lib/Object/IRSymtab.cpp

   1 //===- IRSymtab.cpp - implementation of IR symbol tables ------------------===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8
   9 #include "llvm/Object/IRSymtab.h"
  10 #include "llvm/ADT/ArrayRef.h"
  11 #include "llvm/ADT/DenseMap.h"
  12 #include "llvm/ADT/SmallPtrSet.h"
  13 #include "llvm/ADT/SmallString.h"
  14 #include "llvm/ADT/SmallVector.h"
  15 #include "llvm/ADT/StringRef.h"
  16 #include "llvm/ADT/Triple.h"
  17 #include "llvm/Config/llvm-config.h"
  18 #include "llvm/IR/Comdat.h"
  19 #include "llvm/IR/DataLayout.h"
  20 #include "llvm/IR/GlobalAlias.h"
  21 #include "llvm/IR/GlobalObject.h"
  22 #include "llvm/IR/Mangler.h"
  23 #include "llvm/IR/Metadata.h"
  24 #include "llvm/IR/Module.h"
  25 #include "llvm/Bitcode/BitcodeReader.h"
  26 #include "llvm/MC/StringTableBuilder.h"
  27 #include "llvm/Object/IRObjectFile.h"
  28 #include "llvm/Object/ModuleSymbolTable.h"
  29 #include "llvm/Object/SymbolicFile.h"
  30 #include "llvm/Support/Allocator.h"
  31 #include "llvm/Support/Casting.h"
  32 #include "llvm/Support/Error.h"
  33 #include "llvm/Support/StringSaver.h"
  34 #include "llvm/Support/VCSRevision.h"
  35 #include "llvm/Support/raw_ostream.h"
  36 #include <cassert>
  37 #include <string>
  38 #include <utility>
  39 #include <vector>
  40
  41 using namespace llvm;
  42 using namespace irsymtab;
  43
  44 static const char *LibcallRoutineNames[] = {
  45 #define HANDLE_LIBCALL(code, name) name,
  46 #include "llvm/IR/RuntimeLibcalls.def"
  47 #undef HANDLE_LIBCALL
  48 };
  49
  50 namespace {
  51
  52 const char *getExpectedProducerName() {
  53   static char DefaultName[] = LLVM_VERSION_STRING
  54 #ifdef LLVM_REVISION
  55       " " LLVM_REVISION
  56 #endif
  57       ;
  58   // Allows for testing of the irsymtab writer and upgrade mechanism. This
  59   // environment variable should not be set by users.
  60   if (char *OverrideName = getenv("LLVM_OVERRIDE_PRODUCER"))
  61     return OverrideName;
  62   return DefaultName;
  63 }
  64
  65 const char *kExpectedProducerName = getExpectedProducerName();
  66
  67 /// Stores the temporary state that is required to build an IR symbol table.
  68 struct Builder {
  69   SmallVector<char, 0> &Symtab;
  70   StringTableBuilder &StrtabBuilder;
  71   StringSaver Saver;
  72
  73   // This ctor initializes a StringSaver using the passed in BumpPtrAllocator.
  74   // The StringTableBuilder does not create a copy of any strings added to it,
  75   // so this provides somewhere to store any strings that we create.
  76   Builder(SmallVector<char, 0> &Symtab, StringTableBuilder &StrtabBuilder,
  77           BumpPtrAllocator &Alloc)
  78       : Symtab(Symtab), StrtabBuilder(StrtabBuilder), Saver(Alloc) {}
  79
  80   DenseMap<const Comdat *, int> ComdatMap;
  81   Mangler Mang;
  82   Triple TT;
  83
  84   std::vector<storage::Comdat> Comdats;
  85   std::vector<storage::Module> Mods;
  86   std::vector<storage::Symbol> Syms;
  87   std::vector<storage::Uncommon> Uncommons;
  88
  89   std::string COFFLinkerOpts;
  90   raw_string_ostream COFFLinkerOptsOS{COFFLinkerOpts};
  91
  92   std::vector<storage::Str> DependentLibraries;
  93
  94   void setStr(storage::Str &S, StringRef Value) {
  95     S.Offset = StrtabBuilder.add(Value);
  96     S.Size = Value.size();
  97   }
  98
  99   template <typename T>
 100   void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
 101     R.Offset = Symtab.size();
 102     R.Size = Objs.size();
 103     Symtab.insert(Symtab.end(), reinterpret_cast<const char *>(Objs.data()),
 104                   reinterpret_cast<const char *>(Objs.data() + Objs.size()));
 105   }
 106
 107   Expected<int> getComdatIndex(const Comdat *C, const Module *M);
 108
 109   Error addModule(Module *M);
 110   Error addSymbol(const ModuleSymbolTable &Msymtab,
 111                   const SmallPtrSet<GlobalValue *, 8> &Used,
 112                   ModuleSymbolTable::Symbol Sym);
 113
 114   Error build(ArrayRef<Module *> Mods);
 115 };
 116
 117 Error Builder::addModule(Module *M) {
 118   if (M->getDataLayoutStr().empty())
 119     return make_error<StringError>("input module has no datalayout",
 120                                    inconvertibleErrorCode());
 121
 122   SmallPtrSet<GlobalValue *, 8> Used;
 123   collectUsedGlobalVariables(*M, Used, /*CompilerUsed*/ false);
 124
 125   ModuleSymbolTable Msymtab;
 126   Msymtab.addModule(M);
 127
 128   storage::Module Mod;
 129   Mod.Begin = Syms.size();
 130   Mod.End = Syms.size() + Msymtab.symbols().size();
 131   Mod.UncBegin = Uncommons.size();
 132   Mods.push_back(Mod);
 133
 134   if (TT.isOSBinFormatCOFF()) {
 135     if (auto E = M->materializeMetadata())
 136       return E;
 137     if (NamedMDNode *LinkerOptions =
 138             M->getNamedMetadata("llvm.linker.options")) {
 139       for (MDNode *MDOptions : LinkerOptions->operands())
 140         for (const MDOperand &MDOption : cast<MDNode>(MDOptions)->operands())
 141           COFFLinkerOptsOS << " " << cast<MDString>(MDOption)->getString();
 142     }
 143   }
 144
 145   if (TT.isOSBinFormatELF()) {
 146     if (auto E = M->materializeMetadata())
 147       return E;
 148     if (NamedMDNode *N = M->getNamedMetadata("llvm.dependent-libraries")) {
 149       for (MDNode *MDOptions : N->operands()) {
 150         const auto OperandStr =
 151             cast<MDString>(cast<MDNode>(MDOptions)->getOperand(0))->getString();
 152         storage::Str Specifier;
 153         setStr(Specifier, OperandStr);
 154         DependentLibraries.emplace_back(Specifier);
 155       }
 156     }
 157   }
 158
 159   for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
 160     if (Error Err = addSymbol(Msymtab, Used, Msym))
 161       return Err;
 162
 163   return Error::success();
 164 }
 165
 166 Expected<int> Builder::getComdatIndex(const Comdat *C, const Module *M) {
 167   auto P = ComdatMap.insert(std::make_pair(C, Comdats.size()));
 168   if (P.second) {
 169     std::string Name;
 170     if (TT.isOSBinFormatCOFF()) {
 171       const GlobalValue *GV = M->getNamedValue(C->getName());
 172       if (!GV)
 173         return make_error<StringError>("Could not find leader",
 174                                        inconvertibleErrorCode());
 175       // Internal leaders do not affect symbol resolution, therefore they do not
 176       // appear in the symbol table.
 177       if (GV->hasLocalLinkage()) {
 178         P.first->second = -1;
 179         return -1;
 180       }
 181       llvm::raw_string_ostream OS(Name);
 182       Mang.getNameWithPrefix(OS, GV, false);
 183     } else {
 184       Name = C->getName();
 185     }
 186
 187     storage::Comdat Comdat;
 188     setStr(Comdat.Name, Saver.save(Name));
 189     Comdats.push_back(Comdat);
 190   }
 191
 192   return P.first->second;
 193 }
 194
 195 Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
 196                          const SmallPtrSet<GlobalValue *, 8> &Used,
 197                          ModuleSymbolTable::Symbol Msym) {
 198   Syms.emplace_back();
 199   storage::Symbol &Sym = Syms.back();
 200   Sym = {};
 201
 202   storage::Uncommon *Unc = nullptr;
 203   auto Uncommon = [&]() -> storage::Uncommon & {
 204     if (Unc)
 205       return *Unc;
 206     Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
 207     Uncommons.emplace_back();
 208     Unc = &Uncommons.back();
 209     *Unc = {};
 210     setStr(Unc->COFFWeakExternFallbackName, "");
 211     setStr(Unc->SectionName, "");
 212     return *Unc;
 213   };
 214
 215   SmallString<64> Name;
 216   {
 217     raw_svector_ostream OS(Name);
 218     Msymtab.printSymbolName(OS, Msym);
 219   }
 220   setStr(Sym.Name, Saver.save(StringRef(Name)));
 221
 222   auto Flags = Msymtab.getSymbolFlags(Msym);
 223   if (Flags & object::BasicSymbolRef::SF_Undefined)
 224     Sym.Flags |= 1 << storage::Symbol::FB_undefined;
 225   if (Flags & object::BasicSymbolRef::SF_Weak)
 226     Sym.Flags |= 1 << storage::Symbol::FB_weak;
 227   if (Flags & object::BasicSymbolRef::SF_Common)
 228     Sym.Flags |= 1 << storage::Symbol::FB_common;
 229   if (Flags & object::BasicSymbolRef::SF_Indirect)
 230     Sym.Flags |= 1 << storage::Symbol::FB_indirect;
 231   if (Flags & object::BasicSymbolRef::SF_Global)
 232     Sym.Flags |= 1 << storage::Symbol::FB_global;
 233   if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
 234     Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
 235   if (Flags & object::BasicSymbolRef::SF_Executable)
 236     Sym.Flags |= 1 << storage::Symbol::FB_executable;
 237
 238   Sym.ComdatIndex = -1;
 239   auto *GV = Msym.dyn_cast<GlobalValue *>();
 240   if (!GV) {
 241     // Undefined module asm symbols act as GC roots and are implicitly used.
 242     if (Flags & object::BasicSymbolRef::SF_Undefined)
 243       Sym.Flags |= 1 << storage::Symbol::FB_used;
 244     setStr(Sym.IRName, "");
 245     return Error::success();
 246   }
 247
 248   setStr(Sym.IRName, GV->getName());
 249
 250   bool IsBuiltinFunc = false;
 251
 252   for (const char *LibcallName : LibcallRoutineNames)
 253     if (GV->getName() == LibcallName)
 254       IsBuiltinFunc = true;
 255
 256   if (Used.count(GV) || IsBuiltinFunc)
 257     Sym.Flags |= 1 << storage::Symbol::FB_used;
 258   if (GV->isThreadLocal())
 259     Sym.Flags |= 1 << storage::Symbol::FB_tls;
 260   if (GV->hasGlobalUnnamedAddr())
 261     Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
 262   if (GV->canBeOmittedFromSymbolTable())
 263     Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
 264   Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
 265
 266   if (Flags & object::BasicSymbolRef::SF_Common) {
 267     Uncommon().CommonSize = GV->getParent()->getDataLayout().getTypeAllocSize(
 268         GV->getType()->getElementType());
 269     Uncommon().CommonAlign = GV->getAlignment();
 270   }
 271
 272   const GlobalObject *Base = GV->getBaseObject();
 273   if (!Base)
 274     return make_error<StringError>("Unable to determine comdat of alias!",
 275                                    inconvertibleErrorCode());
 276   if (const Comdat *C = Base->getComdat()) {
 277     Expected<int> ComdatIndexOrErr = getComdatIndex(C, GV->getParent());
 278     if (!ComdatIndexOrErr)
 279       return ComdatIndexOrErr.takeError();
 280     Sym.ComdatIndex = *ComdatIndexOrErr;
 281   }
 282
 283   if (TT.isOSBinFormatCOFF()) {
 284     emitLinkerFlagsForGlobalCOFF(COFFLinkerOptsOS, GV, TT, Mang);
 285
 286     if ((Flags & object::BasicSymbolRef::SF_Weak) &&
 287         (Flags & object::BasicSymbolRef::SF_Indirect)) {
 288       auto *Fallback = dyn_cast<GlobalValue>(
 289           cast<GlobalAlias>(GV)->getAliasee()->stripPointerCasts());
 290       if (!Fallback)
 291         return make_error<StringError>("Invalid weak external",
 292                                        inconvertibleErrorCode());
 293       std::string FallbackName;
 294       raw_string_ostream OS(FallbackName);
 295       Msymtab.printSymbolName(OS, Fallback);
 296       OS.flush();
 297       setStr(Uncommon().COFFWeakExternFallbackName, Saver.save(FallbackName));
 298     }
 299   }
 300
 301   if (!Base->getSection().empty())
 302     setStr(Uncommon().SectionName, Saver.save(Base->getSection()));
 303
 304   return Error::success();
 305 }
 306
 307 Error Builder::build(ArrayRef<Module *> IRMods) {
 308   storage::Header Hdr;
 309
 310   assert(!IRMods.empty());
 311   Hdr.Version = storage::Header::kCurrentVersion;
 312   setStr(Hdr.Producer, kExpectedProducerName);
 313   setStr(Hdr.TargetTriple, IRMods[0]->getTargetTriple());
 314   setStr(Hdr.SourceFileName, IRMods[0]->getSourceFileName());
 315   TT = Triple(IRMods[0]->getTargetTriple());
 316
 317   for (auto *M : IRMods)
 318     if (Error Err = addModule(M))
 319       return Err;
 320
 321   COFFLinkerOptsOS.flush();
 322   setStr(Hdr.COFFLinkerOpts, Saver.save(COFFLinkerOpts));
 323
 324   // We are about to fill in the header's range fields, so reserve space for it
 325   // and copy it in afterwards.
 326   Symtab.resize(sizeof(storage::Header));
 327   writeRange(Hdr.Modules, Mods);
 328   writeRange(Hdr.Comdats, Comdats);
 329   writeRange(Hdr.Symbols, Syms);
 330   writeRange(Hdr.Uncommons, Uncommons);
 331   writeRange(Hdr.DependentLibraries, DependentLibraries);
 332   *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
 333   return Error::success();
 334 }
 335
 336 } // end anonymous namespace
 337
 338 Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
 339                       StringTableBuilder &StrtabBuilder,
 340                       BumpPtrAllocator &Alloc) {
 341   return Builder(Symtab, StrtabBuilder, Alloc).build(Mods);
 342 }
 343
 344 // Upgrade a vector of bitcode modules created by an old version of LLVM by
 345 // creating an irsymtab for them in the current format.
 346 static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
 347   FileContents FC;
 348
 349   LLVMContext Ctx;
 350   std::vector<Module *> Mods;
 351   std::vector<std::unique_ptr<Module>> OwnedMods;
 352   for (auto BM : BMs) {
 353     Expected<std::unique_ptr<Module>> MOrErr =
 354         BM.getLazyModule(Ctx, /*ShouldLazyLoadMetadata*/ true,
 355                          /*IsImporting*/ false);
 356     if (!MOrErr)
 357       return MOrErr.takeError();
 358
 359     Mods.push_back(MOrErr->get());
 360     OwnedMods.push_back(std::move(*MOrErr));
 361   }
 362
 363   StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
 364   BumpPtrAllocator Alloc;
 365   if (Error E = build(Mods, FC.Symtab, StrtabBuilder, Alloc))
 366     return std::move(E);
 367
 368   StrtabBuilder.finalizeInOrder();
 369   FC.Strtab.resize(StrtabBuilder.getSize());
 370   StrtabBuilder.write((uint8_t *)FC.Strtab.data());
 371
 372   FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
 373                   {FC.Strtab.data(), FC.Strtab.size()}};
 374   return std::move(FC);
 375 }
 376
 377 Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
 378   if (BFC.Mods.empty())
 379     return make_error<StringError>("Bitcode file does not contain any modules",
 380                                    inconvertibleErrorCode());
 381
 382   if (BFC.StrtabForSymtab.empty() ||
 383       BFC.Symtab.size() < sizeof(storage::Header))
 384     return upgrade(BFC.Mods);
 385
 386   // We cannot use the regular reader to read the version and producer, because
 387   // it will expect the header to be in the current format. The only thing we
 388   // can rely on is that the version and producer will be present as the first
 389   // struct elements.
 390   auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
 391   unsigned Version = Hdr->Version;
 392   StringRef Producer = Hdr->Producer.get(BFC.StrtabForSymtab);
 393   if (Version != storage::Header::kCurrentVersion ||
 394       Producer != kExpectedProducerName)
 395     return upgrade(BFC.Mods);
 396
 397   FileContents FC;
 398   FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
 399                   {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
 400
 401   // Finally, make sure that the number of modules in the symbol table matches
 402   // the number of modules in the bitcode file. If they differ, it may mean that
 403   // the bitcode file was created by binary concatenation, so we need to create
 404   // a new symbol table from scratch.
 405   if (FC.TheReader.getNumModules() != BFC.Mods.size())
 406     return upgrade(std::move(BFC.Mods));
 407
 408   return std::move(FC);
 409 }