libsanitizer/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cc

   1 //===-- sanitizer_symbolizer_posix_libcdep.cc -----------------------------===//
   2 //
   3 // This file is distributed under the University of Illinois Open Source
   4 // License. See LICENSE.TXT for details.
   5 //
   6 //===----------------------------------------------------------------------===//
   7 //
   8 // This file is shared between AddressSanitizer and ThreadSanitizer
   9 // run-time libraries.
  10 // POSIX-specific implementation of symbolizer parts.
  11 //===----------------------------------------------------------------------===//
  12
  13 #include "sanitizer_platform.h"
  14 #if SANITIZER_POSIX
  15 #include "sanitizer_allocator_internal.h"
  16 #include "sanitizer_common.h"
  17 #include "sanitizer_file.h"
  18 #include "sanitizer_flags.h"
  19 #include "sanitizer_internal_defs.h"
  20 #include "sanitizer_linux.h"
  21 #include "sanitizer_placement_new.h"
  22 #include "sanitizer_posix.h"
  23 #include "sanitizer_procmaps.h"
  24 #include "sanitizer_symbolizer_internal.h"
  25 #include "sanitizer_symbolizer_libbacktrace.h"
  26 #include "sanitizer_symbolizer_mac.h"
  27
  28 #include <dlfcn.h>   // for dlsym()
  29 #include <errno.h>
  30 #include <stdint.h>
  31 #include <stdlib.h>
  32 #include <sys/wait.h>
  33 #include <unistd.h>
  34
  35 #if SANITIZER_MAC
  36 #include <util.h>  // for forkpty()
  37 #endif  // SANITIZER_MAC
  38
  39 // C++ demangling function, as required by Itanium C++ ABI. This is weak,
  40 // because we do not require a C++ ABI library to be linked to a program
  41 // using sanitizers; if it's not present, we'll just use the mangled name.
  42 namespace __cxxabiv1 {
  43   extern "C" SANITIZER_WEAK_ATTRIBUTE
  44   char *__cxa_demangle(const char *mangled, char *buffer,
  45                                   size_t *length, int *status);
  46 }
  47
  48 namespace __sanitizer {
  49
  50 // Attempts to demangle the name via __cxa_demangle from __cxxabiv1.
  51 const char *DemangleCXXABI(const char *name) {
  52   // FIXME: __cxa_demangle aggressively insists on allocating memory.
  53   // There's not much we can do about that, short of providing our
  54   // own demangler (libc++abi's implementation could be adapted so that
  55   // it does not allocate). For now, we just call it anyway, and we leak
  56   // the returned value.
  57   if (&__cxxabiv1::__cxa_demangle)
  58     if (const char *demangled_name =
  59           __cxxabiv1::__cxa_demangle(name, 0, 0, 0))
  60       return demangled_name;
  61
  62   return name;
  63 }
  64
  65 // As of now, there are no headers for the Swift runtime. Once they are
  66 // present, we will weakly link since we do not require Swift runtime to be
  67 // linked.
  68 typedef char *(*swift_demangle_ft)(const char *mangledName,
  69                                    size_t mangledNameLength, char *outputBuffer,
  70                                    size_t *outputBufferSize, uint32_t flags);
  71 static swift_demangle_ft swift_demangle_f;
  72
  73 // This must not happen lazily at symbolication time, because dlsym uses
  74 // malloc and thread-local storage, which is not a good thing to do during
  75 // symbolication.
  76 static void InitializeSwiftDemangler() {
  77   swift_demangle_f = (swift_demangle_ft)dlsym(RTLD_DEFAULT, "swift_demangle");
  78 }
  79
  80 // Attempts to demangle a Swift name. The demangler will return nullptr if a
  81 // non-Swift name is passed in.
  82 const char *DemangleSwift(const char *name) {
  83   if (!name) return nullptr;
  84
  85   // Check if we are dealing with a Swift mangled name first.
  86   if (name[0] != '_' || name[1] != 'T') {
  87     return nullptr;
  88   }
  89
  90   if (swift_demangle_f)
  91     return swift_demangle_f(name, internal_strlen(name), 0, 0, 0);
  92
  93   return nullptr;
  94 }
  95
  96 const char *DemangleSwiftAndCXX(const char *name) {
  97   if (!name) return nullptr;
  98   if (const char *swift_demangled_name = DemangleSwift(name))
  99     return swift_demangled_name;
 100   return DemangleCXXABI(name);
 101 }
 102
 103 static bool CreateTwoHighNumberedPipes(int *infd_, int *outfd_) {
 104   int *infd = NULL;
 105   int *outfd = NULL;
 106   // The client program may close its stdin and/or stdout and/or stderr
 107   // thus allowing socketpair to reuse file descriptors 0, 1 or 2.
 108   // In this case the communication between the forked processes may be
 109   // broken if either the parent or the child tries to close or duplicate
 110   // these descriptors. The loop below produces two pairs of file
 111   // descriptors, each greater than 2 (stderr).
 112   int sock_pair[5][2];
 113   for (int i = 0; i < 5; i++) {
 114     if (pipe(sock_pair[i]) == -1) {
 115       for (int j = 0; j < i; j++) {
 116         internal_close(sock_pair[j][0]);
 117         internal_close(sock_pair[j][1]);
 118       }
 119       return false;
 120     } else if (sock_pair[i][0] > 2 && sock_pair[i][1] > 2) {
 121       if (infd == NULL) {
 122         infd = sock_pair[i];
 123       } else {
 124         outfd = sock_pair[i];
 125         for (int j = 0; j < i; j++) {
 126           if (sock_pair[j] == infd) continue;
 127           internal_close(sock_pair[j][0]);
 128           internal_close(sock_pair[j][1]);
 129         }
 130         break;
 131       }
 132     }
 133   }
 134   CHECK(infd);
 135   CHECK(outfd);
 136   infd_[0] = infd[0];
 137   infd_[1] = infd[1];
 138   outfd_[0] = outfd[0];
 139   outfd_[1] = outfd[1];
 140   return true;
 141 }
 142
 143 bool SymbolizerProcess::StartSymbolizerSubprocess() {
 144   if (!FileExists(path_)) {
 145     if (!reported_invalid_path_) {
 146       Report("WARNING: invalid path to external symbolizer!\n");
 147       reported_invalid_path_ = true;
 148     }
 149     return false;
 150   }
 151
 152   int pid = -1;
 153
 154   int infd[2];
 155   internal_memset(&infd, 0, sizeof(infd));
 156   int outfd[2];
 157   internal_memset(&outfd, 0, sizeof(outfd));
 158   if (!CreateTwoHighNumberedPipes(infd, outfd)) {
 159     Report("WARNING: Can't create a socket pair to start "
 160            "external symbolizer (errno: %d)\n", errno);
 161     return false;
 162   }
 163
 164   if (use_forkpty_) {
 165 #if SANITIZER_MAC
 166     fd_t fd = kInvalidFd;
 167
 168     // forkpty redirects stdout and stderr into a single stream, so we would
 169     // receive error messages as standard replies. To avoid that, let's dup
 170     // stderr and restore it in the child.
 171     int saved_stderr = dup(STDERR_FILENO);
 172     CHECK_GE(saved_stderr, 0);
 173
 174     // We only need one pipe, for stdin of the child.
 175     close(outfd[0]);
 176     close(outfd[1]);
 177
 178     // Use forkpty to disable buffering in the new terminal.
 179     pid = internal_forkpty(&fd);
 180     if (pid == -1) {
 181       // forkpty() failed.
 182       Report("WARNING: failed to fork external symbolizer (errno: %d)\n",
 183              errno);
 184       return false;
 185     } else if (pid == 0) {
 186       // Child subprocess.
 187
 188       // infd[0] is the child's reading end.
 189       close(infd[1]);
 190
 191       // Set up stdin to read from the pipe.
 192       CHECK_GE(dup2(infd[0], STDIN_FILENO), 0);
 193       close(infd[0]);
 194
 195       // Restore stderr.
 196       CHECK_GE(dup2(saved_stderr, STDERR_FILENO), 0);
 197       close(saved_stderr);
 198
 199       const char *argv[kArgVMax];
 200       GetArgV(path_, argv);
 201       execv(path_, const_cast<char **>(&argv[0]));
 202       internal__exit(1);
 203     }
 204
 205     // Input for the child, infd[1] is our writing end.
 206     output_fd_ = infd[1];
 207     close(infd[0]);
 208
 209     // Continue execution in parent process.
 210     input_fd_ = fd;
 211
 212     close(saved_stderr);
 213
 214     // Disable echo in the new terminal, disable CR.
 215     struct termios termflags;
 216     tcgetattr(fd, &termflags);
 217     termflags.c_oflag &= ~ONLCR;
 218     termflags.c_lflag &= ~ECHO;
 219     tcsetattr(fd, TCSANOW, &termflags);
 220 #else  // SANITIZER_MAC
 221     UNIMPLEMENTED();
 222 #endif  // SANITIZER_MAC
 223   } else {
 224     const char *argv[kArgVMax];
 225     GetArgV(path_, argv);
 226     pid = StartSubprocess(path_, argv, /* stdin */ outfd[0],
 227                           /* stdout */ infd[1]);
 228     if (pid < 0) {
 229       internal_close(infd[0]);
 230       internal_close(outfd[1]);
 231       return false;
 232     }
 233
 234     input_fd_ = infd[0];
 235     output_fd_ = outfd[1];
 236   }
 237
 238   CHECK_GT(pid, 0);
 239
 240   // Check that symbolizer subprocess started successfully.
 241   SleepForMillis(kSymbolizerStartupTimeMillis);
 242   if (!IsProcessRunning(pid)) {
 243     // Either waitpid failed, or child has already exited.
 244     Report("WARNING: external symbolizer didn't start up correctly!\n");
 245     return false;
 246   }
 247
 248   return true;
 249 }
 250
 251 class Addr2LineProcess : public SymbolizerProcess {
 252  public:
 253   Addr2LineProcess(const char *path, const char *module_name)
 254       : SymbolizerProcess(path), module_name_(internal_strdup(module_name)) {}
 255
 256   const char *module_name() const { return module_name_; }
 257
 258  private:
 259   void GetArgV(const char *path_to_binary,
 260                const char *(&argv)[kArgVMax]) const override {
 261     int i = 0;
 262     argv[i++] = path_to_binary;
 263     argv[i++] = "-iCfe";
 264     argv[i++] = module_name_;
 265     argv[i++] = nullptr;
 266   }
 267
 268   bool ReachedEndOfOutput(const char *buffer, uptr length) const override;
 269
 270   bool ReadFromSymbolizer(char *buffer, uptr max_length) override {
 271     if (!SymbolizerProcess::ReadFromSymbolizer(buffer, max_length))
 272       return false;
 273     // The returned buffer is empty when output is valid, but exceeds
 274     // max_length.
 275     if (*buffer == '\0')
 276       return true;
 277     // We should cut out output_terminator_ at the end of given buffer,
 278     // appended by addr2line to mark the end of its meaningful output.
 279     // We cannot scan buffer from it's beginning, because it is legal for it
 280     // to start with output_terminator_ in case given offset is invalid. So,
 281     // scanning from second character.
 282     char *garbage = internal_strstr(buffer + 1, output_terminator_);
 283     // This should never be NULL since buffer must end up with
 284     // output_terminator_.
 285     CHECK(garbage);
 286     // Trim the buffer.
 287     garbage[0] = '\0';
 288     return true;
 289   }
 290
 291   const char *module_name_;  // Owned, leaked.
 292   static const char output_terminator_[];
 293 };
 294
 295 const char Addr2LineProcess::output_terminator_[] = "??\n??:0\n";
 296
 297 bool Addr2LineProcess::ReachedEndOfOutput(const char *buffer,
 298                                           uptr length) const {
 299   const size_t kTerminatorLen = sizeof(output_terminator_) - 1;
 300   // Skip, if we read just kTerminatorLen bytes, because Addr2Line output
 301   // should consist at least of two pairs of lines:
 302   // 1. First one, corresponding to given offset to be symbolized
 303   // (may be equal to output_terminator_, if offset is not valid).
 304   // 2. Second one for output_terminator_, itself to mark the end of output.
 305   if (length <= kTerminatorLen) return false;
 306   // Addr2Line output should end up with output_terminator_.
 307   return !internal_memcmp(buffer + length - kTerminatorLen,
 308                           output_terminator_, kTerminatorLen);
 309 }
 310
 311 class Addr2LinePool : public SymbolizerTool {
 312  public:
 313   explicit Addr2LinePool(const char *addr2line_path,
 314                          LowLevelAllocator *allocator)
 315       : addr2line_path_(addr2line_path), allocator_(allocator),
 316         addr2line_pool_(16) {}
 317
 318   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 319     if (const char *buf =
 320             SendCommand(stack->info.module, stack->info.module_offset)) {
 321       ParseSymbolizePCOutput(buf, stack);
 322       return true;
 323     }
 324     return false;
 325   }
 326
 327   bool SymbolizeData(uptr addr, DataInfo *info) override {
 328     return false;
 329   }
 330
 331  private:
 332   const char *SendCommand(const char *module_name, uptr module_offset) {
 333     Addr2LineProcess *addr2line = 0;
 334     for (uptr i = 0; i < addr2line_pool_.size(); ++i) {
 335       if (0 ==
 336           internal_strcmp(module_name, addr2line_pool_[i]->module_name())) {
 337         addr2line = addr2line_pool_[i];
 338         break;
 339       }
 340     }
 341     if (!addr2line) {
 342       addr2line =
 343           new(*allocator_) Addr2LineProcess(addr2line_path_, module_name);
 344       addr2line_pool_.push_back(addr2line);
 345     }
 346     CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
 347     char buffer[kBufferSize];
 348     internal_snprintf(buffer, kBufferSize, "0x%zx\n0x%zx\n",
 349                       module_offset, dummy_address_);
 350     return addr2line->SendCommand(buffer);
 351   }
 352
 353   static const uptr kBufferSize = 64;
 354   const char *addr2line_path_;
 355   LowLevelAllocator *allocator_;
 356   InternalMmapVector<Addr2LineProcess*> addr2line_pool_;
 357   static const uptr dummy_address_ =
 358       FIRST_32_SECOND_64(UINT32_MAX, UINT64_MAX);
 359 };
 360
 361 #if SANITIZER_SUPPORTS_WEAK_HOOKS
 362 extern "C" {
 363 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 364 bool __sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
 365                                 char *Buffer, int MaxLength);
 366 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 367 bool __sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
 368                                 char *Buffer, int MaxLength);
 369 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 370 void __sanitizer_symbolize_flush();
 371 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 372 int __sanitizer_symbolize_demangle(const char *Name, char *Buffer,
 373                                    int MaxLength);
 374 }  // extern "C"
 375
 376 class InternalSymbolizer : public SymbolizerTool {
 377  public:
 378   static InternalSymbolizer *get(LowLevelAllocator *alloc) {
 379     if (__sanitizer_symbolize_code != 0 &&
 380         __sanitizer_symbolize_data != 0) {
 381       return new(*alloc) InternalSymbolizer();
 382     }
 383     return 0;
 384   }
 385
 386   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 387     bool result = __sanitizer_symbolize_code(
 388         stack->info.module, stack->info.module_offset, buffer_, kBufferSize);
 389     if (result) ParseSymbolizePCOutput(buffer_, stack);
 390     return result;
 391   }
 392
 393   bool SymbolizeData(uptr addr, DataInfo *info) override {
 394     bool result = __sanitizer_symbolize_data(info->module, info->module_offset,
 395                                              buffer_, kBufferSize);
 396     if (result) {
 397       ParseSymbolizeDataOutput(buffer_, info);
 398       info->start += (addr - info->module_offset);  // Add the base address.
 399     }
 400     return result;
 401   }
 402
 403   void Flush() override {
 404     if (__sanitizer_symbolize_flush)
 405       __sanitizer_symbolize_flush();
 406   }
 407
 408   const char *Demangle(const char *name) override {
 409     if (__sanitizer_symbolize_demangle) {
 410       for (uptr res_length = 1024;
 411            res_length <= InternalSizeClassMap::kMaxSize;) {
 412         char *res_buff = static_cast<char*>(InternalAlloc(res_length));
 413         uptr req_length =
 414             __sanitizer_symbolize_demangle(name, res_buff, res_length);
 415         if (req_length > res_length) {
 416           res_length = req_length + 1;
 417           InternalFree(res_buff);
 418           continue;
 419         }
 420         return res_buff;
 421       }
 422     }
 423     return name;
 424   }
 425
 426  private:
 427   InternalSymbolizer() { }
 428
 429   static const int kBufferSize = 16 * 1024;
 430   char buffer_[kBufferSize];
 431 };
 432 #else  // SANITIZER_SUPPORTS_WEAK_HOOKS
 433
 434 class InternalSymbolizer : public SymbolizerTool {
 435  public:
 436   static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
 437 };
 438
 439 #endif  // SANITIZER_SUPPORTS_WEAK_HOOKS
 440
 441 const char *Symbolizer::PlatformDemangle(const char *name) {
 442   return DemangleSwiftAndCXX(name);
 443 }
 444
 445 void Symbolizer::PlatformPrepareForSandboxing() {}
 446
 447 static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
 448   const char *path = common_flags()->external_symbolizer_path;
 449   const char *binary_name = path ? StripModuleName(path) : "";
 450   if (path && path[0] == '\0') {
 451     VReport(2, "External symbolizer is explicitly disabled.\n");
 452     return nullptr;
 453   } else if (!internal_strcmp(binary_name, "llvm-symbolizer")) {
 454     VReport(2, "Using llvm-symbolizer at user-specified path: %s\n", path);
 455     return new(*allocator) LLVMSymbolizer(path, allocator);
 456   } else if (!internal_strcmp(binary_name, "atos")) {
 457 #if SANITIZER_MAC
 458     VReport(2, "Using atos at user-specified path: %s\n", path);
 459     return new(*allocator) AtosSymbolizer(path, allocator);
 460 #else  // SANITIZER_MAC
 461     Report("ERROR: Using `atos` is only supported on Darwin.\n");
 462     Die();
 463 #endif  // SANITIZER_MAC
 464   } else if (!internal_strcmp(binary_name, "addr2line")) {
 465     VReport(2, "Using addr2line at user-specified path: %s\n", path);
 466     return new(*allocator) Addr2LinePool(path, allocator);
 467   } else if (path) {
 468     Report("ERROR: External symbolizer path is set to '%s' which isn't "
 469            "a known symbolizer. Please set the path to the llvm-symbolizer "
 470            "binary or other known tool.\n", path);
 471     Die();
 472   }
 473
 474   // Otherwise symbolizer program is unknown, let's search $PATH
 475   CHECK(path == nullptr);
 476 #if SANITIZER_MAC
 477   if (const char *found_path = FindPathToBinary("atos")) {
 478     VReport(2, "Using atos found at: %s\n", found_path);
 479     return new(*allocator) AtosSymbolizer(found_path, allocator);
 480   }
 481 #endif  // SANITIZER_MAC
 482   if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
 483     VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
 484     return new(*allocator) LLVMSymbolizer(found_path, allocator);
 485   }
 486   if (common_flags()->allow_addr2line) {
 487     if (const char *found_path = FindPathToBinary("addr2line")) {
 488       VReport(2, "Using addr2line found at: %s\n", found_path);
 489       return new(*allocator) Addr2LinePool(found_path, allocator);
 490     }
 491   }
 492   return nullptr;
 493 }
 494
 495 static void ChooseSymbolizerTools(IntrusiveList<SymbolizerTool> *list,
 496                                   LowLevelAllocator *allocator) {
 497   if (!common_flags()->symbolize) {
 498     VReport(2, "Symbolizer is disabled.\n");
 499     return;
 500   }
 501   if (IsAllocatorOutOfMemory()) {
 502     VReport(2, "Cannot use internal symbolizer: out of memory\n");
 503   } else if (SymbolizerTool *tool = InternalSymbolizer::get(allocator)) {
 504     VReport(2, "Using internal symbolizer.\n");
 505     list->push_back(tool);
 506     return;
 507   }
 508   if (SymbolizerTool *tool = LibbacktraceSymbolizer::get(allocator)) {
 509     VReport(2, "Using libbacktrace symbolizer.\n");
 510     list->push_back(tool);
 511     return;
 512   }
 513
 514   if (SymbolizerTool *tool = ChooseExternalSymbolizer(allocator)) {
 515     list->push_back(tool);
 516   }
 517
 518 #if SANITIZER_MAC
 519   VReport(2, "Using dladdr symbolizer.\n");
 520   list->push_back(new(*allocator) DlAddrSymbolizer());
 521 #endif  // SANITIZER_MAC
 522 }
 523
 524 Symbolizer *Symbolizer::PlatformInit() {
 525   IntrusiveList<SymbolizerTool> list;
 526   list.clear();
 527   ChooseSymbolizerTools(&list, &symbolizer_allocator_);
 528   return new(symbolizer_allocator_) Symbolizer(list);
 529 }
 530
 531 void Symbolizer::LateInitialize() {
 532   Symbolizer::GetOrInit();
 533   InitializeSwiftDemangler();
 534 }
 535
 536 }  // namespace __sanitizer
 537
 538 #endif  // SANITIZER_POSIX