CacheTokens.cpp revision 36c35ba0aca641e60e5dbee8efbc620c08b9bd61 
1//===--- CacheTokens.cpp - Caching of lexer tokens for PTH support --------===//
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// This provides a possible implementation of PTH support for Clang that is
11// based on caching lexed tokens and identifiers.
12//
13//===----------------------------------------------------------------------===//
14
15#include "clang/Frontend/Utils.h"
16#include "clang/Basic/FileManager.h"
17#include "clang/Basic/SourceManager.h"
18#include "clang/Basic/IdentifierTable.h"
19#include "clang/Basic/Diagnostic.h"
20#include "clang/Basic/OnDiskHashTable.h"
21#include "clang/Lex/Lexer.h"
22#include "clang/Lex/Preprocessor.h"
23#include "llvm/ADT/StringExtras.h"
24#include "llvm/ADT/StringMap.h"
25#include "llvm/Support/MemoryBuffer.h"
26#include "llvm/Support/raw_ostream.h"
27#include "llvm/System/Path.h"
28
29// FIXME: put this somewhere else?
30#ifndef S_ISDIR
31#define S_ISDIR(x) (((x)&_S_IFDIR)!=0)
32#endif
33
34using namespace clang;
35using namespace clang::io;
36
37//===----------------------------------------------------------------------===//
38// PTH-specific stuff.
39//===----------------------------------------------------------------------===//
40
41namespace {
42class PTHEntry {
43  Offset TokenData, PPCondData;
44
45public:
46  PTHEntry() {}
47
48  PTHEntry(Offset td, Offset ppcd)
49    : TokenData(td), PPCondData(ppcd) {}
50
51  Offset getTokenOffset() const { return TokenData; }
52  Offset getPPCondTableOffset() const { return PPCondData; }
53};
54
55
56class PTHEntryKeyVariant {
57  union { const FileEntry* FE; const char* Path; };
58  enum { IsFE = 0x1, IsDE = 0x2, IsNoExist = 0x0 } Kind;
59  struct stat *StatBuf;
60public:
61  PTHEntryKeyVariant(const FileEntry *fe)
62    : FE(fe), Kind(IsFE), StatBuf(0) {}
63
64  PTHEntryKeyVariant(struct stat* statbuf, const char* path)
65    : Path(path), Kind(IsDE), StatBuf(new struct stat(*statbuf)) {}
66
67  explicit PTHEntryKeyVariant(const char* path)
68    : Path(path), Kind(IsNoExist), StatBuf(0) {}
69
70  bool isFile() const { return Kind == IsFE; }
71
72  llvm::StringRef getString() const {
73    return Kind == IsFE ? FE->getName() : Path;
74  }
75
76  unsigned getKind() const { return (unsigned) Kind; }
77
78  void EmitData(llvm::raw_ostream& Out) {
79    switch (Kind) {
80    case IsFE:
81      // Emit stat information.
82      ::Emit32(Out, FE->getInode());
83      ::Emit32(Out, FE->getDevice());
84      ::Emit16(Out, FE->getFileMode());
85      ::Emit64(Out, FE->getModificationTime());
86      ::Emit64(Out, FE->getSize());
87      break;
88    case IsDE:
89      // Emit stat information.
90      ::Emit32(Out, (uint32_t) StatBuf->st_ino);
91      ::Emit32(Out, (uint32_t) StatBuf->st_dev);
92      ::Emit16(Out, (uint16_t) StatBuf->st_mode);
93      ::Emit64(Out, (uint64_t) StatBuf->st_mtime);
94      ::Emit64(Out, (uint64_t) StatBuf->st_size);
95      delete StatBuf;
96      break;
97    default:
98      break;
99    }
100  }
101
102  unsigned getRepresentationLength() const {
103    return Kind == IsNoExist ? 0 : 4 + 4 + 2 + 8 + 8;
104  }
105};
106
107class FileEntryPTHEntryInfo {
108public:
109  typedef PTHEntryKeyVariant key_type;
110  typedef key_type key_type_ref;
111
112  typedef PTHEntry data_type;
113  typedef const PTHEntry& data_type_ref;
114
115  static unsigned ComputeHash(PTHEntryKeyVariant V) {
116    return llvm::HashString(V.getString());
117  }
118
119  static std::pair<unsigned,unsigned>
120  EmitKeyDataLength(llvm::raw_ostream& Out, PTHEntryKeyVariant V,
121                    const PTHEntry& E) {
122
123    unsigned n = V.getString().size() + 1 + 1;
124    ::Emit16(Out, n);
125
126    unsigned m = V.getRepresentationLength() + (V.isFile() ? 4 + 4 : 0);
127    ::Emit8(Out, m);
128
129    return std::make_pair(n, m);
130  }
131
132  static void EmitKey(llvm::raw_ostream& Out, PTHEntryKeyVariant V, unsigned n){
133    // Emit the entry kind.
134    ::Emit8(Out, (unsigned) V.getKind());
135    // Emit the string.
136    Out.write(V.getString().data(), n - 1);
137  }
138
139  static void EmitData(llvm::raw_ostream& Out, PTHEntryKeyVariant V,
140                       const PTHEntry& E, unsigned) {
141
142
143    // For file entries emit the offsets into the PTH file for token data
144    // and the preprocessor blocks table.
145    if (V.isFile()) {
146      ::Emit32(Out, E.getTokenOffset());
147      ::Emit32(Out, E.getPPCondTableOffset());
148    }
149
150    // Emit any other data associated with the key (i.e., stat information).
151    V.EmitData(Out);
152  }
153};
154
155class OffsetOpt {
156  bool valid;
157  Offset off;
158public:
159  OffsetOpt() : valid(false) {}
160  bool hasOffset() const { return valid; }
161  Offset getOffset() const { assert(valid); return off; }
162  void setOffset(Offset o) { off = o; valid = true; }
163};
164} // end anonymous namespace
165
166typedef OnDiskChainedHashTableGenerator<FileEntryPTHEntryInfo> PTHMap;
167typedef llvm::DenseMap<const IdentifierInfo*,uint32_t> IDMap;
168typedef llvm::StringMap<OffsetOpt, llvm::BumpPtrAllocator> CachedStrsTy;
169
170namespace {
171class PTHWriter {
172  IDMap IM;
173  llvm::raw_fd_ostream& Out;
174  Preprocessor& PP;
175  uint32_t idcount;
176  PTHMap PM;
177  CachedStrsTy CachedStrs;
178  Offset CurStrOffset;
179  std::vector<llvm::StringMapEntry<OffsetOpt>*> StrEntries;
180
181  //// Get the persistent id for the given IdentifierInfo*.
182  uint32_t ResolveID(const IdentifierInfo* II);
183
184  /// Emit a token to the PTH file.
185  void EmitToken(const Token& T);
186
187  void Emit8(uint32_t V) { ::Emit8(Out, V); }
188
189  void Emit16(uint32_t V) { ::Emit16(Out, V); }
190
191  void Emit24(uint32_t V) { ::Emit24(Out, V); }
192
193  void Emit32(uint32_t V) { ::Emit32(Out, V); }
194
195  void EmitBuf(const char *Ptr, unsigned NumBytes) {
196    Out.write(Ptr, NumBytes);
197  }
198
199  void EmitString(llvm::StringRef V) {
200    ::Emit16(Out, V.size());
201    EmitBuf(V.data(), V.size());
202  }
203
204  /// EmitIdentifierTable - Emits two tables to the PTH file.  The first is
205  ///  a hashtable mapping from identifier strings to persistent IDs.
206  ///  The second is a straight table mapping from persistent IDs to string data
207  ///  (the keys of the first table).
208  std::pair<Offset, Offset> EmitIdentifierTable();
209
210  /// EmitFileTable - Emit a table mapping from file name strings to PTH
211  /// token data.
212  Offset EmitFileTable() { return PM.Emit(Out); }
213
214  PTHEntry LexTokens(Lexer& L);
215  Offset EmitCachedSpellings();
216
217public:
218  PTHWriter(llvm::raw_fd_ostream& out, Preprocessor& pp)
219    : Out(out), PP(pp), idcount(0), CurStrOffset(0) {}
220
221  PTHMap &getPM() { return PM; }
222  void GeneratePTH(const std::string &MainFile);
223};
224} // end anonymous namespace
225
226uint32_t PTHWriter::ResolveID(const IdentifierInfo* II) {
227  // Null IdentifierInfo's map to the persistent ID 0.
228  if (!II)
229    return 0;
230
231  IDMap::iterator I = IM.find(II);
232  if (I != IM.end())
233    return I->second; // We've already added 1.
234
235  IM[II] = ++idcount; // Pre-increment since '0' is reserved for NULL.
236  return idcount;
237}
238
239void PTHWriter::EmitToken(const Token& T) {
240  // Emit the token kind, flags, and length.
241  Emit32(((uint32_t) T.getKind()) | ((((uint32_t) T.getFlags())) << 8)|
242         (((uint32_t) T.getLength()) << 16));
243
244  if (!T.isLiteral()) {
245    Emit32(ResolveID(T.getIdentifierInfo()));
246  } else {
247    // We cache *un-cleaned* spellings. This gives us 100% fidelity with the
248    // source code.
249    const char* s = T.getLiteralData();
250    unsigned len = T.getLength();
251
252    // Get the string entry.
253    llvm::StringMapEntry<OffsetOpt> *E = &CachedStrs.GetOrCreateValue(s, s+len);
254
255    // If this is a new string entry, bump the PTH offset.
256    if (!E->getValue().hasOffset()) {
257      E->getValue().setOffset(CurStrOffset);
258      StrEntries.push_back(E);
259      CurStrOffset += len + 1;
260    }
261
262    // Emit the relative offset into the PTH file for the spelling string.
263    Emit32(E->getValue().getOffset());
264  }
265
266  // Emit the offset into the original source file of this token so that we
267  // can reconstruct its SourceLocation.
268  Emit32(PP.getSourceManager().getFileOffset(T.getLocation()));
269}
270
271PTHEntry PTHWriter::LexTokens(Lexer& L) {
272  // Pad 0's so that we emit tokens to a 4-byte alignment.
273  // This speed up reading them back in.
274  Pad(Out, 4);
275  Offset off = (Offset) Out.tell();
276
277  // Keep track of matching '#if' ... '#endif'.
278  typedef std::vector<std::pair<Offset, unsigned> > PPCondTable;
279  PPCondTable PPCond;
280  std::vector<unsigned> PPStartCond;
281  bool ParsingPreprocessorDirective = false;
282  Token Tok;
283
284  do {
285    L.LexFromRawLexer(Tok);
286  NextToken:
287
288    if ((Tok.isAtStartOfLine() || Tok.is(tok::eof)) &&
289        ParsingPreprocessorDirective) {
290      // Insert an eom token into the token cache.  It has the same
291      // position as the next token that is not on the same line as the
292      // preprocessor directive.  Observe that we continue processing
293      // 'Tok' when we exit this branch.
294      Token Tmp = Tok;
295      Tmp.setKind(tok::eom);
296      Tmp.clearFlag(Token::StartOfLine);
297      Tmp.setIdentifierInfo(0);
298      EmitToken(Tmp);
299      ParsingPreprocessorDirective = false;
300    }
301
302    if (Tok.is(tok::identifier)) {
303      PP.LookUpIdentifierInfo(Tok);
304      EmitToken(Tok);
305      continue;
306    }
307
308    if (Tok.is(tok::hash) && Tok.isAtStartOfLine()) {
309      // Special processing for #include.  Store the '#' token and lex
310      // the next token.
311      assert(!ParsingPreprocessorDirective);
312      Offset HashOff = (Offset) Out.tell();
313      EmitToken(Tok);
314
315      // Get the next token.
316      L.LexFromRawLexer(Tok);
317
318      // If we see the start of line, then we had a null directive "#".
319      if (Tok.isAtStartOfLine())
320        goto NextToken;
321
322      // Did we see 'include'/'import'/'include_next'?
323      if (Tok.isNot(tok::identifier)) {
324        EmitToken(Tok);
325        continue;
326      }
327
328      IdentifierInfo* II = PP.LookUpIdentifierInfo(Tok);
329      tok::PPKeywordKind K = II->getPPKeywordID();
330
331      ParsingPreprocessorDirective = true;
332
333      switch (K) {
334      case tok::pp_not_keyword:
335        // Invalid directives "#foo" can occur in #if 0 blocks etc, just pass
336        // them through.
337      default:
338        break;
339
340      case tok::pp_include:
341      case tok::pp_import:
342      case tok::pp_include_next: {
343        // Save the 'include' token.
344        EmitToken(Tok);
345        // Lex the next token as an include string.
346        L.setParsingPreprocessorDirective(true);
347        L.LexIncludeFilename(Tok);
348        L.setParsingPreprocessorDirective(false);
349        assert(!Tok.isAtStartOfLine());
350        if (Tok.is(tok::identifier))
351          PP.LookUpIdentifierInfo(Tok);
352
353        break;
354      }
355      case tok::pp_if:
356      case tok::pp_ifdef:
357      case tok::pp_ifndef: {
358        // Add an entry for '#if' and friends.  We initially set the target
359        // index to 0.  This will get backpatched when we hit #endif.
360        PPStartCond.push_back(PPCond.size());
361        PPCond.push_back(std::make_pair(HashOff, 0U));
362        break;
363      }
364      case tok::pp_endif: {
365        // Add an entry for '#endif'.  We set the target table index to itself.
366        // This will later be set to zero when emitting to the PTH file.  We
367        // use 0 for uninitialized indices because that is easier to debug.
368        unsigned index = PPCond.size();
369        // Backpatch the opening '#if' entry.
370        assert(!PPStartCond.empty());
371        assert(PPCond.size() > PPStartCond.back());
372        assert(PPCond[PPStartCond.back()].second == 0);
373        PPCond[PPStartCond.back()].second = index;
374        PPStartCond.pop_back();
375        // Add the new entry to PPCond.
376        PPCond.push_back(std::make_pair(HashOff, index));
377        EmitToken(Tok);
378
379        // Some files have gibberish on the same line as '#endif'.
380        // Discard these tokens.
381        do
382          L.LexFromRawLexer(Tok);
383        while (Tok.isNot(tok::eof) && !Tok.isAtStartOfLine());
384        // We have the next token in hand.
385        // Don't immediately lex the next one.
386        goto NextToken;
387      }
388      case tok::pp_elif:
389      case tok::pp_else: {
390        // Add an entry for #elif or #else.
391        // This serves as both a closing and opening of a conditional block.
392        // This means that its entry will get backpatched later.
393        unsigned index = PPCond.size();
394        // Backpatch the previous '#if' entry.
395        assert(!PPStartCond.empty());
396        assert(PPCond.size() > PPStartCond.back());
397        assert(PPCond[PPStartCond.back()].second == 0);
398        PPCond[PPStartCond.back()].second = index;
399        PPStartCond.pop_back();
400        // Now add '#elif' as a new block opening.
401        PPCond.push_back(std::make_pair(HashOff, 0U));
402        PPStartCond.push_back(index);
403        break;
404      }
405      }
406    }
407
408    EmitToken(Tok);
409  }
410  while (Tok.isNot(tok::eof));
411
412  assert(PPStartCond.empty() && "Error: imblanced preprocessor conditionals.");
413
414  // Next write out PPCond.
415  Offset PPCondOff = (Offset) Out.tell();
416
417  // Write out the size of PPCond so that clients can identifer empty tables.
418  Emit32(PPCond.size());
419
420  for (unsigned i = 0, e = PPCond.size(); i!=e; ++i) {
421    Emit32(PPCond[i].first - off);
422    uint32_t x = PPCond[i].second;
423    assert(x != 0 && "PPCond entry not backpatched.");
424    // Emit zero for #endifs.  This allows us to do checking when
425    // we read the PTH file back in.
426    Emit32(x == i ? 0 : x);
427  }
428
429  return PTHEntry(off, PPCondOff);
430}
431
432Offset PTHWriter::EmitCachedSpellings() {
433  // Write each cached strings to the PTH file.
434  Offset SpellingsOff = Out.tell();
435
436  for (std::vector<llvm::StringMapEntry<OffsetOpt>*>::iterator
437       I = StrEntries.begin(), E = StrEntries.end(); I!=E; ++I)
438    EmitBuf((*I)->getKeyData(), (*I)->getKeyLength()+1 /*nul included*/);
439
440  return SpellingsOff;
441}
442
443void PTHWriter::GeneratePTH(const std::string &MainFile) {
444  // Generate the prologue.
445  Out << "cfe-pth";
446  Emit32(PTHManager::Version);
447
448  // Leave 4 words for the prologue.
449  Offset PrologueOffset = Out.tell();
450  for (unsigned i = 0; i < 4; ++i)
451    Emit32(0);
452
453  // Write the name of the MainFile.
454  if (!MainFile.empty()) {
455  	EmitString(MainFile);
456  } else {
457    // String with 0 bytes.
458    Emit16(0);
459  }
460  Emit8(0);
461
462  // Iterate over all the files in SourceManager.  Create a lexer
463  // for each file and cache the tokens.
464  SourceManager &SM = PP.getSourceManager();
465  const LangOptions &LOpts = PP.getLangOptions();
466
467  for (SourceManager::fileinfo_iterator I = SM.fileinfo_begin(),
468       E = SM.fileinfo_end(); I != E; ++I) {
469    const SrcMgr::ContentCache &C = *I->second;
470    const FileEntry *FE = C.Entry;
471
472    // FIXME: Handle files with non-absolute paths.
473    llvm::sys::Path P(FE->getName());
474    if (!P.isAbsolute())
475      continue;
476
477    const llvm::MemoryBuffer *B = C.getBuffer(PP.getDiagnostics());
478    if (!B) continue;
479
480    FileID FID = SM.createFileID(FE, SourceLocation(), SrcMgr::C_User);
481    const llvm::MemoryBuffer *FromFile = SM.getBuffer(FID);
482    Lexer L(FID, FromFile, SM, LOpts);
483    PM.insert(FE, LexTokens(L));
484  }
485
486  // Write out the identifier table.
487  const std::pair<Offset,Offset> &IdTableOff = EmitIdentifierTable();
488
489  // Write out the cached strings table.
490  Offset SpellingOff = EmitCachedSpellings();
491
492  // Write out the file table.
493  Offset FileTableOff = EmitFileTable();
494
495  // Finally, write the prologue.
496  Out.seek(PrologueOffset);
497  Emit32(IdTableOff.first);
498  Emit32(IdTableOff.second);
499  Emit32(FileTableOff);
500  Emit32(SpellingOff);
501}
502
503namespace {
504/// StatListener - A simple "interpose" object used to monitor stat calls
505/// invoked by FileManager while processing the original sources used
506/// as input to PTH generation.  StatListener populates the PTHWriter's
507/// file map with stat information for directories as well as negative stats.
508/// Stat information for files are populated elsewhere.
509class StatListener : public StatSysCallCache {
510  PTHMap &PM;
511public:
512  StatListener(PTHMap &pm) : PM(pm) {}
513  ~StatListener() {}
514
515  int stat(const char *path, struct stat *buf) {
516    int result = StatSysCallCache::stat(path, buf);
517
518    if (result != 0) // Failed 'stat'.
519      PM.insert(PTHEntryKeyVariant(path), PTHEntry());
520    else if (S_ISDIR(buf->st_mode)) {
521      // Only cache directories with absolute paths.
522      if (!llvm::sys::Path(path).isAbsolute())
523        return result;
524
525      PM.insert(PTHEntryKeyVariant(buf, path), PTHEntry());
526    }
527
528    return result;
529  }
530};
531} // end anonymous namespace
532
533
534void clang::CacheTokens(Preprocessor &PP, llvm::raw_fd_ostream* OS) {
535  // Get the name of the main file.
536  const SourceManager &SrcMgr = PP.getSourceManager();
537  const FileEntry *MainFile = SrcMgr.getFileEntryForID(SrcMgr.getMainFileID());
538  llvm::sys::Path MainFilePath(MainFile->getName());
539
540  MainFilePath.makeAbsolute();
541
542  // Create the PTHWriter.
543  PTHWriter PW(*OS, PP);
544
545  // Install the 'stat' system call listener in the FileManager.
546  StatListener *StatCache = new StatListener(PW.getPM());
547  PP.getFileManager().addStatCache(StatCache, /*AtBeginning=*/true);
548
549  // Lex through the entire file.  This will populate SourceManager with
550  // all of the header information.
551  Token Tok;
552  PP.EnterMainSourceFile();
553  do { PP.Lex(Tok); } while (Tok.isNot(tok::eof));
554
555  // Generate the PTH file.
556  PP.getFileManager().removeStatCache(StatCache);
557  PW.GeneratePTH(MainFilePath.str());
558}
559
560//===----------------------------------------------------------------------===//
561
562namespace {
563class PTHIdKey {
564public:
565  const IdentifierInfo* II;
566  uint32_t FileOffset;
567};
568
569class PTHIdentifierTableTrait {
570public:
571  typedef PTHIdKey* key_type;
572  typedef key_type  key_type_ref;
573
574  typedef uint32_t  data_type;
575  typedef data_type data_type_ref;
576
577  static unsigned ComputeHash(PTHIdKey* key) {
578    return llvm::HashString(key->II->getName());
579  }
580
581  static std::pair<unsigned,unsigned>
582  EmitKeyDataLength(llvm::raw_ostream& Out, const PTHIdKey* key, uint32_t) {
583    unsigned n = key->II->getLength() + 1;
584    ::Emit16(Out, n);
585    return std::make_pair(n, sizeof(uint32_t));
586  }
587
588  static void EmitKey(llvm::raw_ostream& Out, PTHIdKey* key, unsigned n) {
589    // Record the location of the key data.  This is used when generating
590    // the mapping from persistent IDs to strings.
591    key->FileOffset = Out.tell();
592    Out.write(key->II->getNameStart(), n);
593  }
594
595  static void EmitData(llvm::raw_ostream& Out, PTHIdKey*, uint32_t pID,
596                       unsigned) {
597    ::Emit32(Out, pID);
598  }
599};
600} // end anonymous namespace
601
602/// EmitIdentifierTable - Emits two tables to the PTH file.  The first is
603///  a hashtable mapping from identifier strings to persistent IDs.  The second
604///  is a straight table mapping from persistent IDs to string data (the
605///  keys of the first table).
606///
607std::pair<Offset,Offset> PTHWriter::EmitIdentifierTable() {
608  // Build two maps:
609  //  (1) an inverse map from persistent IDs -> (IdentifierInfo*,Offset)
610  //  (2) a map from (IdentifierInfo*, Offset)* -> persistent IDs
611
612  // Note that we use 'calloc', so all the bytes are 0.
613  PTHIdKey *IIDMap = (PTHIdKey*)calloc(idcount, sizeof(PTHIdKey));
614
615  // Create the hashtable.
616  OnDiskChainedHashTableGenerator<PTHIdentifierTableTrait> IIOffMap;
617
618  // Generate mapping from persistent IDs -> IdentifierInfo*.
619  for (IDMap::iterator I = IM.begin(), E = IM.end(); I != E; ++I) {
620    // Decrement by 1 because we are using a vector for the lookup and
621    // 0 is reserved for NULL.
622    assert(I->second > 0);
623    assert(I->second-1 < idcount);
624    unsigned idx = I->second-1;
625
626    // Store the mapping from persistent ID to IdentifierInfo*
627    IIDMap[idx].II = I->first;
628
629    // Store the reverse mapping in a hashtable.
630    IIOffMap.insert(&IIDMap[idx], I->second);
631  }
632
633  // Write out the inverse map first.  This causes the PCIDKey entries to
634  // record PTH file offsets for the string data.  This is used to write
635  // the second table.
636  Offset StringTableOffset = IIOffMap.Emit(Out);
637
638  // Now emit the table mapping from persistent IDs to PTH file offsets.
639  Offset IDOff = Out.tell();
640  Emit32(idcount);  // Emit the number of identifiers.
641  for (unsigned i = 0 ; i < idcount; ++i)
642    Emit32(IIDMap[i].FileOffset);
643
644  // Finally, release the inverse map.
645  free(IIDMap);
646
647  return std::make_pair(IDOff, StringTableOffset);
648}
649