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>
  59     ClPrintInlining("inlining", cl::init(true),
  60                     cl::desc("Print all inlined frames for a given address"));
  61 static cl::alias
  62     ClPrintInliningAliasI("i", cl::desc("Alias for -inlining"),
  63                           cl::NotHidden, cl::aliasopt(ClPrintInlining),
  64                           cl::Grouping);
  65 static cl::alias
  66     ClPrintInliningAliasInlines("inlines", cl::desc("Alias for -inlining"),
  67                                 cl::NotHidden, cl::aliasopt(ClPrintInlining));
  68
  69 // -basenames, -s
  70 static cl::opt<bool> ClBasenames("basenames", cl::init(false),
  71                                  cl::desc("Strip directory names from paths"));
  72 static cl::alias ClBasenamesShort("s", cl::desc("Alias for -basenames"),
  73                                   cl::NotHidden, cl::aliasopt(ClBasenames));
  74
  75 // -demangle, -C, -no-demangle
  76 static cl::opt<bool>
  77 ClDemangle("demangle", cl::init(true), cl::desc("Demangle function names"));
  78 static cl::alias
  79 ClDemangleShort("C", cl::desc("Alias for -demangle"),
  80                 cl::NotHidden, cl::aliasopt(ClDemangle), cl::Grouping);
  81 static cl::opt<bool>
  82 ClNoDemangle("no-demangle", cl::init(false),
  83              cl::desc("Don't demangle function names"));
  84
  85 static cl::opt<std::string> ClDefaultArch("default-arch", cl::init(""),
  86                                           cl::desc("Default architecture "
  87                                                    "(for multi-arch objects)"));
  88
  89 // -obj, -exe, -e
  90 static cl::opt<std::string>
  91 ClBinaryName("obj", cl::init(""),
  92              cl::desc("Path to object file to be symbolized (if not provided, "
  93                       "object file should be specified for each input line)"));
  94 static cl::alias
  95 ClBinaryNameAliasExe("exe", cl::desc("Alias for -obj"),
  96                      cl::NotHidden, cl::aliasopt(ClBinaryName));
  97 static cl::alias ClBinaryNameAliasE("e", cl::desc("Alias for -obj"),
  98                                     cl::NotHidden, cl::Grouping, cl::Prefix,
  99                                     cl::aliasopt(ClBinaryName));
 100
 101 static cl::opt<std::string>
 102     ClDwpName("dwp", cl::init(""),
 103               cl::desc("Path to DWP file to be use for any split CUs"));
 104
 105 static cl::list<std::string>
 106 ClDsymHint("dsym-hint", cl::ZeroOrMore,
 107            cl::desc("Path to .dSYM bundles to search for debug info for the "
 108                     "object files"));
 109
 110 // -print-address, -addresses, -a
 111 static cl::opt<bool>
 112 ClPrintAddress("print-address", cl::init(false),
 113                cl::desc("Show address before line information"));
 114 static cl::alias
 115 ClPrintAddressAliasAddresses("addresses", cl::desc("Alias for -print-address"),
 116                              cl::NotHidden, cl::aliasopt(ClPrintAddress));
 117 static cl::alias
 118 ClPrintAddressAliasA("a", cl::desc("Alias for -print-address"),
 119                      cl::NotHidden, cl::aliasopt(ClPrintAddress), cl::Grouping);
 120
 121 // -pretty-print, -p
 122 static cl::opt<bool>
 123     ClPrettyPrint("pretty-print", cl::init(false),
 124                   cl::desc("Make the output more human friendly"));
 125 static cl::alias ClPrettyPrintShort("p", cl::desc("Alias for -pretty-print"),
 126                                     cl::NotHidden,
 127                                     cl::aliasopt(ClPrettyPrint), cl::Grouping);
 128
 129 static cl::opt<int> ClPrintSourceContextLines(
 130     "print-source-context-lines", cl::init(0),
 131     cl::desc("Print N number of source file context"));
 132
 133 static cl::opt<bool> ClVerbose("verbose", cl::init(false),
 134                                cl::desc("Print verbose line info"));
 135
 136 // -adjust-vma
 137 static cl::opt<uint64_t>
 138     ClAdjustVMA("adjust-vma", cl::init(0), cl::value_desc("offset"),
 139                 cl::desc("Add specified offset to object file addresses"));
 140
 141 static cl::list<std::string> ClInputAddresses(cl::Positional,
 142                                               cl::desc("<input addresses>..."),
 143                                               cl::ZeroOrMore);
 144
 145 static cl::opt<std::string>
 146     ClFallbackDebugPath("fallback-debug-path", cl::init(""),
 147                         cl::desc("Fallback path for debug binaries."));
 148
 149 static cl::opt<DIPrinter::OutputStyle>
 150     ClOutputStyle("output-style", cl::init(DIPrinter::OutputStyle::LLVM),
 151                   cl::desc("Specify print style"),
 152                   cl::values(clEnumValN(DIPrinter::OutputStyle::LLVM, "LLVM",
 153                                         "LLVM default style"),
 154                              clEnumValN(DIPrinter::OutputStyle::GNU, "GNU",
 155                                         "GNU addr2line style")));
 156
 157 template<typename T>
 158 static bool error(Expected<T> &ResOrErr) {
 159   if (ResOrErr)
 160     return false;
 161   logAllUnhandledErrors(ResOrErr.takeError(), errs(),
 162                         "LLVMSymbolizer: error reading file: ");
 163   return true;
 164 }
 165
 166 static bool parseCommand(StringRef InputString, bool &IsData,
 167                          std::string &ModuleName, uint64_t &ModuleOffset) {
 168   const char kDelimiters[] = " \n\r";
 169   ModuleName = "";
 170   if (InputString.consume_front("CODE ")) {
 171     IsData = false;
 172   } else if (InputString.consume_front("DATA ")) {
 173     IsData = true;
 174   } else {
 175     // If no cmd, assume it's CODE.
 176     IsData = false;
 177   }
 178   const char *pos = InputString.data();
 179   // Skip delimiters and parse input filename (if needed).
 180   if (ClBinaryName.empty()) {
 181     pos += strspn(pos, kDelimiters);
 182     if (*pos == '"' || *pos == '\'') {
 183       char quote = *pos;
 184       pos++;
 185       const char *end = strchr(pos, quote);
 186       if (!end)
 187         return false;
 188       ModuleName = std::string(pos, end - pos);
 189       pos = end + 1;
 190     } else {
 191       int name_length = strcspn(pos, kDelimiters);
 192       ModuleName = std::string(pos, name_length);
 193       pos += name_length;
 194     }
 195   } else {
 196     ModuleName = ClBinaryName;
 197   }
 198   // Skip delimiters and parse module offset.
 199   pos += strspn(pos, kDelimiters);
 200   int offset_length = strcspn(pos, kDelimiters);
 201   return !StringRef(pos, offset_length).getAsInteger(0, ModuleOffset);
 202 }
 203
 204 static void symbolizeInput(StringRef InputString, LLVMSymbolizer &Symbolizer,
 205                            DIPrinter &Printer) {
 206   bool IsData = false;
 207   std::string ModuleName;
 208   uint64_t Offset = 0;
 209   if (!parseCommand(StringRef(InputString), IsData, ModuleName, Offset)) {
 210     outs() << InputString;
 211     return;
 212   }
 213
 214   if (ClPrintAddress) {
 215     outs() << "0x";
 216     outs().write_hex(Offset);
 217     StringRef Delimiter = ClPrettyPrint ? ": " : "\n";
 218     outs() << Delimiter;
 219   }
 220   Offset -= ClAdjustVMA;
 221   if (IsData) {
 222     auto ResOrErr = Symbolizer.symbolizeData(
 223         ModuleName, {Offset, object::SectionedAddress::UndefSection});
 224     Printer << (error(ResOrErr) ? DIGlobal() : ResOrErr.get());
 225   } else if (ClPrintInlining) {
 226     auto ResOrErr = Symbolizer.symbolizeInlinedCode(
 227         ModuleName, {Offset, object::SectionedAddress::UndefSection},
 228         ClDwpName);
 229     Printer << (error(ResOrErr) ? DIInliningInfo() : ResOrErr.get());
 230   } else if (ClOutputStyle == DIPrinter::OutputStyle::GNU) {
 231     // With ClPrintFunctions == FunctionNameKind::LinkageName (default)
 232     // and ClUseSymbolTable == true (also default), Symbolizer.symbolizeCode()
 233     // may override the name of an inlined function with the name of the topmost
 234     // caller function in the inlining chain. This contradicts the existing
 235     // behavior of addr2line. Symbolizer.symbolizeInlinedCode() overrides only
 236     // the topmost function, which suits our needs better.
 237     auto ResOrErr = Symbolizer.symbolizeInlinedCode(
 238         ModuleName, {Offset, object::SectionedAddress::UndefSection},
 239         ClDwpName);
 240     Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get().getFrame(0));
 241   } else {
 242     auto ResOrErr = Symbolizer.symbolizeCode(
 243         ModuleName, {Offset, object::SectionedAddress::UndefSection},
 244         ClDwpName);
 245     Printer << (error(ResOrErr) ? DILineInfo() : ResOrErr.get());
 246   }
 247   if (ClOutputStyle == DIPrinter::OutputStyle::LLVM)
 248     outs() << "\n";
 249 }
 250
 251 int main(int argc, char **argv) {
 252   InitLLVM X(argc, argv);
 253
 254   bool IsAddr2Line = sys::path::stem(argv[0]).contains("addr2line");
 255
 256   if (IsAddr2Line) {
 257     ClDemangle.setInitialValue(false);
 258     ClPrintFunctions.setInitialValue(FunctionNameKind::None);
 259     ClPrintInlining.setInitialValue(false);
 260     ClOutputStyle.setInitialValue(DIPrinter::OutputStyle::GNU);
 261   }
 262
 263   llvm::sys::InitializeCOMRAII COM(llvm::sys::COMThreadingMode::MultiThreaded);
 264   cl::ParseCommandLineOptions(argc, argv, IsAddr2Line ? "llvm-addr2line\n"
 265                                                       : "llvm-symbolizer\n");
 266
 267   // If both --demangle and --no-demangle are specified then pick the last one.
 268   if (ClNoDemangle.getPosition() > ClDemangle.getPosition())
 269     ClDemangle = !ClNoDemangle;
 270
 271   LLVMSymbolizer::Options Opts(ClPrintFunctions, ClUseSymbolTable, ClDemangle,
 272                                ClUseRelativeAddress, ClDefaultArch,
 273                                ClFallbackDebugPath);
 274
 275   for (const auto &hint : ClDsymHint) {
 276     if (sys::path::extension(hint) == ".dSYM") {
 277       Opts.DsymHints.push_back(hint);
 278     } else {
 279       errs() << "Warning: invalid dSYM hint: \"" << hint <<
 280                 "\" (must have the '.dSYM' extension).\n";
 281     }
 282   }
 283   LLVMSymbolizer Symbolizer(Opts);
 284
 285   DIPrinter Printer(outs(), ClPrintFunctions != FunctionNameKind::None,
 286                     ClPrettyPrint, ClPrintSourceContextLines, ClVerbose,
 287                     ClBasenames, ClOutputStyle);
 288
 289   if (ClInputAddresses.empty()) {
 290     const int kMaxInputStringLength = 1024;
 291     char InputString[kMaxInputStringLength];
 292
 293     while (fgets(InputString, sizeof(InputString), stdin)) {
 294       symbolizeInput(InputString, Symbolizer, Printer);
 295       outs().flush();
 296     }
 297   } else {
 298     for (StringRef Address : ClInputAddresses)
 299       symbolizeInput(Address, Symbolizer, Printer);
 300   }
 301
 302   return 0;
 303 }