compiler-rt/lib/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 }
  76
  77 // Attempts to demangle a Swift name. The demangler will return nullptr if a
  78 // non-Swift name is passed in.
  79 const char *DemangleSwift(const char *name) {
  80   if (swift_demangle_f)
  81     return swift_demangle_f(name, internal_strlen(name), 0, 0, 0);
  82
  83   return nullptr;
  84 }
  85
  86 const char *DemangleSwiftAndCXX(const char *name) {
  87   if (!name) return nullptr;
  88   if (const char *swift_demangled_name = DemangleSwift(name))
  89     return swift_demangled_name;
  90   return DemangleCXXABI(name);
  91 }
  92
  93 static bool CreateTwoHighNumberedPipes(int *infd_, int *outfd_) {
  94   int *infd = NULL;
  95   int *outfd = NULL;
  96   // The client program may close its stdin and/or stdout and/or stderr
  97   // thus allowing socketpair to reuse file descriptors 0, 1 or 2.
  98   // In this case the communication between the forked processes may be
  99   // broken if either the parent or the child tries to close or duplicate
 100   // these descriptors. The loop below produces two pairs of file
 101   // descriptors, each greater than 2 (stderr).
 102   int sock_pair[5][2];
 103   for (int i = 0; i < 5; i++) {
 104     if (pipe(sock_pair[i]) == -1) {
 105       for (int j = 0; j < i; j++) {
 106         internal_close(sock_pair[j][0]);
 107         internal_close(sock_pair[j][1]);
 108       }
 109       return false;
 110     } else if (sock_pair[i][0] > 2 && sock_pair[i][1] > 2) {
 111       if (infd == NULL) {
 112         infd = sock_pair[i];
 113       } else {
 114         outfd = sock_pair[i];
 115         for (int j = 0; j < i; j++) {
 116           if (sock_pair[j] == infd) continue;
 117           internal_close(sock_pair[j][0]);
 118           internal_close(sock_pair[j][1]);
 119         }
 120         break;
 121       }
 122     }
 123   }
 124   CHECK(infd);
 125   CHECK(outfd);
 126   infd_[0] = infd[0];
 127   infd_[1] = infd[1];
 128   outfd_[0] = outfd[0];
 129   outfd_[1] = outfd[1];
 130   return true;
 131 }
 132
 133 bool SymbolizerProcess::StartSymbolizerSubprocess() {
 134   if (!FileExists(path_)) {
 135     if (!reported_invalid_path_) {
 136       Report("WARNING: invalid path to external symbolizer!\n");
 137       reported_invalid_path_ = true;
 138     }
 139     return false;
 140   }
 141
 142   const char *argv[kArgVMax];
 143   GetArgV(path_, argv);
 144   pid_t pid;
 145
 146   // Report how symbolizer is being launched for debugging purposes.
 147   if (Verbosity() >= 3) {
 148     // Only use `Report` for first line so subsequent prints don't get prefixed
 149     // with current PID.
 150     Report("Launching Symbolizer process: ");
 151     for (unsigned index = 0; index < kArgVMax && argv[index]; ++index)
 152       Printf("%s ", argv[index]);
 153     Printf("\n");
 154   }
 155
 156   if (use_posix_spawn_) {
 157 #if SANITIZER_APPLE
 158     fd_t fd = internal_spawn(argv, const_cast<const char **>(GetEnvP()), &pid);
 159     if (fd == kInvalidFd) {
 160       Report("WARNING: failed to spawn external symbolizer (errno: %d)\n",
 161              errno);
 162       return false;
 163     }
 164
 165     input_fd_ = fd;
 166     output_fd_ = fd;
 167 #else  // SANITIZER_APPLE
 168     UNIMPLEMENTED();
 169 #endif  // SANITIZER_APPLE
 170   } else {
 171     fd_t infd[2] = {}, outfd[2] = {};
 172     if (!CreateTwoHighNumberedPipes(infd, outfd)) {
 173       Report("WARNING: Can't create a socket pair to start "
 174              "external symbolizer (errno: %d)\n", errno);
 175       return false;
 176     }
 177
 178     pid = StartSubprocess(path_, argv, GetEnvP(), /* stdin */ outfd[0],
 179                           /* stdout */ infd[1]);
 180     if (pid < 0) {
 181       internal_close(infd[0]);
 182       internal_close(outfd[1]);
 183       return false;
 184     }
 185
 186     input_fd_ = infd[0];
 187     output_fd_ = outfd[1];
 188   }
 189
 190   CHECK_GT(pid, 0);
 191
 192   // Check that symbolizer subprocess started successfully.
 193   SleepForMillis(kSymbolizerStartupTimeMillis);
 194   if (!IsProcessRunning(pid)) {
 195     // Either waitpid failed, or child has already exited.
 196     Report("WARNING: external symbolizer didn't start up correctly!\n");
 197     return false;
 198   }
 199
 200   return true;
 201 }
 202
 203 class Addr2LineProcess final : public SymbolizerProcess {
 204  public:
 205   Addr2LineProcess(const char *path, const char *module_name)
 206       : SymbolizerProcess(path), module_name_(internal_strdup(module_name)) {}
 207
 208   const char *module_name() const { return module_name_; }
 209
 210  private:
 211   void GetArgV(const char *path_to_binary,
 212                const char *(&argv)[kArgVMax]) const override {
 213     int i = 0;
 214     argv[i++] = path_to_binary;
 215     if (common_flags()->demangle)
 216       argv[i++] = "-C";
 217     if (common_flags()->symbolize_inline_frames)
 218       argv[i++] = "-i";
 219     argv[i++] = "-fe";
 220     argv[i++] = module_name_;
 221     argv[i++] = nullptr;
 222     CHECK_LE(i, kArgVMax);
 223   }
 224
 225   bool ReachedEndOfOutput(const char *buffer, uptr length) const override;
 226
 227   bool ReadFromSymbolizer() override {
 228     if (!SymbolizerProcess::ReadFromSymbolizer())
 229       return false;
 230     auto &buff = GetBuff();
 231     // We should cut out output_terminator_ at the end of given buffer,
 232     // appended by addr2line to mark the end of its meaningful output.
 233     // We cannot scan buffer from it's beginning, because it is legal for it
 234     // to start with output_terminator_ in case given offset is invalid. So,
 235     // scanning from second character.
 236     char *garbage = internal_strstr(buff.data() + 1, output_terminator_);
 237     // This should never be NULL since buffer must end up with
 238     // output_terminator_.
 239     CHECK(garbage);
 240
 241     // Trim the buffer.
 242     uintptr_t new_size = garbage - buff.data();
 243     GetBuff().resize(new_size);
 244     GetBuff().push_back('\0');
 245     return true;
 246   }
 247
 248   const char *module_name_;  // Owned, leaked.
 249   static const char output_terminator_[];
 250 };
 251
 252 const char Addr2LineProcess::output_terminator_[] = "??\n??:0\n";
 253
 254 bool Addr2LineProcess::ReachedEndOfOutput(const char *buffer,
 255                                           uptr length) const {
 256   const size_t kTerminatorLen = sizeof(output_terminator_) - 1;
 257   // Skip, if we read just kTerminatorLen bytes, because Addr2Line output
 258   // should consist at least of two pairs of lines:
 259   // 1. First one, corresponding to given offset to be symbolized
 260   // (may be equal to output_terminator_, if offset is not valid).
 261   // 2. Second one for output_terminator_, itself to mark the end of output.
 262   if (length <= kTerminatorLen) return false;
 263   // Addr2Line output should end up with output_terminator_.
 264   return !internal_memcmp(buffer + length - kTerminatorLen,
 265                           output_terminator_, kTerminatorLen);
 266 }
 267
 268 class Addr2LinePool final : public SymbolizerTool {
 269  public:
 270   explicit Addr2LinePool(const char *addr2line_path,
 271                          LowLevelAllocator *allocator)
 272       : addr2line_path_(addr2line_path), allocator_(allocator) {
 273     addr2line_pool_.reserve(16);
 274   }
 275
 276   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 277     if (const char *buf =
 278             SendCommand(stack->info.module, stack->info.module_offset)) {
 279       ParseSymbolizePCOutput(buf, stack);
 280       return true;
 281     }
 282     return false;
 283   }
 284
 285   bool SymbolizeData(uptr addr, DataInfo *info) override {
 286     return false;
 287   }
 288
 289  private:
 290   const char *SendCommand(const char *module_name, uptr module_offset) {
 291     Addr2LineProcess *addr2line = 0;
 292     for (uptr i = 0; i < addr2line_pool_.size(); ++i) {
 293       if (0 ==
 294           internal_strcmp(module_name, addr2line_pool_[i]->module_name())) {
 295         addr2line = addr2line_pool_[i];
 296         break;
 297       }
 298     }
 299     if (!addr2line) {
 300       addr2line =
 301           new(*allocator_) Addr2LineProcess(addr2line_path_, module_name);
 302       addr2line_pool_.push_back(addr2line);
 303     }
 304     CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
 305     char buffer[kBufferSize];
 306     internal_snprintf(buffer, kBufferSize, "0x%zx\n0x%zx\n",
 307                       module_offset, dummy_address_);
 308     return addr2line->SendCommand(buffer);
 309   }
 310
 311   static const uptr kBufferSize = 64;
 312   const char *addr2line_path_;
 313   LowLevelAllocator *allocator_;
 314   InternalMmapVector<Addr2LineProcess*> addr2line_pool_;
 315   static const uptr dummy_address_ =
 316       FIRST_32_SECOND_64(UINT32_MAX, UINT64_MAX);
 317 };
 318
 319 #  if SANITIZER_SUPPORTS_WEAK_HOOKS
 320 extern "C" {
 321 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 322 __sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
 323                            char *Buffer, int MaxLength);
 324 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 325 __sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
 326                            char *Buffer, int MaxLength);
 327 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
 328 __sanitizer_symbolize_flush();
 329 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE int
 330 __sanitizer_symbolize_demangle(const char *Name, char *Buffer, int MaxLength);
 331 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 332 __sanitizer_symbolize_set_demangle(bool Demangle);
 333 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 334 __sanitizer_symbolize_set_inline_frames(bool InlineFrames);
 335 }  // extern "C"
 336
 337 class InternalSymbolizer final : public SymbolizerTool {
 338  public:
 339   static InternalSymbolizer *get(LowLevelAllocator *alloc) {
 340     if (__sanitizer_symbolize_set_demangle)
 341       CHECK(__sanitizer_symbolize_set_demangle(common_flags()->demangle));
 342     if (__sanitizer_symbolize_set_inline_frames)
 343       CHECK(__sanitizer_symbolize_set_inline_frames(
 344           common_flags()->symbolize_inline_frames));
 345     if (__sanitizer_symbolize_code && __sanitizer_symbolize_data)
 346       return new (*alloc) InternalSymbolizer();
 347     return 0;
 348   }
 349
 350   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 351     bool result = __sanitizer_symbolize_code(
 352         stack->info.module, stack->info.module_offset, buffer_, kBufferSize);
 353     if (result)
 354       ParseSymbolizePCOutput(buffer_, stack);
 355     return result;
 356   }
 357
 358   bool SymbolizeData(uptr addr, DataInfo *info) override {
 359     bool result = __sanitizer_symbolize_data(info->module, info->module_offset,
 360                                              buffer_, kBufferSize);
 361     if (result) {
 362       ParseSymbolizeDataOutput(buffer_, info);
 363       info->start += (addr - info->module_offset);  // Add the base address.
 364     }
 365     return result;
 366   }
 367
 368   void Flush() override {
 369     if (__sanitizer_symbolize_flush)
 370       __sanitizer_symbolize_flush();
 371   }
 372
 373   const char *Demangle(const char *name) override {
 374     if (__sanitizer_symbolize_demangle) {
 375       for (uptr res_length = 1024;
 376            res_length <= InternalSizeClassMap::kMaxSize;) {
 377         char *res_buff = static_cast<char *>(InternalAlloc(res_length));
 378         uptr req_length =
 379             __sanitizer_symbolize_demangle(name, res_buff, res_length);
 380         if (req_length > res_length) {
 381           res_length = req_length + 1;
 382           InternalFree(res_buff);
 383           continue;
 384         }
 385         return res_buff;
 386       }
 387     }
 388     return name;
 389   }
 390
 391  private:
 392   InternalSymbolizer() {}
 393
 394   static const int kBufferSize = 16 * 1024;
 395   char buffer_[kBufferSize];
 396 };
 397 #  else  // SANITIZER_SUPPORTS_WEAK_HOOKS
 398
 399 class InternalSymbolizer final : public SymbolizerTool {
 400  public:
 401   static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
 402 };
 403
 404 #  endif  // SANITIZER_SUPPORTS_WEAK_HOOKS
 405
 406 const char *Symbolizer::PlatformDemangle(const char *name) {
 407   return DemangleSwiftAndCXX(name);
 408 }
 409
 410 static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
 411   const char *path = common_flags()->external_symbolizer_path;
 412
 413   if (path && internal_strchr(path, '%')) {
 414     char *new_path = (char *)InternalAlloc(kMaxPathLength);
 415     SubstituteForFlagValue(path, new_path, kMaxPathLength);
 416     path = new_path;
 417   }
 418
 419   const char *binary_name = path ? StripModuleName(path) : "";
 420   static const char kLLVMSymbolizerPrefix[] = "llvm-symbolizer";
 421   if (path && path[0] == '\0') {
 422     VReport(2, "External symbolizer is explicitly disabled.\n");
 423     return nullptr;
 424   } else if (!internal_strncmp(binary_name, kLLVMSymbolizerPrefix,
 425                                internal_strlen(kLLVMSymbolizerPrefix))) {
 426     VReport(2, "Using llvm-symbolizer at user-specified path: %s\n", path);
 427     return new(*allocator) LLVMSymbolizer(path, allocator);
 428   } else if (!internal_strcmp(binary_name, "atos")) {
 429 #if SANITIZER_APPLE
 430     VReport(2, "Using atos at user-specified path: %s\n", path);
 431     return new(*allocator) AtosSymbolizer(path, allocator);
 432 #else  // SANITIZER_APPLE
 433     Report("ERROR: Using `atos` is only supported on Darwin.\n");
 434     Die();
 435 #endif  // SANITIZER_APPLE
 436   } else if (!internal_strcmp(binary_name, "addr2line")) {
 437     VReport(2, "Using addr2line at user-specified path: %s\n", path);
 438     return new(*allocator) Addr2LinePool(path, allocator);
 439   } else if (path) {
 440     Report("ERROR: External symbolizer path is set to '%s' which isn't "
 441            "a known symbolizer. Please set the path to the llvm-symbolizer "
 442            "binary or other known tool.\n", path);
 443     Die();
 444   }
 445
 446   // Otherwise symbolizer program is unknown, let's search $PATH
 447   CHECK(path == nullptr);
 448 #if SANITIZER_APPLE
 449   if (const char *found_path = FindPathToBinary("atos")) {
 450     VReport(2, "Using atos found at: %s\n", found_path);
 451     return new(*allocator) AtosSymbolizer(found_path, allocator);
 452   }
 453 #endif  // SANITIZER_APPLE
 454   if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
 455     VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
 456     return new(*allocator) LLVMSymbolizer(found_path, allocator);
 457   }
 458   if (common_flags()->allow_addr2line) {
 459     if (const char *found_path = FindPathToBinary("addr2line")) {
 460       VReport(2, "Using addr2line found at: %s\n", found_path);
 461       return new(*allocator) Addr2LinePool(found_path, allocator);
 462     }
 463   }
 464   return nullptr;
 465 }
 466
 467 static void ChooseSymbolizerTools(IntrusiveList<SymbolizerTool> *list,
 468                                   LowLevelAllocator *allocator) {
 469   if (!common_flags()->symbolize) {
 470     VReport(2, "Symbolizer is disabled.\n");
 471     return;
 472   }
 473   if (IsAllocatorOutOfMemory()) {
 474     VReport(2, "Cannot use internal symbolizer: out of memory\n");
 475   } else if (SymbolizerTool *tool = InternalSymbolizer::get(allocator)) {
 476     VReport(2, "Using internal symbolizer.\n");
 477     list->push_back(tool);
 478     return;
 479   }
 480   if (SymbolizerTool *tool = LibbacktraceSymbolizer::get(allocator)) {
 481     VReport(2, "Using libbacktrace symbolizer.\n");
 482     list->push_back(tool);
 483     return;
 484   }
 485
 486   if (SymbolizerTool *tool = ChooseExternalSymbolizer(allocator)) {
 487     list->push_back(tool);
 488   }
 489
 490 #if SANITIZER_APPLE
 491   VReport(2, "Using dladdr symbolizer.\n");
 492   list->push_back(new(*allocator) DlAddrSymbolizer());
 493 #endif  // SANITIZER_APPLE
 494 }
 495
 496 Symbolizer *Symbolizer::PlatformInit() {
 497   IntrusiveList<SymbolizerTool> list;
 498   list.clear();
 499   ChooseSymbolizerTools(&list, &symbolizer_allocator_);
 500   return new(symbolizer_allocator_) Symbolizer(list);
 501 }
 502
 503 void Symbolizer::LateInitialize() {
 504   Symbolizer::GetOrInit();
 505   InitializeSwiftDemangler();
 506 }
 507
 508 }  // namespace __sanitizer
 509
 510 #endif  // SANITIZER_POSIX