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