CacheTokens.cpp revision 738555ca3005c529b4725e72ebd54ca58ec1c93f 
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  /// EmitIdentifierTable - Emits two tables to the PTH file.  The first is
200  ///  a hashtable mapping from identifier strings to persistent IDs.
201  ///  The second is a straight table mapping from persistent IDs to string data
202  ///  (the keys of the first table).
203  std::pair<Offset, Offset> EmitIdentifierTable();
204
205  /// EmitFileTable - Emit a table mapping from file name strings to PTH
206  /// token data.
207  Offset EmitFileTable() { return PM.Emit(Out); }
208
209  PTHEntry LexTokens(Lexer& L);
210  Offset EmitCachedSpellings();
211
212public:
213  PTHWriter(llvm::raw_fd_ostream& out, Preprocessor& pp)
214    : Out(out), PP(pp), idcount(0), CurStrOffset(0) {}
215
216  PTHMap &getPM() { return PM; }
217  void GeneratePTH(const std::string &MainFile);
218};
219} // end anonymous namespace
220
221uint32_t PTHWriter::ResolveID(const IdentifierInfo* II) {
222  // Null IdentifierInfo's map to the persistent ID 0.
223  if (!II)
224    return 0;
225
226  IDMap::iterator I = IM.find(II);
227  if (I != IM.end())
228    return I->second; // We've already added 1.
229
230  IM[II] = ++idcount; // Pre-increment since '0' is reserved for NULL.
231  return idcount;
232}
233
234void PTHWriter::EmitToken(const Token& T) {
235  // Emit the token kind, flags, and length.
236  Emit32(((uint32_t) T.getKind()) | ((((uint32_t) T.getFlags())) << 8)|
237         (((uint32_t) T.getLength()) << 16));
238
239  if (!T.isLiteral()) {
240    Emit32(ResolveID(T.getIdentifierInfo()));
241  } else {
242    // We cache *un-cleaned* spellings. This gives us 100% fidelity with the
243    // source code.
244    const char* s = T.getLiteralData();
245    unsigned len = T.getLength();
246
247    // Get the string entry.
248    llvm::StringMapEntry<OffsetOpt> *E = &CachedStrs.GetOrCreateValue(s, s+len);
249
250    // If this is a new string entry, bump the PTH offset.
251    if (!E->getValue().hasOffset()) {
252      E->getValue().setOffset(CurStrOffset);
253      StrEntries.push_back(E);
254      CurStrOffset += len + 1;
255    }
256
257    // Emit the relative offset into the PTH file for the spelling string.
258    Emit32(E->getValue().getOffset());
259  }
260
261  // Emit the offset into the original source file of this token so that we
262  // can reconstruct its SourceLocation.
263  Emit32(PP.getSourceManager().getFileOffset(T.getLocation()));
264}
265
266PTHEntry PTHWriter::LexTokens(Lexer& L) {
267  // Pad 0's so that we emit tokens to a 4-byte alignment.
268  // This speed up reading them back in.
269  Pad(Out, 4);
270  Offset off = (Offset) Out.tell();
271
272  // Keep track of matching '#if' ... '#endif'.
273  typedef std::vector<std::pair<Offset, unsigned> > PPCondTable;
274  PPCondTable PPCond;
275  std::vector<unsigned> PPStartCond;
276  bool ParsingPreprocessorDirective = false;
277  Token Tok;
278
279  do {
280    L.LexFromRawLexer(Tok);
281  NextToken:
282
283    if ((Tok.isAtStartOfLine() || Tok.is(tok::eof)) &&
284        ParsingPreprocessorDirective) {
285      // Insert an eom token into the token cache.  It has the same
286      // position as the next token that is not on the same line as the
287      // preprocessor directive.  Observe that we continue processing
288      // 'Tok' when we exit this branch.
289      Token Tmp = Tok;
290      Tmp.setKind(tok::eom);
291      Tmp.clearFlag(Token::StartOfLine);
292      Tmp.setIdentifierInfo(0);
293      EmitToken(Tmp);
294      ParsingPreprocessorDirective = false;
295    }
296
297    if (Tok.is(tok::identifier)) {
298      PP.LookUpIdentifierInfo(Tok);
299      EmitToken(Tok);
300      continue;
301    }
302
303    if (Tok.is(tok::hash) && Tok.isAtStartOfLine()) {
304      // Special processing for #include.  Store the '#' token and lex
305      // the next token.
306      assert(!ParsingPreprocessorDirective);
307      Offset HashOff = (Offset) Out.tell();
308      EmitToken(Tok);
309
310      // Get the next token.
311      L.LexFromRawLexer(Tok);
312
313      // If we see the start of line, then we had a null directive "#".
314      if (Tok.isAtStartOfLine())
315        goto NextToken;
316
317      // Did we see 'include'/'import'/'include_next'?
318      if (Tok.isNot(tok::identifier)) {
319        EmitToken(Tok);
320        continue;
321      }
322
323      IdentifierInfo* II = PP.LookUpIdentifierInfo(Tok);
324      tok::PPKeywordKind K = II->getPPKeywordID();
325
326      ParsingPreprocessorDirective = true;
327
328      switch (K) {
329      case tok::pp_not_keyword:
330        // Invalid directives "#foo" can occur in #if 0 blocks etc, just pass
331        // them through.
332      default:
333        break;
334
335      case tok::pp_include:
336      case tok::pp_import:
337      case tok::pp_include_next: {
338        // Save the 'include' token.
339        EmitToken(Tok);
340        // Lex the next token as an include string.
341        L.setParsingPreprocessorDirective(true);
342        L.LexIncludeFilename(Tok);
343        L.setParsingPreprocessorDirective(false);
344        assert(!Tok.isAtStartOfLine());
345        if (Tok.is(tok::identifier))
346          PP.LookUpIdentifierInfo(Tok);
347
348        break;
349      }
350      case tok::pp_if:
351      case tok::pp_ifdef:
352      case tok::pp_ifndef: {
353        // Add an entry for '#if' and friends.  We initially set the target
354        // index to 0.  This will get backpatched when we hit #endif.
355        PPStartCond.push_back(PPCond.size());
356        PPCond.push_back(std::make_pair(HashOff, 0U));
357        break;
358      }
359      case tok::pp_endif: {
360        // Add an entry for '#endif'.  We set the target table index to itself.
361        // This will later be set to zero when emitting to the PTH file.  We
362        // use 0 for uninitialized indices because that is easier to debug.
363        unsigned index = PPCond.size();
364        // Backpatch the opening '#if' entry.
365        assert(!PPStartCond.empty());
366        assert(PPCond.size() > PPStartCond.back());
367        assert(PPCond[PPStartCond.back()].second == 0);
368        PPCond[PPStartCond.back()].second = index;
369        PPStartCond.pop_back();
370        // Add the new entry to PPCond.
371        PPCond.push_back(std::make_pair(HashOff, index));
372        EmitToken(Tok);
373
374        // Some files have gibberish on the same line as '#endif'.
375        // Discard these tokens.
376        do
377          L.LexFromRawLexer(Tok);
378        while (Tok.isNot(tok::eof) && !Tok.isAtStartOfLine());
379        // We have the next token in hand.
380        // Don't immediately lex the next one.
381        goto NextToken;
382      }
383      case tok::pp_elif:
384      case tok::pp_else: {
385        // Add an entry for #elif or #else.
386        // This serves as both a closing and opening of a conditional block.
387        // This means that its entry will get backpatched later.
388        unsigned index = PPCond.size();
389        // Backpatch the previous '#if' entry.
390        assert(!PPStartCond.empty());
391        assert(PPCond.size() > PPStartCond.back());
392        assert(PPCond[PPStartCond.back()].second == 0);
393        PPCond[PPStartCond.back()].second = index;
394        PPStartCond.pop_back();
395        // Now add '#elif' as a new block opening.
396        PPCond.push_back(std::make_pair(HashOff, 0U));
397        PPStartCond.push_back(index);
398        break;
399      }
400      }
401    }
402
403    EmitToken(Tok);
404  }
405  while (Tok.isNot(tok::eof));
406
407  assert(PPStartCond.empty() && "Error: imblanced preprocessor conditionals.");
408
409  // Next write out PPCond.
410  Offset PPCondOff = (Offset) Out.tell();
411
412  // Write out the size of PPCond so that clients can identifer empty tables.
413  Emit32(PPCond.size());
414
415  for (unsigned i = 0, e = PPCond.size(); i!=e; ++i) {
416    Emit32(PPCond[i].first - off);
417    uint32_t x = PPCond[i].second;
418    assert(x != 0 && "PPCond entry not backpatched.");
419    // Emit zero for #endifs.  This allows us to do checking when
420    // we read the PTH file back in.
421    Emit32(x == i ? 0 : x);
422  }
423
424  return PTHEntry(off, PPCondOff);
425}
426
427Offset PTHWriter::EmitCachedSpellings() {
428  // Write each cached strings to the PTH file.
429  Offset SpellingsOff = Out.tell();
430
431  for (std::vector<llvm::StringMapEntry<OffsetOpt>*>::iterator
432       I = StrEntries.begin(), E = StrEntries.end(); I!=E; ++I)
433    EmitBuf((*I)->getKeyData(), (*I)->getKeyLength()+1 /*nul included*/);
434
435  return SpellingsOff;
436}
437
438void PTHWriter::GeneratePTH(const std::string &MainFile) {
439  // Generate the prologue.
440  Out << "cfe-pth";
441  Emit32(PTHManager::Version);
442
443  // Leave 4 words for the prologue.
444  Offset PrologueOffset = Out.tell();
445  for (unsigned i = 0; i < 4; ++i)
446    Emit32(0);
447
448  // Write the name of the MainFile.
449  if (!MainFile.empty()) {
450    Emit16(MainFile.length());
451    EmitBuf(MainFile.data(), MainFile.length());
452  } else {
453    // String with 0 bytes.
454    Emit16(0);
455  }
456  Emit8(0);
457
458  // Iterate over all the files in SourceManager.  Create a lexer
459  // for each file and cache the tokens.
460  SourceManager &SM = PP.getSourceManager();
461  const LangOptions &LOpts = PP.getLangOptions();
462
463  for (SourceManager::fileinfo_iterator I = SM.fileinfo_begin(),
464       E = SM.fileinfo_end(); I != E; ++I) {
465    const SrcMgr::ContentCache &C = *I->second;
466    const FileEntry *FE = C.Entry;
467
468    // FIXME: Handle files with non-absolute paths.
469    llvm::sys::Path P(FE->getName());
470    if (!P.isAbsolute())
471      continue;
472
473    const llvm::MemoryBuffer *B = C.getBuffer();
474    if (!B) continue;
475
476    FileID FID = SM.createFileID(FE, SourceLocation(), SrcMgr::C_User);
477    const llvm::MemoryBuffer *FromFile = SM.getBuffer(FID);
478    Lexer L(FID, FromFile, SM, LOpts);
479    PM.insert(FE, LexTokens(L));
480  }
481
482  // Write out the identifier table.
483  const std::pair<Offset,Offset> &IdTableOff = EmitIdentifierTable();
484
485  // Write out the cached strings table.
486  Offset SpellingOff = EmitCachedSpellings();
487
488  // Write out the file table.
489  Offset FileTableOff = EmitFileTable();
490
491  // Finally, write the prologue.
492  Out.seek(PrologueOffset);
493  Emit32(IdTableOff.first);
494  Emit32(IdTableOff.second);
495  Emit32(FileTableOff);
496  Emit32(SpellingOff);
497}
498
499namespace {
500/// StatListener - A simple "interpose" object used to monitor stat calls
501/// invoked by FileManager while processing the original sources used
502/// as input to PTH generation.  StatListener populates the PTHWriter's
503/// file map with stat information for directories as well as negative stats.
504/// Stat information for files are populated elsewhere.
505class StatListener : public StatSysCallCache {
506  PTHMap &PM;
507public:
508  StatListener(PTHMap &pm) : PM(pm) {}
509  ~StatListener() {}
510
511  int stat(const char *path, struct stat *buf) {
512    int result = StatSysCallCache::stat(path, buf);
513
514    if (result != 0) // Failed 'stat'.
515      PM.insert(PTHEntryKeyVariant(path), PTHEntry());
516    else if (S_ISDIR(buf->st_mode)) {
517      // Only cache directories with absolute paths.
518      if (!llvm::sys::Path(path).isAbsolute())
519        return result;
520
521      PM.insert(PTHEntryKeyVariant(buf, path), PTHEntry());
522    }
523
524    return result;
525  }
526};
527} // end anonymous namespace
528
529
530void clang::CacheTokens(Preprocessor &PP, llvm::raw_fd_ostream* OS) {
531  // Get the name of the main file.
532  const SourceManager &SrcMgr = PP.getSourceManager();
533  const FileEntry *MainFile = SrcMgr.getFileEntryForID(SrcMgr.getMainFileID());
534  llvm::sys::Path MainFilePath(MainFile->getName());
535
536  MainFilePath.makeAbsolute();
537
538  // Create the PTHWriter.
539  PTHWriter PW(*OS, PP);
540
541  // Install the 'stat' system call listener in the FileManager.
542  StatListener *StatCache = new StatListener(PW.getPM());
543  PP.getFileManager().addStatCache(StatCache, /*AtBeginning=*/true);
544
545  // Lex through the entire file.  This will populate SourceManager with
546  // all of the header information.
547  Token Tok;
548  PP.EnterMainSourceFile();
549  do { PP.Lex(Tok); } while (Tok.isNot(tok::eof));
550
551  // Generate the PTH file.
552  PP.getFileManager().removeStatCache(StatCache);
553  PW.GeneratePTH(MainFilePath.str());
554}
555
556//===----------------------------------------------------------------------===//
557
558namespace {
559class PTHIdKey {
560public:
561  const IdentifierInfo* II;
562  uint32_t FileOffset;
563};
564
565class PTHIdentifierTableTrait {
566public:
567  typedef PTHIdKey* key_type;
568  typedef key_type  key_type_ref;
569
570  typedef uint32_t  data_type;
571  typedef data_type data_type_ref;
572
573  static unsigned ComputeHash(PTHIdKey* key) {
574    return llvm::HashString(key->II->getName());
575  }
576
577  static std::pair<unsigned,unsigned>
578  EmitKeyDataLength(llvm::raw_ostream& Out, const PTHIdKey* key, uint32_t) {
579    unsigned n = key->II->getLength() + 1;
580    ::Emit16(Out, n);
581    return std::make_pair(n, sizeof(uint32_t));
582  }
583
584  static void EmitKey(llvm::raw_ostream& Out, PTHIdKey* key, unsigned n) {
585    // Record the location of the key data.  This is used when generating
586    // the mapping from persistent IDs to strings.
587    key->FileOffset = Out.tell();
588    Out.write(key->II->getNameStart(), n);
589  }
590
591  static void EmitData(llvm::raw_ostream& Out, PTHIdKey*, uint32_t pID,
592                       unsigned) {
593    ::Emit32(Out, pID);
594  }
595};
596} // end anonymous namespace
597
598/// EmitIdentifierTable - Emits two tables to the PTH file.  The first is
599///  a hashtable mapping from identifier strings to persistent IDs.  The second
600///  is a straight table mapping from persistent IDs to string data (the
601///  keys of the first table).
602///
603std::pair<Offset,Offset> PTHWriter::EmitIdentifierTable() {
604  // Build two maps:
605  //  (1) an inverse map from persistent IDs -> (IdentifierInfo*,Offset)
606  //  (2) a map from (IdentifierInfo*, Offset)* -> persistent IDs
607
608  // Note that we use 'calloc', so all the bytes are 0.
609  PTHIdKey *IIDMap = (PTHIdKey*)calloc(idcount, sizeof(PTHIdKey));
610
611  // Create the hashtable.
612  OnDiskChainedHashTableGenerator<PTHIdentifierTableTrait> IIOffMap;
613
614  // Generate mapping from persistent IDs -> IdentifierInfo*.
615  for (IDMap::iterator I = IM.begin(), E = IM.end(); I != E; ++I) {
616    // Decrement by 1 because we are using a vector for the lookup and
617    // 0 is reserved for NULL.
618    assert(I->second > 0);
619    assert(I->second-1 < idcount);
620    unsigned idx = I->second-1;
621
622    // Store the mapping from persistent ID to IdentifierInfo*
623    IIDMap[idx].II = I->first;
624
625    // Store the reverse mapping in a hashtable.
626    IIOffMap.insert(&IIDMap[idx], I->second);
627  }
628
629  // Write out the inverse map first.  This causes the PCIDKey entries to
630  // record PTH file offsets for the string data.  This is used to write
631  // the second table.
632  Offset StringTableOffset = IIOffMap.Emit(Out);
633
634  // Now emit the table mapping from persistent IDs to PTH file offsets.
635  Offset IDOff = Out.tell();
636  Emit32(idcount);  // Emit the number of identifiers.
637  for (unsigned i = 0 ; i < idcount; ++i)
638    Emit32(IIDMap[i].FileOffset);
639
640  // Finally, release the inverse map.
641  free(IIDMap);
642
643  return std::make_pair(IDOff, StringTableOffset);
644}
645