tools/llvm-symbolizer/llvm-symbolizer.cpp

   1 //===-- llvm-symbolizer.cpp - Simple addr2line-like symbolizer ------------===//
   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 // This utility works much like "addr2line". It is able of transforming
  10 // tuples (module name, module offset) to code locations (function name,
  11 // file, line number, column number). It is targeted for compiler-rt tools
  12 // (especially AddressSanitizer and ThreadSanitizer) that can use it
  13 // to symbolize stack traces in their error reports.
  14 //
  15 //===----------------------------------------------------------------------===//
  16
  17 #include "llvm/ADT/StringRef.h"
  18 #include "llvm/DebugInfo/Symbolize/DIPrinter.h"
  19 #include "llvm/DebugInfo/Symbolize/Symbolize.h"
  20 #include "llvm/Support/COM.h"
  21 #include "llvm/Support/CommandLine.h"
  22 #include "llvm/Support/Debug.h"
  23 #include "llvm/Support/FileSystem.h"
  24 #include "llvm/Support/InitLLVM.h"
  25 #include "llvm/Support/Path.h"
  26 #include "llvm/Support/raw_ostream.h"
  27 #include <cstdio>
  28 #include <cstring>
  29 #include <string>
  30
  31 using namespace llvm;
  32 using namespace symbolize;
  33
  34 static cl::opt<bool>
  35 ClUseSymbolTable("use-symbol-table", cl::init(true),
  36                  cl::desc("Prefer names in symbol table to names "
  37                           "in debug info"));
  38
  39 static cl::opt<FunctionNameKind> ClPrintFunctions(
  40     "functions", cl::init(FunctionNameKind::LinkageName),
  41     cl::desc("Print function name for a given address"), cl::ValueOptional,
  42     cl::values(clEnumValN(FunctionNameKind::None, "none", "omit function name"),
  43                clEnumValN(FunctionNameKind::ShortName, "short",
  44                           "print short function name"),
  45                clEnumValN(FunctionNameKind::LinkageName, "linkage",
  46                           "print function linkage name"),
  47                // Sentinel value for unspecified value.
  48                clEnumValN(FunctionNameKind::LinkageName, "", "")));
  49 static cl::alias ClPrintFunctionsShort("f", cl::desc("Alias for -functions"),
  50                                        cl::NotHidden, cl::Grouping,
  51                                        cl::aliasopt(ClPrintFunctions));
  52
  53 static cl::opt<bool>
  54     ClUseRelativeAddress("relative-address", cl::init(false),
  55                          cl::desc("Interpret addresses as relative addresses"),
  56                          cl::ReallyHidden);
  57
  58 static cl::opt<bool> ClUntagAddresses(
  59     "untag-addresses", cl::init(true),
  60     cl::desc("Remove memory tags from addresses before symbolization"));
  61
  62 static cl::opt<bool>
  63     ClPrintInlining("inlining", cl::init(true),
  64                     cl::desc("Print all inlined frames for a given address"));
  65 static cl::alias
  66     ClPrintInliningAliasI("i", cl::desc("Alias for -inlining"),
  67                           cl::NotHidden, cl::aliasopt(ClPrintInlining),
  68                           cl::Grouping);
  69 static cl::alias
  70     ClPrintInliningAliasInlines("inlines", cl::desc("Alias for -inlining"),
  71                                 cl::NotHidden, cl::aliasopt(ClPrintInlining));
  72
  73 // -basenames, -s
  74 static cl::opt<bool> ClBasenames("basenames", cl::init(false),
  75                                  cl::desc("Strip directory names from paths"));
  76 static cl::alias ClBasenamesShort("s", cl::desc("Alias for -basenames"),
  77                                   cl::NotHidden, cl::aliasopt(ClBasenames));
  78
  79 // -demangle, -C, -no-demangle
  80 static cl::opt<bool>
  81 ClDemangle("demangle", cl::init(true), cl::desc("Demangle function names"));
  82 static cl::alias
  83 ClDemangleShort("C", cl::desc("Alias for -demangle"),
  84                 cl::NotHidden, cl::aliasopt(ClDemangle), cl::Grouping);
  85 static cl::opt<bool>
  86 ClNoDemangle("no-demangle", cl::init(false),
  87              cl::desc("Don't demangle function names"));
  88
  89 static cl::opt<std::string> ClDefaultArch("default-arch", cl::init(""),
  90                                           cl::desc("Default architecture "
  91                                                    "(for multi-arch objects)"));
  92
  93 // -obj, -exe, -e
  94 static cl::opt<std::string>
  95 ClBinaryName("obj", cl::init(""),
  96              cl::desc("Path to object file to be symbolized (if not provided, "
  97                       "object file should be specified for each input line)"));
  98 static cl::alias
  99 ClBinaryNameAliasExe("exe", cl::desc("Alias for -obj"),
 100                      cl::NotHidden, cl::aliasopt(ClBinaryName));
 101 static cl::alias ClBinaryNameAliasE("e", cl::desc("Alias for -obj"),
 102                                     cl::NotHidden, cl::Grouping, cl::Prefix,
 103                                     cl::aliasopt(ClBinaryName));
 104
 105 static cl::opt<std::string>
 106     ClDwpName("dwp", cl::init(""),
 107               cl::desc("Path to DWP file to be use for any split CUs"));
 108
 109 static cl::list<std::string>
 110 ClDsymHint("dsym-hint", cl::ZeroOrMore,
 111            cl::desc("Path to .dSYM bundles to search for debug info for the "
 112                     "object files"));
 113
 114 // -print-address, -addresses, -a
 115 static cl::opt<bool>
 116 ClPrintAddress("print-address", cl::init(false),
 117                cl::desc("Show address before line information"));
 118 static cl::alias
 119 ClPrintAddressAliasAddresses("addresses", cl::desc("Alias for -print-address"),
 120                              cl::NotHidden, cl::aliasopt(ClPrintAddress));
 121 static cl::alias
 122 ClPrintAddressAliasA("a", cl::desc("Alias for -print-address"),
 123                      cl::NotHidden, cl::aliasopt(ClPrintAddress), cl::Grouping);
 124
 125 // -pretty-print, -p
 126 static cl::opt<bool>
 127     ClPrettyPrint("pretty-print", cl::init(false),
 128                   cl::desc("Make the output more human friendly"));
 129 static cl::alias ClPrettyPrintShort("p", cl::desc("Alias for -pretty-print"),
 130                                     cl::NotHidden,
 131                                     cl::aliasopt(ClPrettyPrint), cl::Grouping);
 132
 133 static cl::opt<int> ClPrintSourceContextLines(
 134     "print-source-context-lines", cl::init(0),
 135     cl::desc("Print N number of source file context"));
 136
 137 static cl::opt<bool> ClVerbose("verbose", cl::init(false),
 138                                cl::desc("Print verbose line info"));
 139
 140 // -adjust-vma
 141 static cl::opt<uint64_t>
 142     ClAdjustVMA("adjust-vma", cl::init(0), cl::value_desc("offset"),
 143                 cl::desc("Add specified offset to object file addresses"));
 144
 145 static cl::list<std::string> ClInputAddresses(cl::Positional,
 146                                               cl::desc("<input addresses>..."),
 147                                               cl::ZeroOrMore);
 148
 149 static cl::opt<std::string>
 150     ClFallbackDebugPath("fallback-debug-path", cl::init(""),
 151                         cl::desc("Fallback path for debug binaries."));
 152
 153 static cl::opt<DIPrinter::OutputStyle>
 154     ClOutputStyle("output-style", cl::init(DIPrinter::OutputStyle::LLVM),
 155                   cl::desc("Specify print style"),
 156                   cl::values(clEnumValN(DIPrinter::OutputStyle::LLVM, "LLVM",
 157                                         "LLVM default style"),
 158                              clEnumValN(DIPrinter::OutputStyle::GNU, "GNU",
 159                                         "GNU addr2line style")));
 160
 161 static cl::extrahelp
 162     HelpResponse("\nPass @FILE as argument to read options from FILE.\n");
 163
 164 template<typename T>
 165 static bool error(Expected<T> &ResOrErr) {
 166   if (ResOrErr)
 167     return false;
 168   logAllUnhandledErrors(ResOrErr.takeError(), errs(),
 169                         "LLVMSymbolizer: error reading file: ");
 170   return true;
 171 }
 172
 173 enum class Command {
 174   Code,
 175   Data,
 176   Frame,
 177 };
 178
 179 static bool parseCommand(StringRef InputString, Command &Cmd,
 180                          std::string &ModuleName, uint64_t &ModuleOffset) {
 181   const char kDelimiters[] = " \n\r";
 182   ModuleName = "";
 183   if (InputString.consume_front("CODE ")) {
 184     Cmd = Command::Code;
 185   } else if (InputString.consume_front("DATA ")) {
 186     Cmd = Command::Data;
 187   } else if (InputString.consume_front("FRAME ")) {
 188     Cmd = Command::Frame;
 189   } else {
 190     // If no cmd, assume it's CODE.
 191     Cmd = Command::Code;
 192   }
 193   const char *pos = InputString.data();
 194   // Skip delimiters and parse input filename (if needed).
 195   if (ClBinaryName.empty()) {
 196     pos += strspn(pos, kDelimiters);
 197     if (*pos == '"' || *pos == '\'') {
 198       char quote = *pos;
 199       pos++;
 200       const char *end = strchr(pos, quote);
 201       if (!end)
 202         return false;
 203       ModuleName = std::string(pos, end - pos);
 204       pos = end + 1;
 205     } else {
 206       int name_length = strcspn(pos, kDelimiters);
 207       ModuleName = std::string(pos, name_length);
 208       pos += name_length;
 209     }
 210   } else {
 211     ModuleName = ClBinaryName;
 212   }
 213   // Skip delimiters and parse module offset.
 214   pos += strspn(pos, kDelimiters);
 215   int offset_length = strcspn(pos, kDelimiters);
 216   return !StringRef(pos, offset_length).getAsInteger(0, ModuleOffset);
 217 }
 218
 219 static void symbolizeInput(StringRef InputString, LLVMSymbolizer &Symbolizer,
 220                            DIPrinter &Printer) {
 221   Command Cmd;
 222   std::string ModuleName;
 223   uint64_t Offset = 0;
 224   if (!parseCommand(StringRef(InputString), Cmd, ModuleName, Offset)) {
 225     outs() << InputString;
 226     return;
 227   }
 228
 229   if (ClPrintAddress) {
 230     outs() << "0x";
 231     outs().write_hex(Offset);
 232     StringRef Delimiter = ClPrettyPrint ? ": " : "\n";
 233     outs() << Delimiter;
 234   }
 235   Offset -= ClAdjustVMA;
 236   if (Cmd == Command::Data) {
 237     auto ResOrErr = Symbolizer.symbolizeData(
 238         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 239     Printer << (error(ResOrErr) ? DIGlobal() : ResOrErr.get());
 240   } else if (Cmd == Command::Frame) {
 241     auto ResOrErr = Symbolizer.symbolizeFrame(
 242         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 243     if (!error(ResOrErr)) {
 244       for (DILocal Local : *ResOrErr)
 245         Printer << Local;
 246       if (ResOrErr->empty())
 247         outs() << "??\n";
 248     }
 249   } else if (ClPrintInlining) {
 250     auto ResOrErr = Symbolizer.symbolizeInlinedCode(
 251         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 252     Printer << (error(ResOrErr) ? DIInliningInfo() : ResOrErr.get());
 253   } else if (ClOutputStyle == DIPrinter::OutputStyle::GNU) {
 254     // With ClPrintFunctions == FunctionNameKind::LinkageName (default)
 255     // and ClUseSymbolTable == true (also default), Symbolizer.symbolizeCode()
 256     // may override the name of an inlined function with the name of the topmost
 257     // caller function in the inlining chain. This contradicts the existing
 258     // behavior of addr2line. Symbolizer.symbolizeInlinedCode() overrides only
 259     // the topmost function, which suits our needs better.
 260     auto ResOrErr = Symbolizer.symbolizeInlinedCode(
 261         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 262     Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get().getFrame(0));
 263   } else {
 264     auto ResOrErr = Symbolizer.symbolizeCode(
 265         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 266     Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get());
 267   }
 268   if (ClOutputStyle == DIPrinter::OutputStyle::LLVM)
 269     outs() << "\n";
 270 }
 271
 272 int main(int argc, char **argv) {
 273   InitLLVM X(argc, argv);
 274
 275   bool IsAddr2Line = sys::path::stem(argv[0]).contains("addr2line");
 276
 277   if (IsAddr2Line) {
 278     ClDemangle.setInitialValue(false);
 279     ClPrintFunctions.setInitialValue(FunctionNameKind::None);
 280     ClPrintInlining.setInitialValue(false);
 281     ClUntagAddresses.setInitialValue(false);
 282     ClOutputStyle.setInitialValue(DIPrinter::OutputStyle::GNU);
 283   }
 284
 285   llvm::sys::InitializeCOMRAII COM(llvm::sys::COMThreadingMode::MultiThreaded);
 286   cl::ParseCommandLineOptions(argc, argv, IsAddr2Line ? "llvm-addr2line\n"
 287                                                       : "llvm-symbolizer\n");
 288
 289   // If both --demangle and --no-demangle are specified then pick the last one.
 290   if (ClNoDemangle.getPosition() > ClDemangle.getPosition())
 291     ClDemangle = !ClNoDemangle;
 292
 293   LLVMSymbolizer::Options Opts;
 294   Opts.PrintFunctions = ClPrintFunctions;
 295   Opts.UseSymbolTable = ClUseSymbolTable;
 296   Opts.Demangle = ClDemangle;
 297   Opts.RelativeAddresses = ClUseRelativeAddress;
 298   Opts.UntagAddresses = ClUntagAddresses;
 299   Opts.DefaultArch = ClDefaultArch;
 300   Opts.FallbackDebugPath = ClFallbackDebugPath;
 301   Opts.DWPName = ClDwpName;
 302
 303   for (const auto &hint : ClDsymHint) {
 304     if (sys::path::extension(hint) == ".dSYM") {
 305       Opts.DsymHints.push_back(hint);
 306     } else {
 307       errs() << "Warning: invalid dSYM hint: \"" << hint <<
 308                 "\" (must have the '.dSYM' extension).\n";
 309     }
 310   }
 311   LLVMSymbolizer Symbolizer(Opts);
 312
 313   DIPrinter Printer(outs(), ClPrintFunctions != FunctionNameKind::None,
 314                     ClPrettyPrint, ClPrintSourceContextLines, ClVerbose,
 315                     ClBasenames, ClOutputStyle);
 316
 317   if (ClInputAddresses.empty()) {
 318     const int kMaxInputStringLength = 1024;
 319     char InputString[kMaxInputStringLength];
 320
 321     while (fgets(InputString, sizeof(InputString), stdin)) {
 322       symbolizeInput(InputString, Symbolizer, Printer);
 323       outs().flush();
 324     }
 325   } else {
 326     for (StringRef Address : ClInputAddresses)
 327       symbolizeInput(Address, Symbolizer, Printer);
 328   }
 329
 330   return 0;
 331 }