID(const FileEntry *File) {
1110   auto It = FileEntries.find(File);
1111   if (It != FileEntries.end())
1112     return It->second;
1113   unsigned FileID = FileEntries.size();
1114   FileEntries.insert(std::make_pair(File, FileID));
1115   return FileID;
1116 }
1117 
1118 void CoverageMappingGen::emitCounterMapping(const Decl *D,
1119                                             llvm::raw_ostream &OS) {
1120   assert(CounterMap);
1121   CounterCoverageMappingBuilder Walker(CVM, *CounterMap, SM, LangOpts);
1122   Walker.VisitDecl(D);
1123   Walker.write(OS);
1124 }
1125 
1126 void CoverageMappingGen::emitEmptyMapping(const Decl *D,
1127                                           llvm::raw_ostream &OS) {
1128   EmptyCoverageMappingBuilder Walker(CVM, SM, LangOpts);
1129   Walker.VisitDecl(D);
1130   Walker.write(OS);
1131 }
1132