libsanitizer/sanitizer_common/sanitizer_symbolizer_posix_libcdep.cpp

   1 //===-- sanitizer_symbolizer_posix_libcdep.cpp ----------------------------===//
   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 file is shared between AddressSanitizer and ThreadSanitizer
  10 // run-time libraries.
  11 // POSIX-specific implementation of symbolizer parts.
  12 //===----------------------------------------------------------------------===//
  13
  14 #include "sanitizer_platform.h"
  15 #if SANITIZER_POSIX
  16 #include "sanitizer_allocator_internal.h"
  17 #include "sanitizer_common.h"
  18 #include "sanitizer_file.h"
  19 #include "sanitizer_flags.h"
  20 #include "sanitizer_internal_defs.h"
  21 #include "sanitizer_linux.h"
  22 #include "sanitizer_placement_new.h"
  23 #include "sanitizer_posix.h"
  24 #include "sanitizer_procmaps.h"
  25 #include "sanitizer_symbolizer_internal.h"
  26 #include "sanitizer_symbolizer_libbacktrace.h"
  27 #include "sanitizer_symbolizer_mac.h"
  28
  29 #include <dlfcn.h>   // for dlsym()
  30 #include <errno.h>
  31 #include <stdint.h>
  32 #include <stdlib.h>
  33 #include <sys/wait.h>
  34 #include <unistd.h>
  35
  36 // C++ demangling function, as required by Itanium C++ ABI. This is weak,
  37 // because we do not require a C++ ABI library to be linked to a program
  38 // using sanitizers; if it's not present, we'll just use the mangled name.
  39 namespace __cxxabiv1 {
  40   extern "C" SANITIZER_WEAK_ATTRIBUTE
  41   char *__cxa_demangle(const char *mangled, char *buffer,
  42                                   size_t *length, int *status);
  43 }
  44
  45 namespace __sanitizer {
  46
  47 // Attempts to demangle the name via __cxa_demangle from __cxxabiv1.
  48 const char *DemangleCXXABI(const char *name) {
  49   // FIXME: __cxa_demangle aggressively insists on allocating memory.
  50   // There's not much we can do about that, short of providing our
  51   // own demangler (libc++abi's implementation could be adapted so that
  52   // it does not allocate). For now, we just call it anyway, and we leak
  53   // the returned value.
  54   if (&__cxxabiv1::__cxa_demangle)
  55     if (const char *demangled_name =
  56           __cxxabiv1::__cxa_demangle(name, 0, 0, 0))
  57       return demangled_name;
  58
  59   return name;
  60 }
  61
  62 // As of now, there are no headers for the Swift runtime. Once they are
  63 // present, we will weakly link since we do not require Swift runtime to be
  64 // linked.
  65 typedef char *(*swift_demangle_ft)(const char *mangledName,
  66                                    size_t mangledNameLength, char *outputBuffer,
  67                                    size_t *outputBufferSize, uint32_t flags);
  68 static swift_demangle_ft swift_demangle_f;
  69
  70 // This must not happen lazily at symbolication time, because dlsym uses
  71 // malloc and thread-local storage, which is not a good thing to do during
  72 // symbolication.
  73 static void InitializeSwiftDemangler() {
  74   swift_demangle_f = (swift_demangle_ft)dlsym(RTLD_DEFAULT, "swift_demangle");
  75   (void)dlerror(); // Cleanup error message in case of failure
  76 }
  77
  78 // Attempts to demangle a Swift name. The demangler will return nullptr if a
  79 // non-Swift name is passed in.
  80 const char *DemangleSwift(const char *name) {
  81   if (swift_demangle_f)
  82     return swift_demangle_f(name, internal_strlen(name), 0, 0, 0);
  83
  84   return nullptr;
  85 }
  86
  87 const char *DemangleSwiftAndCXX(const char *name) {
  88   if (!name) return nullptr;
  89   if (const char *swift_demangled_name = DemangleSwift(name))
  90     return swift_demangled_name;
  91   return DemangleCXXABI(name);
  92 }
  93
  94 static bool CreateTwoHighNumberedPipes(int *infd_, int *outfd_) {
  95   int *infd = NULL;
  96   int *outfd = NULL;
  97   // The client program may close its stdin and/or stdout and/or stderr
  98   // thus allowing socketpair to reuse file descriptors 0, 1 or 2.
  99   // In this case the communication between the forked processes may be
 100   // broken if either the parent or the child tries to close or duplicate
 101   // these descriptors. The loop below produces two pairs of file
 102   // descriptors, each greater than 2 (stderr).
 103   int sock_pair[5][2];
 104   for (int i = 0; i < 5; i++) {
 105     if (pipe(sock_pair[i]) == -1) {
 106       for (int j = 0; j < i; j++) {
 107         internal_close(sock_pair[j][0]);
 108         internal_close(sock_pair[j][1]);
 109       }
 110       return false;
 111     } else if (sock_pair[i][0] > 2 && sock_pair[i][1] > 2) {
 112       if (infd == NULL) {
 113         infd = sock_pair[i];
 114       } else {
 115         outfd = sock_pair[i];
 116         for (int j = 0; j < i; j++) {
 117           if (sock_pair[j] == infd) continue;
 118           internal_close(sock_pair[j][0]);
 119           internal_close(sock_pair[j][1]);
 120         }
 121         break;
 122       }
 123     }
 124   }
 125   CHECK(infd);
 126   CHECK(outfd);
 127   infd_[0] = infd[0];
 128   infd_[1] = infd[1];
 129   outfd_[0] = outfd[0];
 130   outfd_[1] = outfd[1];
 131   return true;
 132 }
 133
 134 bool SymbolizerProcess::StartSymbolizerSubprocess() {
 135   if (!FileExists(path_)) {
 136     if (!reported_invalid_path_) {
 137       Report("WARNING: invalid path to external symbolizer!\n");
 138       reported_invalid_path_ = true;
 139     }
 140     return false;
 141   }
 142
 143   const char *argv[kArgVMax];
 144   GetArgV(path_, argv);
 145   pid_t pid;
 146
 147   // Report how symbolizer is being launched for debugging purposes.
 148   if (Verbosity() >= 3) {
 149     // Only use `Report` for first line so subsequent prints don't get prefixed
 150     // with current PID.
 151     Report("Launching Symbolizer process: ");
 152     for (unsigned index = 0; index < kArgVMax && argv[index]; ++index)
 153       Printf("%s ", argv[index]);
 154     Printf("\n");
 155   }
 156
 157   if (use_posix_spawn_) {
 158 #if SANITIZER_MAC
 159     fd_t fd = internal_spawn(argv, const_cast<const char **>(GetEnvP()), &pid);
 160     if (fd == kInvalidFd) {
 161       Report("WARNING: failed to spawn external symbolizer (errno: %d)\n",
 162              errno);
 163       return false;
 164     }
 165
 166     input_fd_ = fd;
 167     output_fd_ = fd;
 168 #else  // SANITIZER_MAC
 169     UNIMPLEMENTED();
 170 #endif  // SANITIZER_MAC
 171   } else {
 172     fd_t infd[2] = {}, outfd[2] = {};
 173     if (!CreateTwoHighNumberedPipes(infd, outfd)) {
 174       Report("WARNING: Can't create a socket pair to start "
 175              "external symbolizer (errno: %d)\n", errno);
 176       return false;
 177     }
 178
 179     pid = StartSubprocess(path_, argv, GetEnvP(), /* stdin */ outfd[0],
 180                           /* stdout */ infd[1]);
 181     if (pid < 0) {
 182       internal_close(infd[0]);
 183       internal_close(outfd[1]);
 184       return false;
 185     }
 186
 187     input_fd_ = infd[0];
 188     output_fd_ = outfd[1];
 189   }
 190
 191   CHECK_GT(pid, 0);
 192
 193   // Check that symbolizer subprocess started successfully.
 194   SleepForMillis(kSymbolizerStartupTimeMillis);
 195   if (!IsProcessRunning(pid)) {
 196     // Either waitpid failed, or child has already exited.
 197     Report("WARNING: external symbolizer didn't start up correctly!\n");
 198     return false;
 199   }
 200
 201   return true;
 202 }
 203
 204 class Addr2LineProcess final : public SymbolizerProcess {
 205  public:
 206   Addr2LineProcess(const char *path, const char *module_name)
 207       : SymbolizerProcess(path), module_name_(internal_strdup(module_name)) {}
 208
 209   const char *module_name() const { return module_name_; }
 210
 211  private:
 212   void GetArgV(const char *path_to_binary,
 213                const char *(&argv)[kArgVMax]) const override {
 214     int i = 0;
 215     argv[i++] = path_to_binary;
 216     argv[i++] = "-iCfe";
 217     argv[i++] = module_name_;
 218     argv[i++] = nullptr;
 219   }
 220
 221   bool ReachedEndOfOutput(const char *buffer, uptr length) const override;
 222
 223   bool ReadFromSymbolizer(char *buffer, uptr max_length) override {
 224     if (!SymbolizerProcess::ReadFromSymbolizer(buffer, max_length))
 225       return false;
 226     // The returned buffer is empty when output is valid, but exceeds
 227     // max_length.
 228     if (*buffer == '\0')
 229       return true;
 230     // We should cut out output_terminator_ at the end of given buffer,
 231     // appended by addr2line to mark the end of its meaningful output.
 232     // We cannot scan buffer from it's beginning, because it is legal for it
 233     // to start with output_terminator_ in case given offset is invalid. So,
 234     // scanning from second character.
 235     char *garbage = internal_strstr(buffer + 1, output_terminator_);
 236     // This should never be NULL since buffer must end up with
 237     // output_terminator_.
 238     CHECK(garbage);
 239     // Trim the buffer.
 240     garbage[0] = '\0';
 241     return true;
 242   }
 243
 244   const char *module_name_;  // Owned, leaked.
 245   static const char output_terminator_[];
 246 };
 247
 248 const char Addr2LineProcess::output_terminator_[] = "??\n??:0\n";
 249
 250 bool Addr2LineProcess::ReachedEndOfOutput(const char *buffer,
 251                                           uptr length) const {
 252   const size_t kTerminatorLen = sizeof(output_terminator_) - 1;
 253   // Skip, if we read just kTerminatorLen bytes, because Addr2Line output
 254   // should consist at least of two pairs of lines:
 255   // 1. First one, corresponding to given offset to be symbolized
 256   // (may be equal to output_terminator_, if offset is not valid).
 257   // 2. Second one for output_terminator_, itself to mark the end of output.
 258   if (length <= kTerminatorLen) return false;
 259   // Addr2Line output should end up with output_terminator_.
 260   return !internal_memcmp(buffer + length - kTerminatorLen,
 261                           output_terminator_, kTerminatorLen);
 262 }
 263
 264 class Addr2LinePool final : public SymbolizerTool {
 265  public:
 266   explicit Addr2LinePool(const char *addr2line_path,
 267                          LowLevelAllocator *allocator)
 268       : addr2line_path_(addr2line_path), allocator_(allocator) {
 269     addr2line_pool_.reserve(16);
 270   }
 271
 272   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 273     if (const char *buf =
 274             SendCommand(stack->info.module, stack->info.module_offset)) {
 275       ParseSymbolizePCOutput(buf, stack);
 276       return true;
 277     }
 278     return false;
 279   }
 280
 281   bool SymbolizeData(uptr addr, DataInfo *info) override {
 282     return false;
 283   }
 284
 285  private:
 286   const char *SendCommand(const char *module_name, uptr module_offset) {
 287     Addr2LineProcess *addr2line = 0;
 288     for (uptr i = 0; i < addr2line_pool_.size(); ++i) {
 289       if (0 ==
 290           internal_strcmp(module_name, addr2line_pool_[i]->module_name())) {
 291         addr2line = addr2line_pool_[i];
 292         break;
 293       }
 294     }
 295     if (!addr2line) {
 296       addr2line =
 297           new(*allocator_) Addr2LineProcess(addr2line_path_, module_name);
 298       addr2line_pool_.push_back(addr2line);
 299     }
 300     CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
 301     char buffer[kBufferSize];
 302     internal_snprintf(buffer, kBufferSize, "0x%zx\n0x%zx\n",
 303                       module_offset, dummy_address_);
 304     return addr2line->SendCommand(buffer);
 305   }
 306
 307   static const uptr kBufferSize = 64;
 308   const char *addr2line_path_;
 309   LowLevelAllocator *allocator_;
 310   InternalMmapVector<Addr2LineProcess*> addr2line_pool_;
 311   static const uptr dummy_address_ =
 312       FIRST_32_SECOND_64(UINT32_MAX, UINT64_MAX);
 313 };
 314
 315 #if SANITIZER_SUPPORTS_WEAK_HOOKS
 316 extern "C" {
 317 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 318 __sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
 319                            char *Buffer, int MaxLength,
 320                            bool SymbolizeInlineFrames);
 321 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 322 bool __sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
 323                                 char *Buffer, int MaxLength);
 324 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 325 void __sanitizer_symbolize_flush();
 326 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
 327 int __sanitizer_symbolize_demangle(const char *Name, char *Buffer,
 328                                    int MaxLength);
 329 }  // extern "C"
 330
 331 class InternalSymbolizer final : public SymbolizerTool {
 332  public:
 333   static InternalSymbolizer *get(LowLevelAllocator *alloc) {
 334     if (__sanitizer_symbolize_code != 0 &&
 335         __sanitizer_symbolize_data != 0) {
 336       return new(*alloc) InternalSymbolizer();
 337     }
 338     return 0;
 339   }
 340
 341   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 342     bool result = __sanitizer_symbolize_code(
 343         stack->info.module, stack->info.module_offset, buffer_, kBufferSize,
 344         common_flags()->symbolize_inline_frames);
 345     if (result) ParseSymbolizePCOutput(buffer_, stack);
 346     return result;
 347   }
 348
 349   bool SymbolizeData(uptr addr, DataInfo *info) override {
 350     bool result = __sanitizer_symbolize_data(info->module, info->module_offset,
 351                                              buffer_, kBufferSize);
 352     if (result) {
 353       ParseSymbolizeDataOutput(buffer_, info);
 354       info->start += (addr - info->module_offset);  // Add the base address.
 355     }
 356     return result;
 357   }
 358
 359   void Flush() override {
 360     if (__sanitizer_symbolize_flush)
 361       __sanitizer_symbolize_flush();
 362   }
 363
 364   const char *Demangle(const char *name) override {
 365     if (__sanitizer_symbolize_demangle) {
 366       for (uptr res_length = 1024;
 367            res_length <= InternalSizeClassMap::kMaxSize;) {
 368         char *res_buff = static_cast<char*>(InternalAlloc(res_length));
 369         uptr req_length =
 370             __sanitizer_symbolize_demangle(name, res_buff, res_length);
 371         if (req_length > res_length) {
 372           res_length = req_length + 1;
 373           InternalFree(res_buff);
 374           continue;
 375         }
 376         return res_buff;
 377       }
 378     }
 379     return name;
 380   }
 381
 382  private:
 383   InternalSymbolizer() { }
 384
 385   static const int kBufferSize = 16 * 1024;
 386   char buffer_[kBufferSize];
 387 };
 388 #else  // SANITIZER_SUPPORTS_WEAK_HOOKS
 389
 390 class InternalSymbolizer final : public SymbolizerTool {
 391  public:
 392   static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
 393 };
 394
 395 #endif  // SANITIZER_SUPPORTS_WEAK_HOOKS
 396
 397 const char *Symbolizer::PlatformDemangle(const char *name) {
 398   return DemangleSwiftAndCXX(name);
 399 }
 400
 401 static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
 402   const char *path = common_flags()->external_symbolizer_path;
 403
 404   if (path && internal_strchr(path, '%')) {
 405     char *new_path = (char *)InternalAlloc(kMaxPathLength);
 406     SubstituteForFlagValue(path, new_path, kMaxPathLength);
 407     path = new_path;
 408   }
 409
 410   const char *binary_name = path ? StripModuleName(path) : "";
 411   static const char kLLVMSymbolizerPrefix[] = "llvm-symbolizer";
 412   if (path && path[0] == '\0') {
 413     VReport(2, "External symbolizer is explicitly disabled.\n");
 414     return nullptr;
 415   } else if (!internal_strncmp(binary_name, kLLVMSymbolizerPrefix,
 416                                internal_strlen(kLLVMSymbolizerPrefix))) {
 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 SANITIZER_MAC
 440   if (const char *found_path = FindPathToBinary("atos")) {
 441     VReport(2, "Using atos found at: %s\n", found_path);
 442     return new(*allocator) AtosSymbolizer(found_path, allocator);
 443   }
 444 #endif  // SANITIZER_MAC
 445   if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
 446     VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
 447     return new(*allocator) LLVMSymbolizer(found_path, allocator);
 448   }
 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 (IsAllocatorOutOfMemory()) {
 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()->LateInitializeTools();
 496   InitializeSwiftDemangler();
 497 }
 498
 499 }  // namespace __sanitizer
 500
 501 #endif  // SANITIZER_POSIX