llvm/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 *, 4> &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   // Symbols in the llvm.used list will get the FB_Used bit and will not be
 123   // internalized. We do this for llvm.compiler.used as well:
 124   //
 125   // IR symbol table tracks module-level asm symbol references but not inline
 126   // asm. A symbol only referenced by inline asm is not in the IR symbol table,
 127   // so we may not know that the definition (in another translation unit) is
 128   // referenced. That definition may have __attribute__((used)) (which lowers to
 129   // llvm.compiler.used on ELF targets) to communicate to the compiler that it
 130   // may be used by inline asm. The usage is perfectly fine, so we treat
 131   // llvm.compiler.used conservatively as llvm.used to work around our own
 132   // limitation.
 133   SmallVector<GlobalValue *, 4> UsedV;
 134   collectUsedGlobalVariables(*M, UsedV, /*CompilerUsed=*/false);
 135   collectUsedGlobalVariables(*M, UsedV, /*CompilerUsed=*/true);
 136   SmallPtrSet<GlobalValue *, 4> Used(UsedV.begin(), UsedV.end());
 137
 138   ModuleSymbolTable Msymtab;
 139   Msymtab.addModule(M);
 140
 141   storage::Module Mod;
 142   Mod.Begin = Syms.size();
 143   Mod.End = Syms.size() + Msymtab.symbols().size();
 144   Mod.UncBegin = Uncommons.size();
 145   Mods.push_back(Mod);
 146
 147   if (TT.isOSBinFormatCOFF()) {
 148     if (auto E = M->materializeMetadata())
 149       return E;
 150     if (NamedMDNode *LinkerOptions =
 151             M->getNamedMetadata("llvm.linker.options")) {
 152       for (MDNode *MDOptions : LinkerOptions->operands())
 153         for (const MDOperand &MDOption : cast<MDNode>(MDOptions)->operands())
 154           COFFLinkerOptsOS << " " << cast<MDString>(MDOption)->getString();
 155     }
 156   }
 157
 158   if (TT.isOSBinFormatELF()) {
 159     if (auto E = M->materializeMetadata())
 160       return E;
 161     if (NamedMDNode *N = M->getNamedMetadata("llvm.dependent-libraries")) {
 162       for (MDNode *MDOptions : N->operands()) {
 163         const auto OperandStr =
 164             cast<MDString>(cast<MDNode>(MDOptions)->getOperand(0))->getString();
 165         storage::Str Specifier;
 166         setStr(Specifier, OperandStr);
 167         DependentLibraries.emplace_back(Specifier);
 168       }
 169     }
 170   }
 171
 172   for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
 173     if (Error Err = addSymbol(Msymtab, Used, Msym))
 174       return Err;
 175
 176   return Error::success();
 177 }
 178
 179 Expected<int> Builder::getComdatIndex(const Comdat *C, const Module *M) {
 180   auto P = ComdatMap.insert(std::make_pair(C, Comdats.size()));
 181   if (P.second) {
 182     std::string Name;
 183     if (TT.isOSBinFormatCOFF()) {
 184       const GlobalValue *GV = M->getNamedValue(C->getName());
 185       if (!GV)
 186         return make_error<StringError>("Could not find leader",
 187                                        inconvertibleErrorCode());
 188       // Internal leaders do not affect symbol resolution, therefore they do not
 189       // appear in the symbol table.
 190       if (GV->hasLocalLinkage()) {
 191         P.first->second = -1;
 192         return -1;
 193       }
 194       llvm::raw_string_ostream OS(Name);
 195       Mang.getNameWithPrefix(OS, GV, false);
 196     } else {
 197       Name = std::string(C->getName());
 198     }
 199
 200     storage::Comdat Comdat;
 201     setStr(Comdat.Name, Saver.save(Name));
 202     Comdat.SelectionKind = C->getSelectionKind();
 203     Comdats.push_back(Comdat);
 204   }
 205
 206   return P.first->second;
 207 }
 208
 209 Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
 210                          const SmallPtrSet<GlobalValue *, 4> &Used,
 211                          ModuleSymbolTable::Symbol Msym) {
 212   Syms.emplace_back();
 213   storage::Symbol &Sym = Syms.back();
 214   Sym = {};
 215
 216   storage::Uncommon *Unc = nullptr;
 217   auto Uncommon = [&]() -> storage::Uncommon & {
 218     if (Unc)
 219       return *Unc;
 220     Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
 221     Uncommons.emplace_back();
 222     Unc = &Uncommons.back();
 223     *Unc = {};
 224     setStr(Unc->COFFWeakExternFallbackName, "");
 225     setStr(Unc->SectionName, "");
 226     return *Unc;
 227   };
 228
 229   SmallString<64> Name;
 230   {
 231     raw_svector_ostream OS(Name);
 232     Msymtab.printSymbolName(OS, Msym);
 233   }
 234   setStr(Sym.Name, Saver.save(Name.str()));
 235
 236   auto Flags = Msymtab.getSymbolFlags(Msym);
 237   if (Flags & object::BasicSymbolRef::SF_Undefined)
 238     Sym.Flags |= 1 << storage::Symbol::FB_undefined;
 239   if (Flags & object::BasicSymbolRef::SF_Weak)
 240     Sym.Flags |= 1 << storage::Symbol::FB_weak;
 241   if (Flags & object::BasicSymbolRef::SF_Common)
 242     Sym.Flags |= 1 << storage::Symbol::FB_common;
 243   if (Flags & object::BasicSymbolRef::SF_Indirect)
 244     Sym.Flags |= 1 << storage::Symbol::FB_indirect;
 245   if (Flags & object::BasicSymbolRef::SF_Global)
 246     Sym.Flags |= 1 << storage::Symbol::FB_global;
 247   if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
 248     Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
 249   if (Flags & object::BasicSymbolRef::SF_Executable)
 250     Sym.Flags |= 1 << storage::Symbol::FB_executable;
 251
 252   Sym.ComdatIndex = -1;
 253   auto *GV = Msym.dyn_cast<GlobalValue *>();
 254   if (!GV) {
 255     // Undefined module asm symbols act as GC roots and are implicitly used.
 256     if (Flags & object::BasicSymbolRef::SF_Undefined)
 257       Sym.Flags |= 1 << storage::Symbol::FB_used;
 258     setStr(Sym.IRName, "");
 259     return Error::success();
 260   }
 261
 262   setStr(Sym.IRName, GV->getName());
 263
 264   bool IsBuiltinFunc = llvm::is_contained(LibcallRoutineNames, GV->getName());
 265
 266   if (Used.count(GV) || IsBuiltinFunc)
 267     Sym.Flags |= 1 << storage::Symbol::FB_used;
 268   if (GV->isThreadLocal())
 269     Sym.Flags |= 1 << storage::Symbol::FB_tls;
 270   if (GV->hasGlobalUnnamedAddr())
 271     Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
 272   if (GV->canBeOmittedFromSymbolTable())
 273     Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
 274   Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
 275
 276   if (Flags & object::BasicSymbolRef::SF_Common) {
 277     auto *GVar = dyn_cast<GlobalVariable>(GV);
 278     if (!GVar)
 279       return make_error<StringError>("Only variables can have common linkage!",
 280                                      inconvertibleErrorCode());
 281     Uncommon().CommonSize =
 282         GV->getParent()->getDataLayout().getTypeAllocSize(GV->getValueType());
 283     Uncommon().CommonAlign = GVar->getAlignment();
 284   }
 285
 286   const GlobalObject *Base = GV->getBaseObject();
 287   if (!Base)
 288     return make_error<StringError>("Unable to determine comdat of alias!",
 289                                    inconvertibleErrorCode());
 290   if (const Comdat *C = Base->getComdat()) {
 291     Expected<int> ComdatIndexOrErr = getComdatIndex(C, GV->getParent());
 292     if (!ComdatIndexOrErr)
 293       return ComdatIndexOrErr.takeError();
 294     Sym.ComdatIndex = *ComdatIndexOrErr;
 295   }
 296
 297   if (TT.isOSBinFormatCOFF()) {
 298     emitLinkerFlagsForGlobalCOFF(COFFLinkerOptsOS, GV, TT, Mang);
 299
 300     if ((Flags & object::BasicSymbolRef::SF_Weak) &&
 301         (Flags & object::BasicSymbolRef::SF_Indirect)) {
 302       auto *Fallback = dyn_cast<GlobalValue>(
 303           cast<GlobalAlias>(GV)->getAliasee()->stripPointerCasts());
 304       if (!Fallback)
 305         return make_error<StringError>("Invalid weak external",
 306                                        inconvertibleErrorCode());
 307       std::string FallbackName;
 308       raw_string_ostream OS(FallbackName);
 309       Msymtab.printSymbolName(OS, Fallback);
 310       OS.flush();
 311       setStr(Uncommon().COFFWeakExternFallbackName, Saver.save(FallbackName));
 312     }
 313   }
 314
 315   if (!Base->getSection().empty())
 316     setStr(Uncommon().SectionName, Saver.save(Base->getSection()));
 317
 318   return Error::success();
 319 }
 320
 321 Error Builder::build(ArrayRef<Module *> IRMods) {
 322   storage::Header Hdr;
 323
 324   assert(!IRMods.empty());
 325   Hdr.Version = storage::Header::kCurrentVersion;
 326   setStr(Hdr.Producer, kExpectedProducerName);
 327   setStr(Hdr.TargetTriple, IRMods[0]->getTargetTriple());
 328   setStr(Hdr.SourceFileName, IRMods[0]->getSourceFileName());
 329   TT = Triple(IRMods[0]->getTargetTriple());
 330
 331   for (auto *M : IRMods)
 332     if (Error Err = addModule(M))
 333       return Err;
 334
 335   COFFLinkerOptsOS.flush();
 336   setStr(Hdr.COFFLinkerOpts, Saver.save(COFFLinkerOpts));
 337
 338   // We are about to fill in the header's range fields, so reserve space for it
 339   // and copy it in afterwards.
 340   Symtab.resize(sizeof(storage::Header));
 341   writeRange(Hdr.Modules, Mods);
 342   writeRange(Hdr.Comdats, Comdats);
 343   writeRange(Hdr.Symbols, Syms);
 344   writeRange(Hdr.Uncommons, Uncommons);
 345   writeRange(Hdr.DependentLibraries, DependentLibraries);
 346   *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
 347   return Error::success();
 348 }
 349
 350 } // end anonymous namespace
 351
 352 Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
 353                       StringTableBuilder &StrtabBuilder,
 354                       BumpPtrAllocator &Alloc) {
 355   return Builder(Symtab, StrtabBuilder, Alloc).build(Mods);
 356 }
 357
 358 // Upgrade a vector of bitcode modules created by an old version of LLVM by
 359 // creating an irsymtab for them in the current format.
 360 static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
 361   FileContents FC;
 362
 363   LLVMContext Ctx;
 364   std::vector<Module *> Mods;
 365   std::vector<std::unique_ptr<Module>> OwnedMods;
 366   for (auto BM : BMs) {
 367     Expected<std::unique_ptr<Module>> MOrErr =
 368         BM.getLazyModule(Ctx, /*ShouldLazyLoadMetadata*/ true,
 369                          /*IsImporting*/ false);
 370     if (!MOrErr)
 371       return MOrErr.takeError();
 372
 373     Mods.push_back(MOrErr->get());
 374     OwnedMods.push_back(std::move(*MOrErr));
 375   }
 376
 377   StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
 378   BumpPtrAllocator Alloc;
 379   if (Error E = build(Mods, FC.Symtab, StrtabBuilder, Alloc))
 380     return std::move(E);
 381
 382   StrtabBuilder.finalizeInOrder();
 383   FC.Strtab.resize(StrtabBuilder.getSize());
 384   StrtabBuilder.write((uint8_t *)FC.Strtab.data());
 385
 386   FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
 387                   {FC.Strtab.data(), FC.Strtab.size()}};
 388   return std::move(FC);
 389 }
 390
 391 Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
 392   if (BFC.Mods.empty())
 393     return make_error<StringError>("Bitcode file does not contain any modules",
 394                                    inconvertibleErrorCode());
 395
 396   if (BFC.StrtabForSymtab.empty() ||
 397       BFC.Symtab.size() < sizeof(storage::Header))
 398     return upgrade(BFC.Mods);
 399
 400   // We cannot use the regular reader to read the version and producer, because
 401   // it will expect the header to be in the current format. The only thing we
 402   // can rely on is that the version and producer will be present as the first
 403   // struct elements.
 404   auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
 405   unsigned Version = Hdr->Version;
 406   StringRef Producer = Hdr->Producer.get(BFC.StrtabForSymtab);
 407   if (Version != storage::Header::kCurrentVersion ||
 408       Producer != kExpectedProducerName)
 409     return upgrade(BFC.Mods);
 410
 411   FileContents FC;
 412   FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
 413                   {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
 414
 415   // Finally, make sure that the number of modules in the symbol table matches
 416   // the number of modules in the bitcode file. If they differ, it may mean that
 417   // the bitcode file was created by binary concatenation, so we need to create
 418   // a new symbol table from scratch.
 419   if (FC.TheReader.getNumModules() != BFC.Mods.size())
 420     return upgrade(std::move(BFC.Mods));
 421
 422   return std::move(FC);
 423 }