tools/llvm-nm/llvm-nm.cpp

   1 //===-- llvm-nm.cpp - Symbol table dumping utility for llvm ---------------===//
   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 program is a utility that works like traditional Unix "nm",
  11 // that is, it prints out the names of symbols in a bitcode file,
  12 // along with some information about each symbol.
  13 //
  14 // This "nm" does not print symbols' addresses. It supports many of
  15 // the features of GNU "nm", including its different output formats.
  16 //
  17 //===----------------------------------------------------------------------===//
  18
  19 #include "llvm/LLVMContext.h"
  20 #include "llvm/Module.h"
  21 #include "llvm/Bitcode/ReaderWriter.h"
  22 #include "llvm/Bitcode/Archive.h"
  23 #include "llvm/Object/ObjectFile.h"
  24 #include "llvm/Support/CommandLine.h"
  25 #include "llvm/Support/FileSystem.h"
  26 #include "llvm/Support/ManagedStatic.h"
  27 #include "llvm/Support/MemoryBuffer.h"
  28 #include "llvm/Support/PrettyStackTrace.h"
  29 #include "llvm/Support/raw_ostream.h"
  30 #include "llvm/Support/Signals.h"
  31 #include "llvm/Support/Format.h"
  32 #include "llvm/Support/system_error.h"
  33 #include <algorithm>
  34 #include <cctype>
  35 #include <cerrno>
  36 #include <cstring>
  37 #include <vector>
  38 using namespace llvm;
  39 using namespace object;
  40
  41 namespace {
  42   enum OutputFormatTy { bsd, sysv, posix };
  43   cl::opt<OutputFormatTy>
  44   OutputFormat("format",
  45        cl::desc("Specify output format"),
  46          cl::values(clEnumVal(bsd,   "BSD format"),
  47                     clEnumVal(sysv,  "System V format"),
  48                     clEnumVal(posix, "POSIX.2 format"),
  49                     clEnumValEnd), cl::init(bsd));
  50   cl::alias OutputFormat2("f", cl::desc("Alias for --format"),
  51                           cl::aliasopt(OutputFormat));
  52
  53   cl::list<std::string>
  54   InputFilenames(cl::Positional, cl::desc("<input bitcode files>"),
  55                  cl::ZeroOrMore);
  56
  57   cl::opt<bool> UndefinedOnly("undefined-only",
  58                               cl::desc("Show only undefined symbols"));
  59   cl::alias UndefinedOnly2("u", cl::desc("Alias for --undefined-only"),
  60                            cl::aliasopt(UndefinedOnly));
  61
  62   cl::opt<bool> DefinedOnly("defined-only",
  63                             cl::desc("Show only defined symbols"));
  64
  65   cl::opt<bool> ExternalOnly("extern-only",
  66                              cl::desc("Show only external symbols"));
  67   cl::alias ExternalOnly2("g", cl::desc("Alias for --extern-only"),
  68                           cl::aliasopt(ExternalOnly));
  69
  70   cl::opt<bool> BSDFormat("B", cl::desc("Alias for --format=bsd"));
  71   cl::opt<bool> POSIXFormat("P", cl::desc("Alias for --format=posix"));
  72
  73   cl::opt<bool> PrintFileName("print-file-name",
  74     cl::desc("Precede each symbol with the object file it came from"));
  75
  76   cl::alias PrintFileNameA("A", cl::desc("Alias for --print-file-name"),
  77                                 cl::aliasopt(PrintFileName));
  78   cl::alias PrintFileNameo("o", cl::desc("Alias for --print-file-name"),
  79                                 cl::aliasopt(PrintFileName));
  80
  81   cl::opt<bool> DebugSyms("debug-syms",
  82     cl::desc("Show all symbols, even debugger only"));
  83   cl::alias DebugSymsa("a", cl::desc("Alias for --debug-syms"),
  84                             cl::aliasopt(DebugSyms));
  85
  86   cl::opt<bool> NumericSort("numeric-sort",
  87     cl::desc("Sort symbols by address"));
  88   cl::alias NumericSortn("n", cl::desc("Alias for --numeric-sort"),
  89                               cl::aliasopt(NumericSort));
  90   cl::alias NumericSortv("v", cl::desc("Alias for --numeric-sort"),
  91                               cl::aliasopt(NumericSort));
  92
  93   cl::opt<bool> NoSort("no-sort",
  94     cl::desc("Show symbols in order encountered"));
  95   cl::alias NoSortp("p", cl::desc("Alias for --no-sort"),
  96                          cl::aliasopt(NoSort));
  97
  98   cl::opt<bool> PrintSize("print-size",
  99     cl::desc("Show symbol size instead of address"));
 100   cl::alias PrintSizeS("S", cl::desc("Alias for --print-size"),
 101                             cl::aliasopt(PrintSize));
 102
 103   cl::opt<bool> SizeSort("size-sort", cl::desc("Sort symbols by size"));
 104
 105   bool PrintAddress = true;
 106
 107   bool MultipleFiles = false;
 108
 109   std::string ToolName;
 110 }
 111
 112 namespace {
 113   struct NMSymbol {
 114     uint64_t  Address;
 115     uint64_t  Size;
 116     char      TypeChar;
 117     StringRef Name;
 118   };
 119
 120   static bool CompareSymbolAddress(const NMSymbol &a, const NMSymbol &b) {
 121     if (a.Address < b.Address)
 122       return true;
 123     else if (a.Address == b.Address && a.Name < b.Name)
 124       return true;
 125     else
 126       return false;
 127
 128   }
 129
 130   static bool CompareSymbolSize(const NMSymbol &a, const NMSymbol &b) {
 131     if (a.Size < b.Size)
 132       return true;
 133     else if (a.Size == b.Size && a.Name < b.Name)
 134       return true;
 135     else
 136       return false;
 137   }
 138
 139   static bool CompareSymbolName(const NMSymbol &a, const NMSymbol &b) {
 140     return a.Name < b.Name;
 141   }
 142
 143   StringRef CurrentFilename;
 144   typedef std::vector<NMSymbol> SymbolListT;
 145   SymbolListT SymbolList;
 146
 147   bool error(error_code ec) {
 148     if (!ec) return false;
 149
 150     outs() << ToolName << ": error reading file: " << ec.message() << ".\n";
 151     outs().flush();
 152     return true;
 153   }
 154 }
 155
 156 static void SortAndPrintSymbolList() {
 157   if (!NoSort) {
 158     if (NumericSort)
 159       std::sort(SymbolList.begin(), SymbolList.end(), CompareSymbolAddress);
 160     else if (SizeSort)
 161       std::sort(SymbolList.begin(), SymbolList.end(), CompareSymbolSize);
 162     else
 163       std::sort(SymbolList.begin(), SymbolList.end(), CompareSymbolName);
 164   }
 165
 166   if (OutputFormat == posix && MultipleFiles) {
 167     outs() << '\n' << CurrentFilename << ":\n";
 168   } else if (OutputFormat == bsd && MultipleFiles) {
 169     outs() << "\n" << CurrentFilename << ":\n";
 170   } else if (OutputFormat == sysv) {
 171     outs() << "\n\nSymbols from " << CurrentFilename << ":\n\n"
 172            << "Name                  Value   Class        Type"
 173            << "         Size   Line  Section\n";
 174   }
 175
 176   for (SymbolListT::iterator i = SymbolList.begin(),
 177                              e = SymbolList.end(); i != e; ++i) {
 178     if ((i->TypeChar != 'U') && UndefinedOnly)
 179       continue;
 180     if ((i->TypeChar == 'U') && DefinedOnly)
 181       continue;
 182     if (SizeSort && !PrintAddress && i->Size == UnknownAddressOrSize)
 183       continue;
 184
 185     char SymbolAddrStr[10] = "";
 186     char SymbolSizeStr[10] = "";
 187
 188     if (OutputFormat == sysv || i->Address == object::UnknownAddressOrSize)
 189       strcpy(SymbolAddrStr, "        ");
 190     if (OutputFormat == sysv)
 191       strcpy(SymbolSizeStr, "        ");
 192
 193     if (i->Address != object::UnknownAddressOrSize)
 194       format("%08x", i->Address).print(SymbolAddrStr, sizeof(SymbolAddrStr));
 195     if (i->Size != object::UnknownAddressOrSize)
 196       format("%08x", i->Size).print(SymbolSizeStr, sizeof(SymbolSizeStr));
 197
 198     if (OutputFormat == posix) {
 199       outs() << i->Name << " " << i->TypeChar << " "
 200              << SymbolAddrStr << SymbolSizeStr << "\n";
 201     } else if (OutputFormat == bsd) {
 202       if (PrintAddress)
 203         outs() << SymbolAddrStr << ' ';
 204       if (PrintSize) {
 205         outs() << SymbolSizeStr;
 206         if (i->Size != object::UnknownAddressOrSize)
 207           outs() << ' ';
 208       }
 209       outs() << i->TypeChar << " " << i->Name  << "\n";
 210     } else if (OutputFormat == sysv) {
 211       std::string PaddedName (i->Name);
 212       while (PaddedName.length () < 20)
 213         PaddedName += " ";
 214       outs() << PaddedName << "|" << SymbolAddrStr << "|   "
 215              << i->TypeChar
 216              << "  |                  |" << SymbolSizeStr << "|     |\n";
 217     }
 218   }
 219
 220   SymbolList.clear();
 221 }
 222
 223 static char TypeCharForSymbol(GlobalValue &GV) {
 224   if (GV.isDeclaration())                                  return 'U';
 225   if (GV.hasLinkOnceLinkage())                             return 'C';
 226   if (GV.hasCommonLinkage())                               return 'C';
 227   if (GV.hasWeakLinkage())                                 return 'W';
 228   if (isa<Function>(GV) && GV.hasInternalLinkage())        return 't';
 229   if (isa<Function>(GV))                                   return 'T';
 230   if (isa<GlobalVariable>(GV) && GV.hasInternalLinkage())  return 'd';
 231   if (isa<GlobalVariable>(GV))                             return 'D';
 232   if (const GlobalAlias *GA = dyn_cast<GlobalAlias>(&GV)) {
 233     const GlobalValue *AliasedGV = GA->getAliasedGlobal();
 234     if (isa<Function>(AliasedGV))                          return 'T';
 235     if (isa<GlobalVariable>(AliasedGV))                    return 'D';
 236   }
 237                                                            return '?';
 238 }
 239
 240 static void DumpSymbolNameForGlobalValue(GlobalValue &GV) {
 241   // Private linkage and available_externally linkage don't exist in symtab.
 242   if (GV.hasPrivateLinkage() ||
 243       GV.hasLinkerPrivateLinkage() ||
 244       GV.hasLinkerPrivateWeakLinkage() ||
 245       GV.hasLinkerPrivateWeakDefAutoLinkage() ||
 246       GV.hasAvailableExternallyLinkage())
 247     return;
 248   char TypeChar = TypeCharForSymbol(GV);
 249   if (GV.hasLocalLinkage () && ExternalOnly)
 250     return;
 251
 252   NMSymbol s;
 253   s.Address = object::UnknownAddressOrSize;
 254   s.Size = object::UnknownAddressOrSize;
 255   s.TypeChar = TypeChar;
 256   s.Name     = GV.getName();
 257   SymbolList.push_back(s);
 258 }
 259
 260 static void DumpSymbolNamesFromModule(Module *M) {
 261   CurrentFilename = M->getModuleIdentifier();
 262   std::for_each (M->begin(), M->end(), DumpSymbolNameForGlobalValue);
 263   std::for_each (M->global_begin(), M->global_end(),
 264                  DumpSymbolNameForGlobalValue);
 265   std::for_each (M->alias_begin(), M->alias_end(),
 266                  DumpSymbolNameForGlobalValue);
 267
 268   SortAndPrintSymbolList();
 269 }
 270
 271 static void DumpSymbolNamesFromObject(ObjectFile *obj) {
 272   error_code ec;
 273   for (ObjectFile::symbol_iterator i = obj->begin_symbols(),
 274                                    e = obj->end_symbols();
 275                                    i != e; i.increment(ec)) {
 276     if (error(ec)) break;
 277     bool internal;
 278     if (error(i->isInternal(internal))) break;
 279     if (!DebugSyms && internal)
 280       continue;
 281     NMSymbol s;
 282     s.Size = object::UnknownAddressOrSize;
 283     s.Address = object::UnknownAddressOrSize;
 284     if (PrintSize || SizeSort) {
 285       if (error(i->getSize(s.Size))) break;
 286     }
 287     if (PrintAddress)
 288       if (error(i->getAddress(s.Address))) break;
 289     if (error(i->getNMTypeChar(s.TypeChar))) break;
 290     if (error(i->getName(s.Name))) break;
 291     SymbolList.push_back(s);
 292   }
 293
 294   CurrentFilename = obj->getFileName();
 295   SortAndPrintSymbolList();
 296 }
 297
 298 static void DumpSymbolNamesFromFile(std::string &Filename) {
 299   LLVMContext &Context = getGlobalContext();
 300   std::string ErrorMessage;
 301   sys::Path aPath(Filename);
 302   bool exists;
 303   if (sys::fs::exists(aPath.str(), exists) || !exists)
 304     errs() << ToolName << ": '" << Filename << "': " << "No such file\n";
 305   // Note: Currently we do not support reading an archive from stdin.
 306   if (Filename == "-" || aPath.isBitcodeFile()) {
 307     OwningPtr<MemoryBuffer> Buffer;
 308     if (error_code ec = MemoryBuffer::getFileOrSTDIN(Filename, Buffer))
 309       ErrorMessage = ec.message();
 310     Module *Result = 0;
 311     if (Buffer.get())
 312       Result = ParseBitcodeFile(Buffer.get(), Context, &ErrorMessage);
 313
 314     if (Result) {
 315       DumpSymbolNamesFromModule(Result);
 316       delete Result;
 317     } else
 318       errs() << ToolName << ": " << Filename << ": " << ErrorMessage << "\n";
 319
 320   } else if (aPath.isArchive()) {
 321     std::string ErrMsg;
 322     Archive* archive = Archive::OpenAndLoad(sys::Path(Filename), Context,
 323                                             &ErrorMessage);
 324     if (!archive)
 325       errs() << ToolName << ": " << Filename << ": " << ErrorMessage << "\n";
 326     std::vector<Module *> Modules;
 327     if (archive->getAllModules(Modules, &ErrorMessage)) {
 328       errs() << ToolName << ": " << Filename << ": " << ErrorMessage << "\n";
 329       return;
 330     }
 331     MultipleFiles = true;
 332     std::for_each (Modules.begin(), Modules.end(), DumpSymbolNamesFromModule);
 333   } else if (aPath.isObjectFile()) {
 334     OwningPtr<Binary> obj;
 335     if (error_code ec = object::createBinary(aPath.str(), obj)) {
 336       errs() << ToolName << ": " << Filename << ": " << ec.message() << ".\n";
 337       return;
 338     }
 339     if (object::ObjectFile *o = dyn_cast<ObjectFile>(obj.get()))
 340       DumpSymbolNamesFromObject(o);
 341   } else {
 342     errs() << ToolName << ": " << Filename << ": "
 343            << "unrecognizable file type\n";
 344     return;
 345   }
 346 }
 347
 348 int main(int argc, char **argv) {
 349   // Print a stack trace if we signal out.
 350   sys::PrintStackTraceOnErrorSignal();
 351   PrettyStackTraceProgram X(argc, argv);
 352
 353   llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
 354   cl::ParseCommandLineOptions(argc, argv, "llvm symbol table dumper\n");
 355
 356   ToolName = argv[0];
 357   if (BSDFormat) OutputFormat = bsd;
 358   if (POSIXFormat) OutputFormat = posix;
 359
 360   // The relative order of these is important. If you pass --size-sort it should
 361   // only print out the size. However, if you pass -S --size-sort, it should
 362   // print out both the size and address.
 363   if (SizeSort && !PrintSize) PrintAddress = false;
 364   if (OutputFormat == sysv || SizeSort) PrintSize = true;
 365
 366   switch (InputFilenames.size()) {
 367   case 0: InputFilenames.push_back("-");
 368   case 1: break;
 369   default: MultipleFiles = true;
 370   }
 371
 372   std::for_each(InputFilenames.begin(), InputFilenames.end(),
 373                 DumpSymbolNamesFromFile);
 374   return 0;
 375 }