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   void setStr(storage::Str &S, StringRef Value) {
  93     S.Offset = StrtabBuilder.add(Value);
  94     S.Size = Value.size();
  95   }
  96
  97   template <typename T>
  98   void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
  99     R.Offset = Symtab.size();
 100     R.Size = Objs.size();
 101     Symtab.insert(Symtab.end(), reinterpret_cast<const char *>(Objs.data()),
 102                   reinterpret_cast<const char *>(Objs.data() + Objs.size()));
 103   }
 104
 105   Expected<int> getComdatIndex(const Comdat *C, const Module *M);
 106
 107   Error addModule(Module *M);
 108   Error addSymbol(const ModuleSymbolTable &Msymtab,
 109                   const SmallPtrSet<GlobalValue *, 8> &Used,
 110                   ModuleSymbolTable::Symbol Sym);
 111
 112   Error build(ArrayRef<Module *> Mods);
 113 };
 114
 115 Error Builder::addModule(Module *M) {
 116   if (M->getDataLayoutStr().empty())
 117     return make_error<StringError>("input module has no datalayout",
 118                                    inconvertibleErrorCode());
 119
 120   SmallPtrSet<GlobalValue *, 8> Used;
 121   collectUsedGlobalVariables(*M, Used, /*CompilerUsed*/ false);
 122
 123   ModuleSymbolTable Msymtab;
 124   Msymtab.addModule(M);
 125
 126   storage::Module Mod;
 127   Mod.Begin = Syms.size();
 128   Mod.End = Syms.size() + Msymtab.symbols().size();
 129   Mod.UncBegin = Uncommons.size();
 130   Mods.push_back(Mod);
 131
 132   if (TT.isOSBinFormatCOFF()) {
 133     if (auto E = M->materializeMetadata())
 134       return E;
 135     if (NamedMDNode *LinkerOptions =
 136             M->getNamedMetadata("llvm.linker.options")) {
 137       for (MDNode *MDOptions : LinkerOptions->operands())
 138         for (const MDOperand &MDOption : cast<MDNode>(MDOptions)->operands())
 139           COFFLinkerOptsOS << " " << cast<MDString>(MDOption)->getString();
 140     }
 141   }
 142
 143   for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
 144     if (Error Err = addSymbol(Msymtab, Used, Msym))
 145       return Err;
 146
 147   return Error::success();
 148 }
 149
 150 Expected<int> Builder::getComdatIndex(const Comdat *C, const Module *M) {
 151   auto P = ComdatMap.insert(std::make_pair(C, Comdats.size()));
 152   if (P.second) {
 153     std::string Name;
 154     if (TT.isOSBinFormatCOFF()) {
 155       const GlobalValue *GV = M->getNamedValue(C->getName());
 156       if (!GV)
 157         return make_error<StringError>("Could not find leader",
 158                                        inconvertibleErrorCode());
 159       // Internal leaders do not affect symbol resolution, therefore they do not
 160       // appear in the symbol table.
 161       if (GV->hasLocalLinkage()) {
 162         P.first->second = -1;
 163         return -1;
 164       }
 165       llvm::raw_string_ostream OS(Name);
 166       Mang.getNameWithPrefix(OS, GV, false);
 167     } else {
 168       Name = C->getName();
 169     }
 170
 171     storage::Comdat Comdat;
 172     setStr(Comdat.Name, Saver.save(Name));
 173     Comdats.push_back(Comdat);
 174   }
 175
 176   return P.first->second;
 177 }
 178
 179 Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
 180                          const SmallPtrSet<GlobalValue *, 8> &Used,
 181                          ModuleSymbolTable::Symbol Msym) {
 182   Syms.emplace_back();
 183   storage::Symbol &Sym = Syms.back();
 184   Sym = {};
 185
 186   storage::Uncommon *Unc = nullptr;
 187   auto Uncommon = [&]() -> storage::Uncommon & {
 188     if (Unc)
 189       return *Unc;
 190     Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
 191     Uncommons.emplace_back();
 192     Unc = &Uncommons.back();
 193     *Unc = {};
 194     setStr(Unc->COFFWeakExternFallbackName, "");
 195     setStr(Unc->SectionName, "");
 196     return *Unc;
 197   };
 198
 199   SmallString<64> Name;
 200   {
 201     raw_svector_ostream OS(Name);
 202     Msymtab.printSymbolName(OS, Msym);
 203   }
 204   setStr(Sym.Name, Saver.save(StringRef(Name)));
 205
 206   auto Flags = Msymtab.getSymbolFlags(Msym);
 207   if (Flags & object::BasicSymbolRef::SF_Undefined)
 208     Sym.Flags |= 1 << storage::Symbol::FB_undefined;
 209   if (Flags & object::BasicSymbolRef::SF_Weak)
 210     Sym.Flags |= 1 << storage::Symbol::FB_weak;
 211   if (Flags & object::BasicSymbolRef::SF_Common)
 212     Sym.Flags |= 1 << storage::Symbol::FB_common;
 213   if (Flags & object::BasicSymbolRef::SF_Indirect)
 214     Sym.Flags |= 1 << storage::Symbol::FB_indirect;
 215   if (Flags & object::BasicSymbolRef::SF_Global)
 216     Sym.Flags |= 1 << storage::Symbol::FB_global;
 217   if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
 218     Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
 219   if (Flags & object::BasicSymbolRef::SF_Executable)
 220     Sym.Flags |= 1 << storage::Symbol::FB_executable;
 221
 222   Sym.ComdatIndex = -1;
 223   auto *GV = Msym.dyn_cast<GlobalValue *>();
 224   if (!GV) {
 225     // Undefined module asm symbols act as GC roots and are implicitly used.
 226     if (Flags & object::BasicSymbolRef::SF_Undefined)
 227       Sym.Flags |= 1 << storage::Symbol::FB_used;
 228     setStr(Sym.IRName, "");
 229     return Error::success();
 230   }
 231
 232   setStr(Sym.IRName, GV->getName());
 233
 234   bool IsBuiltinFunc = false;
 235
 236   for (const char *LibcallName : LibcallRoutineNames)
 237     if (GV->getName() == LibcallName)
 238       IsBuiltinFunc = true;
 239
 240   if (Used.count(GV) || IsBuiltinFunc)
 241     Sym.Flags |= 1 << storage::Symbol::FB_used;
 242   if (GV->isThreadLocal())
 243     Sym.Flags |= 1 << storage::Symbol::FB_tls;
 244   if (GV->hasGlobalUnnamedAddr())
 245     Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
 246   if (GV->canBeOmittedFromSymbolTable())
 247     Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
 248   Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
 249
 250   if (Flags & object::BasicSymbolRef::SF_Common) {
 251     Uncommon().CommonSize = GV->getParent()->getDataLayout().getTypeAllocSize(
 252         GV->getType()->getElementType());
 253     Uncommon().CommonAlign = GV->getAlignment();
 254   }
 255
 256   const GlobalObject *Base = GV->getBaseObject();
 257   if (!Base)
 258     return make_error<StringError>("Unable to determine comdat of alias!",
 259                                    inconvertibleErrorCode());
 260   if (const Comdat *C = Base->getComdat()) {
 261     Expected<int> ComdatIndexOrErr = getComdatIndex(C, GV->getParent());
 262     if (!ComdatIndexOrErr)
 263       return ComdatIndexOrErr.takeError();
 264     Sym.ComdatIndex = *ComdatIndexOrErr;
 265   }
 266
 267   if (TT.isOSBinFormatCOFF()) {
 268     emitLinkerFlagsForGlobalCOFF(COFFLinkerOptsOS, GV, TT, Mang);
 269
 270     if ((Flags & object::BasicSymbolRef::SF_Weak) &&
 271         (Flags & object::BasicSymbolRef::SF_Indirect)) {
 272       auto *Fallback = dyn_cast<GlobalValue>(
 273           cast<GlobalAlias>(GV)->getAliasee()->stripPointerCasts());
 274       if (!Fallback)
 275         return make_error<StringError>("Invalid weak external",
 276                                        inconvertibleErrorCode());
 277       std::string FallbackName;
 278       raw_string_ostream OS(FallbackName);
 279       Msymtab.printSymbolName(OS, Fallback);
 280       OS.flush();
 281       setStr(Uncommon().COFFWeakExternFallbackName, Saver.save(FallbackName));
 282     }
 283   }
 284
 285   if (!Base->getSection().empty())
 286     setStr(Uncommon().SectionName, Saver.save(Base->getSection()));
 287
 288   return Error::success();
 289 }
 290
 291 Error Builder::build(ArrayRef<Module *> IRMods) {
 292   storage::Header Hdr;
 293
 294   assert(!IRMods.empty());
 295   Hdr.Version = storage::Header::kCurrentVersion;
 296   setStr(Hdr.Producer, kExpectedProducerName);
 297   setStr(Hdr.TargetTriple, IRMods[0]->getTargetTriple());
 298   setStr(Hdr.SourceFileName, IRMods[0]->getSourceFileName());
 299   TT = Triple(IRMods[0]->getTargetTriple());
 300
 301   for (auto *M : IRMods)
 302     if (Error Err = addModule(M))
 303       return Err;
 304
 305   COFFLinkerOptsOS.flush();
 306   setStr(Hdr.COFFLinkerOpts, Saver.save(COFFLinkerOpts));
 307
 308   // We are about to fill in the header's range fields, so reserve space for it
 309   // and copy it in afterwards.
 310   Symtab.resize(sizeof(storage::Header));
 311   writeRange(Hdr.Modules, Mods);
 312   writeRange(Hdr.Comdats, Comdats);
 313   writeRange(Hdr.Symbols, Syms);
 314   writeRange(Hdr.Uncommons, Uncommons);
 315
 316   *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
 317   return Error::success();
 318 }
 319
 320 } // end anonymous namespace
 321
 322 Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
 323                       StringTableBuilder &StrtabBuilder,
 324                       BumpPtrAllocator &Alloc) {
 325   return Builder(Symtab, StrtabBuilder, Alloc).build(Mods);
 326 }
 327
 328 // Upgrade a vector of bitcode modules created by an old version of LLVM by
 329 // creating an irsymtab for them in the current format.
 330 static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
 331   FileContents FC;
 332
 333   LLVMContext Ctx;
 334   std::vector<Module *> Mods;
 335   std::vector<std::unique_ptr<Module>> OwnedMods;
 336   for (auto BM : BMs) {
 337     Expected<std::unique_ptr<Module>> MOrErr =
 338         BM.getLazyModule(Ctx, /*ShouldLazyLoadMetadata*/ true,
 339                          /*IsImporting*/ false);
 340     if (!MOrErr)
 341       return MOrErr.takeError();
 342
 343     Mods.push_back(MOrErr->get());
 344     OwnedMods.push_back(std::move(*MOrErr));
 345   }
 346
 347   StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
 348   BumpPtrAllocator Alloc;
 349   if (Error E = build(Mods, FC.Symtab, StrtabBuilder, Alloc))
 350     return std::move(E);
 351
 352   StrtabBuilder.finalizeInOrder();
 353   FC.Strtab.resize(StrtabBuilder.getSize());
 354   StrtabBuilder.write((uint8_t *)FC.Strtab.data());
 355
 356   FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
 357                   {FC.Strtab.data(), FC.Strtab.size()}};
 358   return std::move(FC);
 359 }
 360
 361 Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
 362   if (BFC.Mods.empty())
 363     return make_error<StringError>("Bitcode file does not contain any modules",
 364                                    inconvertibleErrorCode());
 365
 366   if (BFC.StrtabForSymtab.empty() ||
 367       BFC.Symtab.size() < sizeof(storage::Header))
 368     return upgrade(BFC.Mods);
 369
 370   // We cannot use the regular reader to read the version and producer, because
 371   // it will expect the header to be in the current format. The only thing we
 372   // can rely on is that the version and producer will be present as the first
 373   // struct elements.
 374   auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
 375   unsigned Version = Hdr->Version;
 376   StringRef Producer = Hdr->Producer.get(BFC.StrtabForSymtab);
 377   if (Version != storage::Header::kCurrentVersion ||
 378       Producer != kExpectedProducerName)
 379     return upgrade(BFC.Mods);
 380
 381   FileContents FC;
 382   FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
 383                   {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
 384
 385   // Finally, make sure that the number of modules in the symbol table matches
 386   // the number of modules in the bitcode file. If they differ, it may mean that
 387   // the bitcode file was created by binary concatenation, so we need to create
 388   // a new symbol table from scratch.
 389   if (FC.TheReader.getNumModules() != BFC.Mods.size())
 390     return upgrade(std::move(BFC.Mods));
 391
 392   return std::move(FC);
 393 }