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 #include "sanitizer_symbolizer_markup.h"
  16 #if SANITIZER_POSIX
  17 #  include <dlfcn.h>  // for dlsym()
  18 #  include <errno.h>
  19 #  include <stdint.h>
  20 #  include <stdlib.h>
  21 #  include <sys/wait.h>
  22 #  include <unistd.h>
  23
  24 #  include "sanitizer_allocator_internal.h"
  25 #  include "sanitizer_common.h"
  26 #  include "sanitizer_file.h"
  27 #  include "sanitizer_flags.h"
  28 #  include "sanitizer_internal_defs.h"
  29 #  include "sanitizer_linux.h"
  30 #  include "sanitizer_placement_new.h"
  31 #  include "sanitizer_posix.h"
  32 #  include "sanitizer_procmaps.h"
  33 #  include "sanitizer_symbolizer_internal.h"
  34 #  include "sanitizer_symbolizer_libbacktrace.h"
  35 #  include "sanitizer_symbolizer_mac.h"
  36
  37 // C++ demangling function, as required by Itanium C++ ABI. This is weak,
  38 // because we do not require a C++ ABI library to be linked to a program
  39 // using sanitizers; if it's not present, we'll just use the mangled name.
  40 namespace __cxxabiv1 {
  41   extern "C" SANITIZER_WEAK_ATTRIBUTE
  42   char *__cxa_demangle(const char *mangled, char *buffer,
  43                                   size_t *length, int *status);
  44 }
  45
  46 namespace __sanitizer {
  47
  48 // Attempts to demangle the name via __cxa_demangle from __cxxabiv1.
  49 const char *DemangleCXXABI(const char *name) {
  50   // FIXME: __cxa_demangle aggressively insists on allocating memory.
  51   // There's not much we can do about that, short of providing our
  52   // own demangler (libc++abi's implementation could be adapted so that
  53   // it does not allocate). For now, we just call it anyway, and we leak
  54   // the returned value.
  55   if (&__cxxabiv1::__cxa_demangle)
  56     if (const char *demangled_name =
  57           __cxxabiv1::__cxa_demangle(name, 0, 0, 0))
  58       return demangled_name;
  59
  60   return nullptr;
  61 }
  62
  63 // As of now, there are no headers for the Swift runtime. Once they are
  64 // present, we will weakly link since we do not require Swift runtime to be
  65 // linked.
  66 typedef char *(*swift_demangle_ft)(const char *mangledName,
  67                                    size_t mangledNameLength, char *outputBuffer,
  68                                    size_t *outputBufferSize, uint32_t flags);
  69 static swift_demangle_ft swift_demangle_f;
  70
  71 // This must not happen lazily at symbolication time, because dlsym uses
  72 // malloc and thread-local storage, which is not a good thing to do during
  73 // symbolication.
  74 static void InitializeSwiftDemangler() {
  75   swift_demangle_f = (swift_demangle_ft)dlsym(RTLD_DEFAULT, "swift_demangle");
  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_APPLE
 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_APPLE
 169     UNIMPLEMENTED();
 170 #endif  // SANITIZER_APPLE
 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     if (common_flags()->demangle)
 217       argv[i++] = "-C";
 218     if (common_flags()->symbolize_inline_frames)
 219       argv[i++] = "-i";
 220     argv[i++] = "-fe";
 221     argv[i++] = module_name_;
 222     argv[i++] = nullptr;
 223     CHECK_LE(i, kArgVMax);
 224   }
 225
 226   bool ReachedEndOfOutput(const char *buffer, uptr length) const override;
 227
 228   bool ReadFromSymbolizer() override {
 229     if (!SymbolizerProcess::ReadFromSymbolizer())
 230       return false;
 231     auto &buff = GetBuff();
 232     // We should cut out output_terminator_ at the end of given buffer,
 233     // appended by addr2line to mark the end of its meaningful output.
 234     // We cannot scan buffer from it's beginning, because it is legal for it
 235     // to start with output_terminator_ in case given offset is invalid. So,
 236     // scanning from second character.
 237     char *garbage = internal_strstr(buff.data() + 1, output_terminator_);
 238     // This should never be NULL since buffer must end up with
 239     // output_terminator_.
 240     CHECK(garbage);
 241
 242     // Trim the buffer.
 243     uintptr_t new_size = garbage - buff.data();
 244     GetBuff().resize(new_size);
 245     GetBuff().push_back('\0');
 246     return true;
 247   }
 248
 249   const char *module_name_;  // Owned, leaked.
 250   static const char output_terminator_[];
 251 };
 252
 253 const char Addr2LineProcess::output_terminator_[] = "??\n??:0\n";
 254
 255 bool Addr2LineProcess::ReachedEndOfOutput(const char *buffer,
 256                                           uptr length) const {
 257   const size_t kTerminatorLen = sizeof(output_terminator_) - 1;
 258   // Skip, if we read just kTerminatorLen bytes, because Addr2Line output
 259   // should consist at least of two pairs of lines:
 260   // 1. First one, corresponding to given offset to be symbolized
 261   // (may be equal to output_terminator_, if offset is not valid).
 262   // 2. Second one for output_terminator_, itself to mark the end of output.
 263   if (length <= kTerminatorLen) return false;
 264   // Addr2Line output should end up with output_terminator_.
 265   return !internal_memcmp(buffer + length - kTerminatorLen,
 266                           output_terminator_, kTerminatorLen);
 267 }
 268
 269 class Addr2LinePool final : public SymbolizerTool {
 270  public:
 271   explicit Addr2LinePool(const char *addr2line_path,
 272                          LowLevelAllocator *allocator)
 273       : addr2line_path_(addr2line_path), allocator_(allocator) {
 274     addr2line_pool_.reserve(16);
 275   }
 276
 277   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 278     if (const char *buf =
 279             SendCommand(stack->info.module, stack->info.module_offset)) {
 280       ParseSymbolizePCOutput(buf, stack);
 281       return true;
 282     }
 283     return false;
 284   }
 285
 286   bool SymbolizeData(uptr addr, DataInfo *info) override {
 287     return false;
 288   }
 289
 290  private:
 291   const char *SendCommand(const char *module_name, uptr module_offset) {
 292     Addr2LineProcess *addr2line = 0;
 293     for (uptr i = 0; i < addr2line_pool_.size(); ++i) {
 294       if (0 ==
 295           internal_strcmp(module_name, addr2line_pool_[i]->module_name())) {
 296         addr2line = addr2line_pool_[i];
 297         break;
 298       }
 299     }
 300     if (!addr2line) {
 301       addr2line =
 302           new(*allocator_) Addr2LineProcess(addr2line_path_, module_name);
 303       addr2line_pool_.push_back(addr2line);
 304     }
 305     CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
 306     char buffer[kBufferSize];
 307     internal_snprintf(buffer, kBufferSize, "0x%zx\n0x%zx\n",
 308                       module_offset, dummy_address_);
 309     return addr2line->SendCommand(buffer);
 310   }
 311
 312   static const uptr kBufferSize = 64;
 313   const char *addr2line_path_;
 314   LowLevelAllocator *allocator_;
 315   InternalMmapVector<Addr2LineProcess*> addr2line_pool_;
 316   static const uptr dummy_address_ =
 317       FIRST_32_SECOND_64(UINT32_MAX, UINT64_MAX);
 318 };
 319
 320 #  if SANITIZER_SUPPORTS_WEAK_HOOKS
 321 extern "C" {
 322 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 323 __sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
 324                            char *Buffer, int MaxLength);
 325 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 326 __sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
 327                            char *Buffer, int MaxLength);
 328 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 329 __sanitizer_symbolize_frame(const char *ModuleName, u64 ModuleOffset,
 330                             char *Buffer, int MaxLength);
 331 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
 332 __sanitizer_symbolize_flush();
 333 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 334 __sanitizer_symbolize_demangle(const char *Name, char *Buffer, int MaxLength);
 335 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 336 __sanitizer_symbolize_set_demangle(bool Demangle);
 337 SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
 338 __sanitizer_symbolize_set_inline_frames(bool InlineFrames);
 339 }  // extern "C"
 340
 341 class InternalSymbolizer final : public SymbolizerTool {
 342  public:
 343   static InternalSymbolizer *get(LowLevelAllocator *alloc) {
 344     // These one is the most used one, so we will use it to detect a presence of
 345     // internal symbolizer.
 346     if (&__sanitizer_symbolize_code == nullptr)
 347       return nullptr;
 348     CHECK(__sanitizer_symbolize_set_demangle(common_flags()->demangle));
 349     CHECK(__sanitizer_symbolize_set_inline_frames(
 350         common_flags()->symbolize_inline_frames));
 351     return new (*alloc) InternalSymbolizer();
 352   }
 353
 354   bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
 355     bool result = __sanitizer_symbolize_code(
 356         stack->info.module, stack->info.module_offset, buffer_, sizeof(buffer_));
 357     if (result)
 358       ParseSymbolizePCOutput(buffer_, stack);
 359     return result;
 360   }
 361
 362   bool SymbolizeData(uptr addr, DataInfo *info) override {
 363     bool result = __sanitizer_symbolize_data(info->module, info->module_offset,
 364                                              buffer_, sizeof(buffer_));
 365     if (result) {
 366       ParseSymbolizeDataOutput(buffer_, info);
 367       info->start += (addr - info->module_offset);  // Add the base address.
 368     }
 369     return result;
 370   }
 371
 372   bool SymbolizeFrame(uptr addr, FrameInfo *info) override {
 373     bool result = __sanitizer_symbolize_frame(info->module, info->module_offset,
 374                                               buffer_, sizeof(buffer_));
 375     if (result)
 376       ParseSymbolizeFrameOutput(buffer_, &info->locals);
 377     return result;
 378   }
 379
 380   void Flush() override { __sanitizer_symbolize_flush(); }
 381
 382   const char *Demangle(const char *name) override {
 383     if (__sanitizer_symbolize_demangle(name, buffer_, sizeof(buffer_))) {
 384       char *res_buff = nullptr;
 385       ExtractToken(buffer_, "", &res_buff);
 386       return res_buff;
 387     }
 388     return nullptr;
 389   }
 390
 391  private:
 392   InternalSymbolizer() {}
 393
 394   char buffer_[16 * 1024];
 395 };
 396 #  else  // SANITIZER_SUPPORTS_WEAK_HOOKS
 397
 398 class InternalSymbolizer final : public SymbolizerTool {
 399  public:
 400   static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
 401 };
 402
 403 #  endif  // SANITIZER_SUPPORTS_WEAK_HOOKS
 404
 405 const char *Symbolizer::PlatformDemangle(const char *name) {
 406   return DemangleSwiftAndCXX(name);
 407 }
 408
 409 static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
 410   const char *path = common_flags()->external_symbolizer_path;
 411
 412   if (path && internal_strchr(path, '%')) {
 413     char *new_path = (char *)InternalAlloc(kMaxPathLength);
 414     SubstituteForFlagValue(path, new_path, kMaxPathLength);
 415     path = new_path;
 416   }
 417
 418   const char *binary_name = path ? StripModuleName(path) : "";
 419   static const char kLLVMSymbolizerPrefix[] = "llvm-symbolizer";
 420   if (path && path[0] == '\0') {
 421     VReport(2, "External symbolizer is explicitly disabled.\n");
 422     return nullptr;
 423   } else if (!internal_strncmp(binary_name, kLLVMSymbolizerPrefix,
 424                                internal_strlen(kLLVMSymbolizerPrefix))) {
 425     VReport(2, "Using llvm-symbolizer at user-specified path: %s\n", path);
 426     return new(*allocator) LLVMSymbolizer(path, allocator);
 427   } else if (!internal_strcmp(binary_name, "atos")) {
 428 #if SANITIZER_APPLE
 429     VReport(2, "Using atos at user-specified path: %s\n", path);
 430     return new(*allocator) AtosSymbolizer(path, allocator);
 431 #else  // SANITIZER_APPLE
 432     Report("ERROR: Using `atos` is only supported on Darwin.\n");
 433     Die();
 434 #endif  // SANITIZER_APPLE
 435   } else if (!internal_strcmp(binary_name, "addr2line")) {
 436     VReport(2, "Using addr2line at user-specified path: %s\n", path);
 437     return new(*allocator) Addr2LinePool(path, allocator);
 438   } else if (path) {
 439     Report("ERROR: External symbolizer path is set to '%s' which isn't "
 440            "a known symbolizer. Please set the path to the llvm-symbolizer "
 441            "binary or other known tool.\n", path);
 442     Die();
 443   }
 444
 445   // Otherwise symbolizer program is unknown, let's search $PATH
 446   CHECK(path == nullptr);
 447 #if SANITIZER_APPLE
 448   if (const char *found_path = FindPathToBinary("atos")) {
 449     VReport(2, "Using atos found at: %s\n", found_path);
 450     return new(*allocator) AtosSymbolizer(found_path, allocator);
 451   }
 452 #endif  // SANITIZER_APPLE
 453   if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
 454     VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
 455     return new(*allocator) LLVMSymbolizer(found_path, allocator);
 456   }
 457   if (common_flags()->allow_addr2line) {
 458     if (const char *found_path = FindPathToBinary("addr2line")) {
 459       VReport(2, "Using addr2line found at: %s\n", found_path);
 460       return new(*allocator) Addr2LinePool(found_path, allocator);
 461     }
 462   }
 463   return nullptr;
 464 }
 465
 466 static void ChooseSymbolizerTools(IntrusiveList<SymbolizerTool> *list,
 467                                   LowLevelAllocator *allocator) {
 468   if (!common_flags()->symbolize) {
 469     VReport(2, "Symbolizer is disabled.\n");
 470     return;
 471   }
 472   if (common_flags()->enable_symbolizer_markup) {
 473     VReport(2, "Using symbolizer markup");
 474     SymbolizerTool *tool = new (*allocator) MarkupSymbolizerTool();
 475     CHECK(tool);
 476     list->push_back(tool);
 477   }
 478   if (IsAllocatorOutOfMemory()) {
 479     VReport(2, "Cannot use internal symbolizer: out of memory\n");
 480   } else if (SymbolizerTool *tool = InternalSymbolizer::get(allocator)) {
 481     VReport(2, "Using internal symbolizer.\n");
 482     list->push_back(tool);
 483     return;
 484   }
 485   if (SymbolizerTool *tool = LibbacktraceSymbolizer::get(allocator)) {
 486     VReport(2, "Using libbacktrace symbolizer.\n");
 487     list->push_back(tool);
 488     return;
 489   }
 490
 491   if (SymbolizerTool *tool = ChooseExternalSymbolizer(allocator)) {
 492     list->push_back(tool);
 493   }
 494
 495 #if SANITIZER_APPLE
 496   VReport(2, "Using dladdr symbolizer.\n");
 497   list->push_back(new(*allocator) DlAddrSymbolizer());
 498 #endif  // SANITIZER_APPLE
 499 }
 500
 501 Symbolizer *Symbolizer::PlatformInit() {
 502   IntrusiveList<SymbolizerTool> list;
 503   list.clear();
 504   ChooseSymbolizerTools(&list, &symbolizer_allocator_);
 505   return new(symbolizer_allocator_) Symbolizer(list);
 506 }
 507
 508 void Symbolizer::LateInitialize() {
 509   Symbolizer::GetOrInit();
 510   InitializeSwiftDemangler();
 511 }
 512
 513 }  // namespace __sanitizer
 514
 515 #endif  // SANITIZER_POSIX