tools/lto/LTOModule.cpp

   1 //===-LTOModule.cpp - LLVM Link Time Optimizer ----------------------------===//
   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 file implements the Link Time Optimization library. This library is
  11 // intended to be used by linker to optimize code at link time.
  12 //
  13 //===----------------------------------------------------------------------===//
  14
  15 #include "LTOModule.h"
  16
  17 #include "llvm/Constants.h"
  18 #include "llvm/LLVMContext.h"
  19 #include "llvm/Module.h"
  20 #include "llvm/ModuleProvider.h"
  21 #include "llvm/ADT/OwningPtr.h"
  22 #include "llvm/Bitcode/ReaderWriter.h"
  23 #include "llvm/Support/SystemUtils.h"
  24 #include "llvm/Support/Mangler.h"
  25 #include "llvm/Support/MemoryBuffer.h"
  26 #include "llvm/Support/MathExtras.h"
  27 #include "llvm/System/Host.h"
  28 #include "llvm/System/Path.h"
  29 #include "llvm/System/Process.h"
  30 #include "llvm/Target/SubtargetFeature.h"
  31 #include "llvm/MC/MCAsmInfo.h"
  32 #include "llvm/Target/TargetMachine.h"
  33 #include "llvm/Target/TargetRegistry.h"
  34 #include "llvm/Target/TargetSelect.h"
  35
  36 using namespace llvm;
  37
  38 bool LTOModule::isBitcodeFile(const void* mem, size_t length)
  39 {
  40     return ( llvm::sys::IdentifyFileType((char*)mem, length)
  41                                             == llvm::sys::Bitcode_FileType );
  42 }
  43
  44 bool LTOModule::isBitcodeFile(const char* path)
  45 {
  46     return llvm::sys::Path(path).isBitcodeFile();
  47 }
  48
  49 bool LTOModule::isBitcodeFileForTarget(const void* mem, size_t length,
  50                                        const char* triplePrefix)
  51 {
  52     MemoryBuffer* buffer = makeBuffer(mem, length);
  53     if ( buffer == NULL )
  54         return false;
  55     return isTargetMatch(buffer, triplePrefix);
  56 }
  57
  58
  59 bool LTOModule::isBitcodeFileForTarget(const char* path,
  60                                        const char* triplePrefix)
  61 {
  62     MemoryBuffer *buffer = MemoryBuffer::getFile(path);
  63     if (buffer == NULL)
  64         return false;
  65     return isTargetMatch(buffer, triplePrefix);
  66 }
  67
  68 // takes ownership of buffer
  69 bool LTOModule::isTargetMatch(MemoryBuffer* buffer, const char* triplePrefix)
  70 {
  71     OwningPtr<ModuleProvider> mp(getBitcodeModuleProvider(buffer,
  72                                                           getGlobalContext()));
  73     // on success, mp owns buffer and both are deleted at end of this method
  74     if ( !mp ) {
  75         delete buffer;
  76         return false;
  77     }
  78     std::string actualTarget = mp->getModule()->getTargetTriple();
  79     return ( strncmp(actualTarget.c_str(), triplePrefix,
  80                     strlen(triplePrefix)) == 0);
  81 }
  82
  83
  84 LTOModule::LTOModule(Module* m, TargetMachine* t)
  85  : _module(m), _target(t), _symbolsParsed(false)
  86 {
  87 }
  88
  89 LTOModule* LTOModule::makeLTOModule(const char* path,
  90                                     std::string& errMsg)
  91 {
  92     OwningPtr<MemoryBuffer> buffer(MemoryBuffer::getFile(path, &errMsg));
  93     if ( !buffer )
  94         return NULL;
  95     return makeLTOModule(buffer.get(), errMsg);
  96 }
  97
  98 /// makeBuffer - create a MemoryBuffer from a memory range.
  99 /// MemoryBuffer requires the byte past end of the buffer to be a zero.
 100 /// We might get lucky and already be that way, otherwise make a copy.
 101 /// Also if next byte is on a different page, don't assume it is readable.
 102 MemoryBuffer* LTOModule::makeBuffer(const void* mem, size_t length)
 103 {
 104     const char* startPtr = (char*)mem;
 105     const char* endPtr = startPtr+length;
 106     if ( (((uintptr_t)endPtr & (sys::Process::GetPageSize()-1)) == 0)
 107         || (*endPtr != 0) )
 108         return MemoryBuffer::getMemBufferCopy(startPtr, endPtr);
 109     else
 110         return MemoryBuffer::getMemBuffer(startPtr, endPtr);
 111 }
 112
 113
 114 LTOModule* LTOModule::makeLTOModule(const void* mem, size_t length,
 115                                     std::string& errMsg)
 116 {
 117     OwningPtr<MemoryBuffer> buffer(makeBuffer(mem, length));
 118     if ( !buffer )
 119         return NULL;
 120     return makeLTOModule(buffer.get(), errMsg);
 121 }
 122
 123 /// getFeatureString - Return a string listing the features associated with the
 124 /// target triple.
 125 ///
 126 /// FIXME: This is an inelegant way of specifying the features of a
 127 /// subtarget. It would be better if we could encode this information into the
 128 /// IR. See <rdar://5972456>.
 129 std::string getFeatureString(const char *TargetTriple) {
 130   InitializeAllTargets();
 131
 132   SubtargetFeatures Features;
 133
 134   if (strncmp(TargetTriple, "powerpc-apple-", 14) == 0) {
 135     Features.AddFeature("altivec", true);
 136   } else if (strncmp(TargetTriple, "powerpc64-apple-", 16) == 0) {
 137     Features.AddFeature("64bit", true);
 138     Features.AddFeature("altivec", true);
 139   }
 140
 141   return Features.getString();
 142 }
 143
 144 LTOModule* LTOModule::makeLTOModule(MemoryBuffer* buffer,
 145                                     std::string& errMsg)
 146 {
 147     InitializeAllTargets();
 148
 149     // parse bitcode buffer
 150     OwningPtr<Module> m(ParseBitcodeFile(buffer, getGlobalContext(), &errMsg));
 151     if ( !m )
 152         return NULL;
 153
 154     std::string Triple = m->getTargetTriple();
 155     if (Triple.empty())
 156       Triple = sys::getHostTriple();
 157
 158     // find machine architecture for this module
 159     const Target* march = TargetRegistry::lookupTarget(Triple, errMsg);
 160     if ( march == NULL )
 161         return NULL;
 162
 163     // construct LTModule, hand over ownership of module and target
 164     std::string FeatureStr = getFeatureString(Triple.c_str());
 165     TargetMachine* target = march->createTargetMachine(Triple, FeatureStr);
 166     return new LTOModule(m.take(), target);
 167 }
 168
 169
 170 const char* LTOModule::getTargetTriple()
 171 {
 172     return _module->getTargetTriple().c_str();
 173 }
 174
 175 void LTOModule::addDefinedFunctionSymbol(Function* f, Mangler &mangler)
 176 {
 177     // add to list of defined symbols
 178     addDefinedSymbol(f, mangler, true);
 179
 180     // add external symbols referenced by this function.
 181     for (Function::iterator b = f->begin(); b != f->end(); ++b) {
 182         for (BasicBlock::iterator i = b->begin(); i != b->end(); ++i) {
 183             for (unsigned count = 0, total = i->getNumOperands();
 184                                         count != total; ++count) {
 185                 findExternalRefs(i->getOperand(count), mangler);
 186             }
 187         }
 188     }
 189 }
 190
 191 // get string that data pointer points to
 192 bool LTOModule::objcClassNameFromExpression(Constant* c, std::string& name)
 193 {
 194     if (ConstantExpr* ce = dyn_cast<ConstantExpr>(c)) {
 195         Constant* op = ce->getOperand(0);
 196         if (GlobalVariable* gvn = dyn_cast<GlobalVariable>(op)) {
 197             Constant* cn = gvn->getInitializer();
 198             if (ConstantArray* ca = dyn_cast<ConstantArray>(cn)) {
 199                 if ( ca->isCString() ) {
 200                     name = ".objc_class_name_" + ca->getAsString();
 201                     return true;
 202                 }
 203             }
 204         }
 205     }
 206     return false;
 207 }
 208
 209 // parse i386/ppc ObjC class data structure
 210 void LTOModule::addObjCClass(GlobalVariable* clgv)
 211 {
 212     if (ConstantStruct* c = dyn_cast<ConstantStruct>(clgv->getInitializer())) {
 213         // second slot in __OBJC,__class is pointer to superclass name
 214         std::string superclassName;
 215         if ( objcClassNameFromExpression(c->getOperand(1), superclassName) ) {
 216             NameAndAttributes info;
 217             if ( _undefines.find(superclassName.c_str()) == _undefines.end() ) {
 218                 const char* symbolName = ::strdup(superclassName.c_str());
 219                 info.name = ::strdup(symbolName);
 220                 info.attributes = LTO_SYMBOL_DEFINITION_UNDEFINED;
 221                 // string is owned by _undefines
 222                 _undefines[info.name] = info;
 223             }
 224         }
 225         // third slot in __OBJC,__class is pointer to class name
 226         std::string className;
 227          if ( objcClassNameFromExpression(c->getOperand(2), className) ) {
 228             const char* symbolName = ::strdup(className.c_str());
 229             NameAndAttributes info;
 230             info.name = symbolName;
 231             info.attributes = (lto_symbol_attributes)
 232                 (LTO_SYMBOL_PERMISSIONS_DATA |
 233                  LTO_SYMBOL_DEFINITION_REGULAR |
 234                  LTO_SYMBOL_SCOPE_DEFAULT);
 235             _symbols.push_back(info);
 236             _defines[info.name] = 1;
 237          }
 238     }
 239 }
 240
 241
 242 // parse i386/ppc ObjC category data structure
 243 void LTOModule::addObjCCategory(GlobalVariable* clgv)
 244 {
 245     if (ConstantStruct* c = dyn_cast<ConstantStruct>(clgv->getInitializer())) {
 246         // second slot in __OBJC,__category is pointer to target class name
 247         std::string targetclassName;
 248         if ( objcClassNameFromExpression(c->getOperand(1), targetclassName) ) {
 249             NameAndAttributes info;
 250             if ( _undefines.find(targetclassName.c_str()) == _undefines.end() ){
 251                 const char* symbolName = ::strdup(targetclassName.c_str());
 252                 info.name = ::strdup(symbolName);
 253                 info.attributes = LTO_SYMBOL_DEFINITION_UNDEFINED;
 254                 // string is owned by _undefines
 255                _undefines[info.name] = info;
 256             }
 257         }
 258     }
 259 }
 260
 261
 262 // parse i386/ppc ObjC class list data structure
 263 void LTOModule::addObjCClassRef(GlobalVariable* clgv)
 264 {
 265     std::string targetclassName;
 266     if ( objcClassNameFromExpression(clgv->getInitializer(), targetclassName) ){
 267         NameAndAttributes info;
 268         if ( _undefines.find(targetclassName.c_str()) == _undefines.end() ) {
 269             const char* symbolName = ::strdup(targetclassName.c_str());
 270             info.name = ::strdup(symbolName);
 271             info.attributes = LTO_SYMBOL_DEFINITION_UNDEFINED;
 272             // string is owned by _undefines
 273             _undefines[info.name] = info;
 274         }
 275     }
 276 }
 277
 278
 279 void LTOModule::addDefinedDataSymbol(GlobalValue* v, Mangler& mangler)
 280 {
 281     // add to list of defined symbols
 282     addDefinedSymbol(v, mangler, false);
 283
 284     // Special case i386/ppc ObjC data structures in magic sections:
 285     // The issue is that the old ObjC object format did some strange
 286     // contortions to avoid real linker symbols.  For instance, the
 287     // ObjC class data structure is allocated statically in the executable
 288     // that defines that class.  That data structures contains a pointer to
 289     // its superclass.  But instead of just initializing that part of the
 290     // struct to the address of its superclass, and letting the static and
 291     // dynamic linkers do the rest, the runtime works by having that field
 292     // instead point to a C-string that is the name of the superclass.
 293     // At runtime the objc initialization updates that pointer and sets
 294     // it to point to the actual super class.  As far as the linker
 295     // knows it is just a pointer to a string.  But then someone wanted the
 296     // linker to issue errors at build time if the superclass was not found.
 297     // So they figured out a way in mach-o object format to use an absolute
 298     // symbols (.objc_class_name_Foo = 0) and a floating reference
 299     // (.reference .objc_class_name_Bar) to cause the linker into erroring when
 300     // a class was missing.
 301     // The following synthesizes the implicit .objc_* symbols for the linker
 302     // from the ObjC data structures generated by the front end.
 303     if ( v->hasSection() /* && isTargetDarwin */ ) {
 304         // special case if this data blob is an ObjC class definition
 305         if ( v->getSection().compare(0, 15, "__OBJC,__class,") == 0 ) {
 306             if (GlobalVariable* gv = dyn_cast<GlobalVariable>(v)) {
 307                 addObjCClass(gv);
 308             }
 309         }
 310
 311         // special case if this data blob is an ObjC category definition
 312         else if ( v->getSection().compare(0, 18, "__OBJC,__category,") == 0 ) {
 313             if (GlobalVariable* gv = dyn_cast<GlobalVariable>(v)) {
 314                 addObjCCategory(gv);
 315             }
 316         }
 317
 318         // special case if this data blob is the list of referenced classes
 319         else if ( v->getSection().compare(0, 18, "__OBJC,__cls_refs,") == 0 ) {
 320             if (GlobalVariable* gv = dyn_cast<GlobalVariable>(v)) {
 321                 addObjCClassRef(gv);
 322             }
 323         }
 324     }
 325
 326     // add external symbols referenced by this data.
 327     for (unsigned count = 0, total = v->getNumOperands();
 328                                                 count != total; ++count) {
 329         findExternalRefs(v->getOperand(count), mangler);
 330     }
 331 }
 332
 333
 334 void LTOModule::addDefinedSymbol(GlobalValue* def, Mangler &mangler,
 335                                  bool isFunction)
 336 {
 337     // ignore all llvm.* symbols
 338     if (def->getName().startswith("llvm."))
 339         return;
 340
 341     // string is owned by _defines
 342     const char* symbolName = ::strdup(mangler.getMangledName(def).c_str());
 343
 344     // set alignment part log2() can have rounding errors
 345     uint32_t align = def->getAlignment();
 346     uint32_t attr = align ? CountTrailingZeros_32(def->getAlignment()) : 0;
 347
 348     // set permissions part
 349     if ( isFunction )
 350         attr |= LTO_SYMBOL_PERMISSIONS_CODE;
 351     else {
 352         GlobalVariable* gv = dyn_cast<GlobalVariable>(def);
 353         if ( (gv != NULL) && gv->isConstant() )
 354             attr |= LTO_SYMBOL_PERMISSIONS_RODATA;
 355         else
 356             attr |= LTO_SYMBOL_PERMISSIONS_DATA;
 357     }
 358
 359     // set definition part
 360     if ( def->hasWeakLinkage() || def->hasLinkOnceLinkage() ) {
 361         attr |= LTO_SYMBOL_DEFINITION_WEAK;
 362     }
 363     else if ( def->hasCommonLinkage()) {
 364         attr |= LTO_SYMBOL_DEFINITION_TENTATIVE;
 365     }
 366     else {
 367         attr |= LTO_SYMBOL_DEFINITION_REGULAR;
 368     }
 369
 370     // set scope part
 371     if ( def->hasHiddenVisibility() )
 372         attr |= LTO_SYMBOL_SCOPE_HIDDEN;
 373     else if ( def->hasProtectedVisibility() )
 374         attr |= LTO_SYMBOL_SCOPE_PROTECTED;
 375     else if ( def->hasExternalLinkage() || def->hasWeakLinkage()
 376               || def->hasLinkOnceLinkage() || def->hasCommonLinkage() )
 377         attr |= LTO_SYMBOL_SCOPE_DEFAULT;
 378     else
 379         attr |= LTO_SYMBOL_SCOPE_INTERNAL;
 380
 381     // add to table of symbols
 382     NameAndAttributes info;
 383     info.name = symbolName;
 384     info.attributes = (lto_symbol_attributes)attr;
 385     _symbols.push_back(info);
 386     _defines[info.name] = 1;
 387 }
 388
 389 void LTOModule::addAsmGlobalSymbol(const char *name) {
 390     // only add new define if not already defined
 391     if ( _defines.count(name) == 0 )
 392         return;
 393
 394     // string is owned by _defines
 395     const char *symbolName = ::strdup(name);
 396     uint32_t attr = LTO_SYMBOL_DEFINITION_REGULAR;
 397     attr |= LTO_SYMBOL_SCOPE_DEFAULT;
 398     NameAndAttributes info;
 399     info.name = symbolName;
 400     info.attributes = (lto_symbol_attributes)attr;
 401     _symbols.push_back(info);
 402     _defines[info.name] = 1;
 403 }
 404
 405 void LTOModule::addPotentialUndefinedSymbol(GlobalValue* decl, Mangler &mangler)
 406 {
 407     // ignore all llvm.* symbols
 408     if (decl->getName().startswith("llvm."))
 409         return;
 410
 411     // ignore all aliases
 412     if (isa<GlobalAlias>(decl))
 413         return;
 414
 415     std::string name = mangler.getMangledName(decl);
 416
 417     // we already have the symbol
 418     if (_undefines.find(name) != _undefines.end())
 419       return;
 420
 421     NameAndAttributes info;
 422     // string is owned by _undefines
 423     info.name = ::strdup(name.c_str());
 424     if (decl->hasExternalWeakLinkage())
 425       info.attributes = LTO_SYMBOL_DEFINITION_WEAKUNDEF;
 426     else
 427       info.attributes = LTO_SYMBOL_DEFINITION_UNDEFINED;
 428     _undefines[name] = info;
 429 }
 430
 431
 432
 433 // Find external symbols referenced by VALUE. This is a recursive function.
 434 void LTOModule::findExternalRefs(Value* value, Mangler &mangler) {
 435
 436     if (GlobalValue* gv = dyn_cast<GlobalValue>(value)) {
 437         if ( !gv->hasExternalLinkage() )
 438             addPotentialUndefinedSymbol(gv, mangler);
 439         // If this is a variable definition, do not recursively process
 440         // initializer.  It might contain a reference to this variable
 441         // and cause an infinite loop.  The initializer will be
 442         // processed in addDefinedDataSymbol().
 443         return;
 444     }
 445
 446     // GlobalValue, even with InternalLinkage type, may have operands with
 447     // ExternalLinkage type. Do not ignore these operands.
 448     if (Constant* c = dyn_cast<Constant>(value)) {
 449         // Handle ConstantExpr, ConstantStruct, ConstantArry etc..
 450         for (unsigned i = 0, e = c->getNumOperands(); i != e; ++i)
 451             findExternalRefs(c->getOperand(i), mangler);
 452     }
 453 }
 454
 455 void LTOModule::lazyParseSymbols()
 456 {
 457     if ( !_symbolsParsed ) {
 458         _symbolsParsed = true;
 459
 460         // Use mangler to add GlobalPrefix to names to match linker names.
 461         Mangler mangler(*_module, _target->getMCAsmInfo()->getGlobalPrefix());
 462         // add chars used in ObjC method names so method names aren't mangled
 463         mangler.markCharAcceptable('[');
 464         mangler.markCharAcceptable(']');
 465         mangler.markCharAcceptable('(');
 466         mangler.markCharAcceptable(')');
 467         mangler.markCharAcceptable('-');
 468         mangler.markCharAcceptable('+');
 469         mangler.markCharAcceptable(' ');
 470
 471         // add functions
 472         for (Module::iterator f = _module->begin(); f != _module->end(); ++f) {
 473             if ( f->isDeclaration() )
 474                 addPotentialUndefinedSymbol(f, mangler);
 475             else
 476                 addDefinedFunctionSymbol(f, mangler);
 477         }
 478
 479         // add data
 480         for (Module::global_iterator v = _module->global_begin(),
 481                                     e = _module->global_end(); v !=  e; ++v) {
 482             if ( v->isDeclaration() )
 483                 addPotentialUndefinedSymbol(v, mangler);
 484             else
 485                 addDefinedDataSymbol(v, mangler);
 486         }
 487
 488         // add asm globals
 489         const std::string &inlineAsm = _module->getModuleInlineAsm();
 490         const std::string glbl = ".globl";
 491         std::string asmSymbolName;
 492         std::string::size_type pos = inlineAsm.find(glbl, 0);
 493         while (pos != std::string::npos) {
 494           // eat .globl
 495           pos = pos + 6;
 496
 497           // skip white space between .globl and symbol name
 498           std::string::size_type pbegin = inlineAsm.find_first_not_of(' ', pos);
 499           if (pbegin == std::string::npos)
 500             break;
 501
 502           // find end-of-line
 503           std::string::size_type pend = inlineAsm.find_first_of('\n', pbegin);
 504           if (pend == std::string::npos)
 505             break;
 506
 507           asmSymbolName.assign(inlineAsm, pbegin, pend - pbegin);
 508           addAsmGlobalSymbol(asmSymbolName.c_str());
 509
 510           // search next .globl
 511           pos = inlineAsm.find(glbl, pend);
 512         }
 513
 514         // make symbols for all undefines
 515         for (StringMap<NameAndAttributes>::iterator it=_undefines.begin();
 516                                                 it != _undefines.end(); ++it) {
 517             // if this symbol also has a definition, then don't make an undefine
 518             // because it is a tentative definition
 519             if ( _defines.count(it->getKey()) == 0 ) {
 520               NameAndAttributes info = it->getValue();
 521               _symbols.push_back(info);
 522             }
 523         }
 524     }
 525 }
 526
 527
 528 uint32_t LTOModule::getSymbolCount()
 529 {
 530     lazyParseSymbols();
 531     return _symbols.size();
 532 }
 533
 534
 535 lto_symbol_attributes LTOModule::getSymbolAttributes(uint32_t index)
 536 {
 537     lazyParseSymbols();
 538     if ( index < _symbols.size() )
 539         return _symbols[index].attributes;
 540     else
 541         return lto_symbol_attributes(0);
 542 }
 543
 544 const char* LTOModule::getSymbolName(uint32_t index)
 545 {
 546     lazyParseSymbols();
 547     if ( index < _symbols.size() )
 548         return _symbols[index].name;
 549     else
 550         return NULL;
 551 }