src/MachOReaderDylib.hpp

   1 /* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
   2  *
   3  * Copyright (c) 2005-2006 Apple Computer, Inc. All rights reserved.
   4  *
   5  * @APPLE_LICENSE_HEADER_START@
   6  *
   7  * This file contains Original Code and/or Modifications of Original Code
   8  * as defined in and that are subject to the Apple Public Source License
   9  * Version 2.0 (the 'License'). You may not use this file except in
  10  * compliance with the License. Please obtain a copy of the License at
  11  * http://www.opensource.apple.com/apsl/ and read it before using this
  12  * file.
  13  *
  14  * The Original Code and all software distributed under the License are
  15  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  16  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  17  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  18  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  19  * Please see the License for the specific language governing rights and
  20  * limitations under the License.
  21  *
  22  * @APPLE_LICENSE_HEADER_END@
  23  */
  24
  25 #ifndef __OBJECT_FILE_DYLIB_MACH_O__
  26 #define __OBJECT_FILE_DYLIB_MACH_O__
  27
  28 #include <stdint.h>
  29 #include <math.h>
  30 #include <unistd.h>
  31 #include <sys/param.h>
  32
  33
  34 #include <vector>
  35 #include <set>
  36 #include <algorithm>
  37 #include <ext/hash_map>
  38
  39 #include "MachOFileAbstraction.hpp"
  40 #include "ObjectFile.h"
  41
  42 //
  43 //
  44 //      To implement architecture xxx, you must write template specializations for the following method:
  45 //                      Reader<xxx>::validFile()
  46 //
  47 //
  48
  49
  50
  51
  52 namespace mach_o {
  53 namespace dylib {
  54
  55
  56 // forward reference
  57 template <typename A> class Reader;
  58
  59
  60 class Segment : public ObjectFile::Segment
  61 {
  62 public:
  63                                                                 Segment(const char* name)               { fName = name; }
  64         virtual const char*                     getName() const                                 { return fName; }
  65         virtual bool                            isContentReadable() const               { return true; }
  66         virtual bool                            isContentWritable() const               { return false; }
  67         virtual bool                            isContentExecutable() const             { return false; }
  68 private:
  69         const char*                                     fName;
  70 };
  71
  72
  73 //
  74 // An ExportAtom has no content.  It exists so that the linker can track which imported
  75 // symbols can from which dynamic libraries.
  76 //
  77 template <typename A>
  78 class ExportAtom : public ObjectFile::Atom
  79 {
  80 public:
  81         virtual ObjectFile::Reader*                                     getFile() const                         { return &fOwner; }
  82         virtual bool                                                            getTranslationUnitSource(const char** dir, const char** name) const { return false; }
  83         virtual const char*                                                     getName() const                         { return fName; }
  84         virtual const char*                                                     getDisplayName() const          { return fName; }
  85         virtual Scope                                                           getScope() const                        { return ObjectFile::Atom::scopeGlobal; }
  86         virtual DefinitionKind                                          getDefinitionKind() const       { return fWeakDefinition ? kExternalWeakDefinition : kExternalDefinition; }
  87         virtual SymbolTableInclusion                            getSymbolTableInclusion() const { return ObjectFile::Atom::kSymbolTableIn; }
  88         virtual bool                                                            isZeroFill() const                      { return false; }
  89         virtual uint64_t                                                        getSize() const                         { return 0; }
  90         virtual std::vector<ObjectFile::Reference*>&  getReferences() const             { return fgEmptyReferenceList; }
  91         virtual bool                                                            mustRemainInSection() const { return false; }
  92         virtual const char*                                                     getSectionName() const          { return "._imports"; }
  93         virtual Segment&                                                        getSegment() const                      { return fgImportSegment; }
  94         virtual bool                                                            requiresFollowOnAtom() const{ return false; }
  95         virtual ObjectFile::Atom&                                       getFollowOnAtom() const         { return *((ObjectFile::Atom*)NULL); }
  96         virtual std::vector<ObjectFile::LineInfo>*      getLineInfo() const                     { return NULL; }
  97         virtual uint8_t                                                         getAlignment() const            { return 0; }
  98         virtual void                                                            copyRawContent(uint8_t buffer[]) const  {}
  99
 100         virtual void                                                            setScope(Scope)                         { }
 101
 102 protected:
 103         friend class Reader<A>;
 104         typedef typename A::P                                   P;
 105
 106                                                                                         ExportAtom(ObjectFile::Reader& owner, const char* name, bool weak)
 107                                                                                                 : fOwner(owner), fName(name), fWeakDefinition(weak) {}
 108         virtual                                                                 ~ExportAtom() {}
 109
 110         ObjectFile::Reader&                                             fOwner;
 111         const char*                                                             fName;
 112         bool                                                                    fWeakDefinition;
 113
 114         static std::vector<ObjectFile::Reference*>      fgEmptyReferenceList;
 115         static Segment                                                          fgImportSegment;
 116 };
 117
 118 template <typename A>
 119 Segment                                                         ExportAtom<A>::fgImportSegment("__LINKEDIT");
 120
 121 template <typename A>
 122 std::vector<ObjectFile::Reference*>     ExportAtom<A>::fgEmptyReferenceList;
 123
 124
 125 //
 126 // The reader for a dylib extracts all exported symbols names from the memory-mapped
 127 // dylib, builds a hash table, then unmaps the file.  This is an important memory
 128 // savings for large dylibs.
 129 //
 130 template <typename A>
 131 class Reader : public ObjectFile::Reader
 132 {
 133 public:
 134         static bool                                                                             validFile(const uint8_t* fileContent);
 135         static Reader<A>*                                                               make(const uint8_t* fileContent, uint64_t fileLength, const char* path, const ObjectFile::ReaderOptions& options)
 136                                                                                                                 { return new Reader<A>(fileContent, fileLength, path, options); }
 137         virtual                                                                         ~Reader() {}
 138
 139         virtual const char*                                                             getPath()                                       { return fPath; }
 140         virtual time_t                                                                  getModificationTime()           { return 0; }
 141         virtual DebugInfoKind                                                   getDebugInfoKind()                      { return ObjectFile::Reader::kDebugInfoNone; }
 142         virtual std::vector<class ObjectFile::Atom*>&   getAtoms();
 143         virtual std::vector<class ObjectFile::Atom*>*   getJustInTimeAtomsFor(const char* name);
 144         virtual std::vector<Stab>*                                              getStabs()                                      { return NULL; }
 145         virtual const char*                                                             getInstallPath()                        { return fDylibInstallPath; }
 146         virtual uint32_t                                                                getTimestamp()                          { return fDylibTimeStamp; }
 147         virtual uint32_t                                                                getCurrentVersion()                     { return fDylibtCurrentVersion; }
 148         virtual uint32_t                                                                getCompatibilityVersion()       { return fDylibCompatibilityVersion; }
 149         virtual std::vector<const char*>*                               getDependentLibraryPaths();
 150         virtual bool                                                                    reExports(ObjectFile::Reader*);
 151         virtual std::vector<const char*>*                               getAllowableClients();
 152
 153 protected:
 154                 const char*                                                                     parentUmbrella()                        { return fParentUmbrella; }
 155
 156 private:
 157         typedef typename A::P                                           P;
 158         typedef typename A::P::E                                        E;
 159
 160         class CStringEquals
 161         {
 162         public:
 163                 bool operator()(const char* left, const char* right) const { return (strcmp(left, right) == 0); }
 164         };
 165         struct AtomAndWeak { ObjectFile::Atom* atom; bool weak; };
 166         typedef __gnu_cxx::hash_map<const char*, AtomAndWeak, __gnu_cxx::hash<const char*>, CStringEquals> NameToAtomMap;
 167         typedef typename NameToAtomMap::iterator                NameToAtomMapIterator;
 168
 169         struct PathAndFlag { const char* path; bool reExport; };
 170
 171                                                                                                 Reader(const uint8_t* fileContent, uint64_t fileLength, const char* path, const ObjectFile::ReaderOptions& options);
 172
 173         const char*                                                                     fPath;
 174         const char*                                                                     fParentUmbrella;
 175         std::vector<const char*>                                        fAllowableClients;
 176         const char*                                                                     fDylibInstallPath;
 177         uint32_t                                                                        fDylibTimeStamp;
 178         uint32_t                                                                        fDylibtCurrentVersion;
 179         uint32_t                                                                        fDylibCompatibilityVersion;
 180         std::vector<PathAndFlag>                                        fDependentLibraryPaths;
 181         NameToAtomMap                                                           fAtoms;
 182
 183         static bool                                                                     fgLogHashtable;
 184         static std::vector<class ObjectFile::Atom*>     fgEmptyAtomList;
 185 };
 186
 187 template <typename A>
 188 std::vector<class ObjectFile::Atom*>    Reader<A>::fgEmptyAtomList;
 189 template <typename A>
 190 bool                                                                    Reader<A>::fgLogHashtable = false;
 191
 192
 193 template <typename A>
 194 Reader<A>::Reader(const uint8_t* fileContent, uint64_t fileLength, const char* path, const ObjectFile::ReaderOptions& options)
 195         : fParentUmbrella(NULL), fDylibInstallPath(NULL), fDylibTimeStamp(0), fDylibtCurrentVersion(0), fDylibCompatibilityVersion(0)
 196 {
 197         // sanity check
 198         if ( ! validFile(fileContent) )
 199                 throw "not a valid mach-o object file";
 200
 201         fPath = strdup(path);
 202
 203         const macho_header<P>* header = (const macho_header<P>*)fileContent;
 204         const uint32_t cmd_count = header->ncmds();
 205         const macho_load_command<P>* const cmds = (macho_load_command<P>*)((char*)header + sizeof(macho_header<P>));
 206
 207         // a "blank" stub has zero load commands
 208         if ( (header->filetype() == MH_DYLIB_STUB) && (cmd_count == 0) ) {
 209                 // no further processing needed
 210                 munmap((caddr_t)fileContent, fileLength);
 211                 return;
 212         }
 213
 214         // pass 1 builds list of all dependent libraries
 215         const macho_load_command<P>* cmd = cmds;
 216         for (uint32_t i = 0; i < cmd_count; ++i) {
 217                 switch (cmd->cmd()) {
 218                         case LC_LOAD_DYLIB:
 219                         case LC_LOAD_WEAK_DYLIB:
 220                                 PathAndFlag entry;
 221                                 entry.path = strdup(((struct macho_dylib_command<P>*)cmd)->name());
 222                                 entry.reExport = false;
 223                                 fDependentLibraryPaths.push_back(entry);
 224                                 break;
 225                 }
 226                 cmd = (const macho_load_command<P>*)(((char*)cmd)+cmd->cmdsize());
 227         }
 228
 229         // pass 2 determines re-export info
 230         const macho_dysymtab_command<P>* dynamicInfo = NULL;
 231         const macho_nlist<P>* symbolTable = NULL;
 232         const char*     strings = NULL;
 233         cmd = cmds;
 234         for (uint32_t i = 0; i < cmd_count; ++i) {
 235                 switch (cmd->cmd()) {
 236                         case LC_SYMTAB:
 237                                 {
 238                                         const macho_symtab_command<P>* symtab = (macho_symtab_command<P>*)cmd;
 239                                         symbolTable = (const macho_nlist<P>*)((char*)header + symtab->symoff());
 240                                         strings = (char*)header + symtab->stroff();
 241                                 }
 242                                 break;
 243                         case LC_DYSYMTAB:
 244                                 dynamicInfo = (macho_dysymtab_command<P>*)cmd;
 245                                 break;
 246                         case LC_ID_DYLIB:
 247                                 macho_dylib_command<P>* dylibID = (macho_dylib_command<P>*)cmd;
 248                                 fDylibInstallPath                       = strdup(dylibID->name());
 249                                 fDylibTimeStamp                         = dylibID->timestamp();
 250                                 fDylibtCurrentVersion           = dylibID->current_version();
 251                                 fDylibCompatibilityVersion      = dylibID->compatibility_version();
 252                                 break;
 253                         case LC_SUB_UMBRELLA:
 254                                 if ( !options.fFlatNamespace ) {
 255                                         const char* frameworkLeafName = ((macho_sub_umbrella_command<P>*)cmd)->sub_umbrella();
 256                                         for (typename std::vector<PathAndFlag>::iterator it = fDependentLibraryPaths.begin(); it != fDependentLibraryPaths.end(); it++) {
 257                                                 const char* dylibName = it->path;
 258                                                 const char* lastSlash = strrchr(dylibName, '/');
 259                                                 if ( (lastSlash != NULL) && (strcmp(&lastSlash[1], frameworkLeafName) == 0) )
 260                                                         it->reExport = true;
 261                                         }
 262                                 }
 263                                 break;
 264                         case LC_SUB_LIBRARY:
 265                                 if ( !options.fFlatNamespace ) {
 266                                         const char* dylibBaseName = ((macho_sub_library_command<P>*)cmd)->sub_library();
 267                                         for (typename std::vector<PathAndFlag>::iterator it = fDependentLibraryPaths.begin(); it != fDependentLibraryPaths.end(); it++) {
 268                                                 const char* dylibName = it->path;
 269                                                 const char* lastSlash = strrchr(dylibName, '/');
 270                                                 const char* leafStart = &lastSlash[1];
 271                                                 if ( lastSlash == NULL )
 272                                                         leafStart = dylibName;
 273                                                 const char* firstDot = strchr(leafStart, '.');
 274                                                 int len = strlen(leafStart);
 275                                                 if ( firstDot != NULL )
 276                                                         len = firstDot - leafStart;
 277                                                 if ( strncmp(leafStart, dylibBaseName, len) == 0 )
 278                                                         it->reExport = true;
 279                                         }
 280                                 }
 281                                 break;
 282                         case LC_SUB_FRAMEWORK:
 283                                 fParentUmbrella = strdup(((macho_sub_framework_command<P>*)cmd)->umbrella());
 284                                 break;
 285                 }
 286
 287                 cmd = (const macho_load_command<P>*)(((char*)cmd)+cmd->cmdsize());
 288         }
 289         // Process the rest of the commands here.
 290         cmd = cmds;
 291         for (uint32_t i = 0; i < cmd_count; ++i) {
 292                 switch (cmd->cmd()) {
 293                 case LC_SUB_CLIENT:
 294                         const char *temp = strdup(((macho_sub_client_command<P>*)cmd)->client());
 295
 296                         fAllowableClients.push_back(temp);
 297                         break;
 298                 }
 299
 300                 cmd = (const macho_load_command<P>*)(((char*)cmd)+cmd->cmdsize());
 301         }
 302
 303         // validate minimal load commands
 304         if ( fDylibInstallPath == NULL )
 305                 throw "dylib missing LC_ID_DYLIB load command";
 306         if ( symbolTable == NULL )
 307                 throw "dylib missing LC_SYMTAB load command";
 308         if ( dynamicInfo == NULL )
 309                 throw "dylib missing LC_DYSYMTAB load command";
 310
 311         // build hash table
 312         if ( dynamicInfo->tocoff() == 0 ) {
 313                 if ( fgLogHashtable ) fprintf(stderr, "ld64: building hashtable of %u toc entries for %s\n", dynamicInfo->nextdefsym(), path);
 314                 const macho_nlist<P>* start = &symbolTable[dynamicInfo->iextdefsym()];
 315                 const macho_nlist<P>* end = &start[dynamicInfo->nextdefsym()];
 316                 fAtoms.resize(dynamicInfo->nextdefsym()); // set initial bucket count
 317                 for (const macho_nlist<P>* sym=start; sym < end; ++sym) {
 318                         AtomAndWeak bucket;
 319                         bucket.atom = NULL;
 320                         bucket.weak = ((sym->n_desc() & N_WEAK_DEF) != 0);
 321                         const char* name = strdup(&strings[sym->n_strx()]);
 322                         if ( fgLogHashtable ) fprintf(stderr, "  adding %s to hash table for %s\n", name, this->getPath());
 323                         fAtoms[name] = bucket;
 324                 }
 325         }
 326         else {
 327                 int32_t count = dynamicInfo->ntoc();
 328                 fAtoms.resize(count); // set initial bucket count
 329                 if ( fgLogHashtable ) fprintf(stderr, "ld64: building hashtable of %u entries for %s\n", count, path);
 330                 const struct dylib_table_of_contents* toc = (dylib_table_of_contents*)((char*)header + dynamicInfo->tocoff());
 331                 for (int32_t i = 0; i < count; ++i) {
 332                         const uint32_t index = E::get32(toc[i].symbol_index);
 333                         const macho_nlist<P>* sym = &symbolTable[index];
 334                         AtomAndWeak bucket;
 335                         bucket.atom = NULL;
 336                         bucket.weak = ((sym->n_desc() & N_WEAK_DEF) != 0);
 337                         const char* name = strdup(&strings[sym->n_strx()]);
 338                         if ( fgLogHashtable ) fprintf(stderr, "  adding %s to hash table for %s\n", name, this->getPath());
 339                         fAtoms[name] = bucket;
 340                 }
 341         }
 342
 343         // unmap file
 344         munmap((caddr_t)fileContent, fileLength);
 345 }
 346
 347 template <typename A>
 348 std::vector<class ObjectFile::Atom*>& Reader<A>::getAtoms()
 349 {
 350         // TO DO:       for flat-namespace libraries, when linking flat_namespace
 351         //                      we need to create an atom which references all undefines
 352         return fgEmptyAtomList;
 353 }
 354
 355
 356 template <typename A>
 357 std::vector<class ObjectFile::Atom*>* Reader<A>::getJustInTimeAtomsFor(const char* name)
 358 {
 359         std::vector<class ObjectFile::Atom*>* atoms = NULL;
 360
 361         NameToAtomMapIterator pos = fAtoms.find(name);
 362         if ( pos != fAtoms.end() ) {
 363                 if ( pos->second.atom == NULL ) {
 364                         // instantiate atom and update hash table
 365                         pos->second.atom = new ExportAtom<A>(*this, name, pos->second.weak);
 366                         if ( fgLogHashtable ) fprintf(stderr, "getJustInTimeAtomsFor: %s found in %s\n", name, this->getPath());
 367                 }
 368                 // return a vector of one atom
 369                 atoms = new std::vector<class ObjectFile::Atom*>;
 370                 atoms->push_back(pos->second.atom);
 371         }
 372         else {
 373                 if ( fgLogHashtable ) fprintf(stderr, "getJustInTimeAtomsFor: %s NOT found in %s\n", name, this->getPath());
 374         }
 375         return atoms;
 376 }
 377
 378
 379
 380 template <typename A>
 381 std::vector<const char*>* Reader<A>::getDependentLibraryPaths()
 382 {
 383         std::vector<const char*>* result = new std::vector<const char*>;
 384         for (typename std::vector<PathAndFlag>::iterator it = fDependentLibraryPaths.begin(); it != fDependentLibraryPaths.end(); it++) {
 385                 result->push_back(it->path);
 386         }
 387         return result;
 388 }
 389
 390 template <typename A>
 391 std::vector<const char*>* Reader<A>::getAllowableClients()
 392 {
 393         std::vector<const char*>* result = new std::vector<const char*>;
 394         for (typename std::vector<const char*>::iterator it = fAllowableClients.begin();
 395                  it != fAllowableClients.end();
 396                  it++) {
 397                 result->push_back(*it);
 398         }
 399         return (fAllowableClients.size() != 0 ? result : NULL);
 400 }
 401
 402 template <typename A>
 403 bool Reader<A>::reExports(ObjectFile::Reader* child)
 404 {
 405         // A dependent dylib is re-exported under two conditions:
 406         //  1) parent contains LC_SUB_UMBRELLA or LC_SUB_LIBRARY with child name
 407         for (typename std::vector<PathAndFlag>::iterator it = fDependentLibraryPaths.begin(); it != fDependentLibraryPaths.end(); it++) {
 408                 if ( it->reExport && (strcmp(it->path, child->getPath()) == 0) )
 409                         return true;
 410         }
 411
 412         //  2) child contains LC_SUB_FRAMEWORK with parent name
 413         const char* parentUmbrellaName = ((Reader<A>*)child)->parentUmbrella();
 414         if ( parentUmbrellaName != NULL ) {
 415                 const char* parentName = this->getPath();
 416                 const char* lastSlash = strrchr(parentName, '/');
 417                 if ( (lastSlash != NULL) && (strcmp(&lastSlash[1], parentUmbrellaName) == 0) )
 418                         return true;
 419         }
 420
 421         return false;
 422 }
 423
 424 template <>
 425 bool Reader<ppc>::validFile(const uint8_t* fileContent)
 426 {
 427         const macho_header<P>* header = (const macho_header<P>*)fileContent;
 428         if ( header->magic() != MH_MAGIC )
 429                 return false;
 430         if ( header->cputype() != CPU_TYPE_POWERPC )
 431                 return false;
 432         if ( (header->filetype() != MH_DYLIB) && (header->filetype() != MH_DYLIB_STUB)  )
 433                 return false;
 434         return true;
 435 }
 436
 437 template <>
 438 bool Reader<ppc64>::validFile(const uint8_t* fileContent)
 439 {
 440         const macho_header<P>* header = (const macho_header<P>*)fileContent;
 441         if ( header->magic() != MH_MAGIC_64 )
 442                 return false;
 443         if ( header->cputype() != CPU_TYPE_POWERPC64 )
 444                 return false;
 445         if ( (header->filetype() != MH_DYLIB) && (header->filetype() != MH_DYLIB_STUB)  )
 446                 return false;
 447         return true;
 448 }
 449
 450 template <>
 451 bool Reader<x86>::validFile(const uint8_t* fileContent)
 452 {
 453         const macho_header<P>* header = (const macho_header<P>*)fileContent;
 454         if ( header->magic() != MH_MAGIC )
 455                 return false;
 456         if ( header->cputype() != CPU_TYPE_I386 )
 457                 return false;
 458         if ( (header->filetype() != MH_DYLIB) && (header->filetype() != MH_DYLIB_STUB)  )
 459                 return false;
 460         return true;
 461 }
 462
 463
 464
 465
 466 }; // namespace dylib
 467 }; // namespace mach_o
 468
 469
 470 #endif // __OBJECT_FILE_DYLIB_MACH_O__