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.
40namespace __cxxabiv1 {
41extern "C" SANITIZER_WEAK_ATTRIBUTE char *__cxa_demangle(const char *mangled,
42 char *buffer,
43 size_t *length,
44 int *status);
45}
46
47namespace __sanitizer {
48
49// Attempts to demangle the name via __cxa_demangle from __cxxabiv1.
50const 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 = __cxxabiv1::__cxa_demangle(mangled: name, buffer: 0, length: 0, status: 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.
66typedef char *(*swift_demangle_ft)(const char *mangledName,
67 size_t mangledNameLength, char *outputBuffer,
68 size_t *outputBufferSize, uint32_t flags);
69static 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.
74static void InitializeSwiftDemangler() {
75 swift_demangle_f = (swift_demangle_ft)dlsym(RTLD_DEFAULT, name: "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.
80const char *DemangleSwift(const char *name) {
81 if (swift_demangle_f)
82 return swift_demangle_f(name, internal_strlen(s: name), 0, 0, 0);
83
84 return nullptr;
85}
86
87const char *DemangleSwiftAndCXX(const char *name) {
88 if (!name)
89 return nullptr;
90 if (const char *swift_demangled_name = DemangleSwift(name))
91 return swift_demangled_name;
92 return DemangleCXXABI(name);
93}
94
95static bool CreateTwoHighNumberedPipes(int *infd_, int *outfd_) {
96 int *infd = NULL;
97 int *outfd = NULL;
98 // The client program may close its stdin and/or stdout and/or stderr
99 // thus allowing socketpair to reuse file descriptors 0, 1 or 2.
100 // In this case the communication between the forked processes may be
101 // broken if either the parent or the child tries to close or duplicate
102 // these descriptors. The loop below produces two pairs of file
103 // descriptors, each greater than 2 (stderr).
104 int sock_pair[5][2];
105 for (int i = 0; i < 5; i++) {
106 if (pipe(pipedes: sock_pair[i]) == -1) {
107 for (int j = 0; j < i; j++) {
108 internal_close(fd: sock_pair[j][0]);
109 internal_close(fd: sock_pair[j][1]);
110 }
111 return false;
112 } else if (sock_pair[i][0] > 2 && sock_pair[i][1] > 2) {
113 if (infd == NULL) {
114 infd = sock_pair[i];
115 } else {
116 outfd = sock_pair[i];
117 for (int j = 0; j < i; j++) {
118 if (sock_pair[j] == infd)
119 continue;
120 internal_close(fd: sock_pair[j][0]);
121 internal_close(fd: sock_pair[j][1]);
122 }
123 break;
124 }
125 }
126 }
127 CHECK(infd);
128 CHECK(outfd);
129 infd_[0] = infd[0];
130 infd_[1] = infd[1];
131 outfd_[0] = outfd[0];
132 outfd_[1] = outfd[1];
133 return true;
134}
135
136bool SymbolizerProcess::StartSymbolizerSubprocess() {
137 if (!FileExists(filename: path_)) {
138 if (!reported_invalid_path_) {
139 Report(format: "WARNING: invalid path to external symbolizer!\n");
140 reported_invalid_path_ = true;
141 }
142 return false;
143 }
144
145 const char *argv[kArgVMax];
146 GetArgV(path_to_binary: path_, argv);
147 pid_t pid;
148
149 // Report how symbolizer is being launched for debugging purposes.
150 if (Verbosity() >= 3) {
151 // Only use `Report` for first line so subsequent prints don't get prefixed
152 // with current PID.
153 Report(format: "Launching Symbolizer process: ");
154 for (unsigned index = 0; index < kArgVMax && argv[index]; ++index)
155 Printf(format: "%s ", argv[index]);
156 Printf(format: "\n");
157 }
158
159 if (use_posix_spawn_) {
160# if SANITIZER_APPLE
161 fd_t fd = internal_spawn(argv, const_cast<const char **>(GetEnvP()), &pid);
162 if (fd == kInvalidFd) {
163 Report("WARNING: failed to spawn external symbolizer (errno: %d)\n",
164 errno);
165 return false;
166 }
167
168 input_fd_ = fd;
169 output_fd_ = fd;
170# else // SANITIZER_APPLE
171 UNIMPLEMENTED();
172# endif // SANITIZER_APPLE
173 } else {
174 fd_t infd[2] = {}, outfd[2] = {};
175 if (!CreateTwoHighNumberedPipes(infd_: infd, outfd_: outfd)) {
176 Report(
177 format: "WARNING: Can't create a socket pair to start "
178 "external symbolizer (errno: %d)\n",
179 errno);
180 return false;
181 }
182
183 pid = StartSubprocess(filename: path_, argv, envp: GetEnvP(), /* stdin */ stdin_fd: outfd[0],
184 /* stdout */ stdout_fd: infd[1]);
185 if (pid < 0) {
186 internal_close(fd: infd[0]);
187 internal_close(fd: outfd[1]);
188 return false;
189 }
190
191 input_fd_ = infd[0];
192 output_fd_ = outfd[1];
193 }
194
195 CHECK_GT(pid, 0);
196
197 // Check that symbolizer subprocess started successfully.
198 SleepForMillis(millis: kSymbolizerStartupTimeMillis);
199 if (!IsProcessRunning(pid)) {
200 // Either waitpid failed, or child has already exited.
201 Report(format: "WARNING: external symbolizer didn't start up correctly!\n");
202 return false;
203 }
204
205 return true;
206}
207
208class Addr2LineProcess final : public SymbolizerProcess {
209 public:
210 Addr2LineProcess(const char *path, const char *module_name)
211 : SymbolizerProcess(path), module_name_(internal_strdup(s: module_name)) {}
212
213 const char *module_name() const { return module_name_; }
214
215 private:
216 void GetArgV(const char *path_to_binary,
217 const char *(&argv)[kArgVMax]) const override {
218 int i = 0;
219 argv[i++] = path_to_binary;
220 if (common_flags()->demangle)
221 argv[i++] = "-C";
222 if (common_flags()->symbolize_inline_frames)
223 argv[i++] = "-i";
224 argv[i++] = "-fe";
225 argv[i++] = module_name_;
226 argv[i++] = nullptr;
227 CHECK_LE(i, kArgVMax);
228 }
229
230 bool ReachedEndOfOutput(const char *buffer, uptr length) const override;
231
232 bool ReadFromSymbolizer() override {
233 if (!SymbolizerProcess::ReadFromSymbolizer())
234 return false;
235 auto &buff = GetBuff();
236 // We should cut out output_terminator_ at the end of given buffer,
237 // appended by addr2line to mark the end of its meaningful output.
238 // We cannot scan buffer from it's beginning, because it is legal for it
239 // to start with output_terminator_ in case given offset is invalid. So,
240 // scanning from second character.
241 char *garbage = internal_strstr(haystack: buff.data() + 1, needle: output_terminator_);
242 // This should never be NULL since buffer must end up with
243 // output_terminator_.
244 CHECK(garbage);
245
246 // Trim the buffer.
247 uintptr_t new_size = garbage - buff.data();
248 GetBuff().resize(new_size);
249 GetBuff().push_back(element: '\0');
250 return true;
251 }
252
253 const char *module_name_; // Owned, leaked.
254 static const char output_terminator_[];
255};
256
257const char Addr2LineProcess::output_terminator_[] = "??\n??:0\n";
258
259bool 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)
268 return false;
269 // Addr2Line output should end up with output_terminator_.
270 return !internal_memcmp(s1: buffer + length - kTerminatorLen, s2: output_terminator_,
271 n: kTerminatorLen);
272}
273
274class Addr2LinePool final : public SymbolizerTool {
275 public:
276 explicit Addr2LinePool(const char *addr2line_path,
277 LowLevelAllocator *allocator)
278 : addr2line_path_(addr2line_path), allocator_(allocator) {
279 addr2line_pool_.reserve(new_size: 16);
280 }
281
282 bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
283 if (const char *buf =
284 SendCommand(module_name: stack->info.module, module_offset: stack->info.module_offset)) {
285 ParseSymbolizePCOutput(str: buf, res: stack);
286 return true;
287 }
288 return false;
289 }
290
291 bool SymbolizeData(uptr addr, DataInfo *info) override { return false; }
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(s1: module_name, s2: 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(element: addr2line);
307 }
308 CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
309 char buffer[kBufferSize];
310 internal_snprintf(buffer, length: kBufferSize, format: "0x%zx\n0x%zx\n", module_offset,
311 dummy_address_);
312 return addr2line->SendCommand(command: 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_ = FIRST_32_SECOND_64(UINT32_MAX, UINT64_MAX);
320};
321
322# if SANITIZER_SUPPORTS_WEAK_HOOKS
323extern "C" {
324SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
325__sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
326 char *Buffer, int MaxLength);
327SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
328__sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
329 char *Buffer, int MaxLength);
330SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
331__sanitizer_symbolize_frame(const char *ModuleName, u64 ModuleOffset,
332 char *Buffer, int MaxLength);
333SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE void
334__sanitizer_symbolize_flush();
335SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
336__sanitizer_symbolize_demangle(const char *Name, char *Buffer, int MaxLength);
337SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
338__sanitizer_symbolize_set_demangle(bool Demangle);
339SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE bool
340__sanitizer_symbolize_set_inline_frames(bool InlineFrames);
341} // extern "C"
342
343class InternalSymbolizer final : public SymbolizerTool {
344 public:
345 static InternalSymbolizer *get(LowLevelAllocator *alloc) {
346 // These one is the most used one, so we will use it to detect a presence of
347 // internal symbolizer.
348 if (&__sanitizer_symbolize_code == nullptr)
349 return nullptr;
350 CHECK(__sanitizer_symbolize_set_demangle(common_flags()->demangle));
351 CHECK(__sanitizer_symbolize_set_inline_frames(
352 common_flags()->symbolize_inline_frames));
353 return new (*alloc) InternalSymbolizer();
354 }
355
356 bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
357 bool result = __sanitizer_symbolize_code(ModuleName: stack->info.module,
358 ModuleOffset: stack->info.module_offset, Buffer: buffer_,
359 MaxLength: sizeof(buffer_));
360 if (result)
361 ParseSymbolizePCOutput(str: buffer_, res: stack);
362 return result;
363 }
364
365 bool SymbolizeData(uptr addr, DataInfo *info) override {
366 bool result = __sanitizer_symbolize_data(ModuleName: info->module, ModuleOffset: info->module_offset,
367 Buffer: buffer_, MaxLength: sizeof(buffer_));
368 if (result) {
369 ParseSymbolizeDataOutput(str: buffer_, info);
370 info->start += (addr - info->module_offset); // Add the base address.
371 }
372 return result;
373 }
374
375 bool SymbolizeFrame(uptr addr, FrameInfo *info) override {
376 bool result = __sanitizer_symbolize_frame(ModuleName: info->module, ModuleOffset: info->module_offset,
377 Buffer: buffer_, MaxLength: sizeof(buffer_));
378 if (result)
379 ParseSymbolizeFrameOutput(str: buffer_, locals: &info->locals);
380 return result;
381 }
382
383 void Flush() override { __sanitizer_symbolize_flush(); }
384
385 const char *Demangle(const char *name) override {
386 if (__sanitizer_symbolize_demangle(Name: name, Buffer: buffer_, MaxLength: sizeof(buffer_))) {
387 char *res_buff = nullptr;
388 ExtractToken(str: buffer_, delims: "", result: &res_buff);
389 return res_buff;
390 }
391 return nullptr;
392 }
393
394 private:
395 InternalSymbolizer() {}
396
397 char buffer_[16 * 1024];
398};
399# else // SANITIZER_SUPPORTS_WEAK_HOOKS
400
401class InternalSymbolizer final : public SymbolizerTool {
402 public:
403 static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
404};
405
406# endif // SANITIZER_SUPPORTS_WEAK_HOOKS
407
408const char *Symbolizer::PlatformDemangle(const char *name) {
409 return DemangleSwiftAndCXX(name);
410}
411
412static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
413 const char *path = common_flags()->external_symbolizer_path;
414
415 if (path && internal_strchr(s: path, c: '%')) {
416 char *new_path = (char *)InternalAlloc(size: kMaxPathLength);
417 SubstituteForFlagValue(s: path, out: new_path, out_size: kMaxPathLength);
418 path = new_path;
419 }
420
421 const char *binary_name = path ? StripModuleName(module: path) : "";
422 static const char kLLVMSymbolizerPrefix[] = "llvm-symbolizer";
423 if (path && path[0] == '\0') {
424 VReport(2, "External symbolizer is explicitly disabled.\n");
425 return nullptr;
426 } else if (!internal_strncmp(s1: binary_name, s2: kLLVMSymbolizerPrefix,
427 n: internal_strlen(s: kLLVMSymbolizerPrefix))) {
428 VReport(2, "Using llvm-symbolizer at user-specified path: %s\n", path);
429 return new (*allocator) LLVMSymbolizer(path, allocator);
430 } else if (!internal_strcmp(s1: binary_name, s2: "atos")) {
431# if SANITIZER_APPLE
432 VReport(2, "Using atos at user-specified path: %s\n", path);
433 return new (*allocator) AtosSymbolizer(path, allocator);
434# else // SANITIZER_APPLE
435 Report(format: "ERROR: Using `atos` is only supported on Darwin.\n");
436 Die();
437# endif // SANITIZER_APPLE
438 } else if (!internal_strcmp(s1: binary_name, s2: "addr2line")) {
439 VReport(2, "Using addr2line at user-specified path: %s\n", path);
440 return new (*allocator) Addr2LinePool(path, allocator);
441 } else if (path) {
442 Report(
443 format: "ERROR: External symbolizer path is set to '%s' which isn't "
444 "a known symbolizer. Please set the path to the llvm-symbolizer "
445 "binary or other known tool.\n",
446 path);
447 Die();
448 }
449
450 // Otherwise symbolizer program is unknown, let's search $PATH
451# ifdef SANITIZER_DISABLE_SYMBOLIZER_PATH_SEARCH
452 VReport(2,
453 "Symbolizer path search is disabled in the runtime "
454 "build configuration.\n");
455 return nullptr;
456# else
457 CHECK(path == nullptr);
458# if SANITIZER_APPLE
459 if (const char *found_path = FindPathToBinary("atos")) {
460 VReport(2, "Using atos found at: %s\n", found_path);
461 return new (*allocator) AtosSymbolizer(found_path, allocator);
462 }
463# endif // SANITIZER_APPLE
464 if (const char *found_path = FindPathToBinary(name: "llvm-symbolizer")) {
465 VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
466 return new (*allocator) LLVMSymbolizer(found_path, allocator);
467 }
468 if (common_flags()->allow_addr2line) {
469 if (const char *found_path = FindPathToBinary(name: "addr2line")) {
470 VReport(2, "Using addr2line found at: %s\n", found_path);
471 return new (*allocator) Addr2LinePool(found_path, allocator);
472 }
473 }
474 return nullptr;
475# endif // SANITIZER_DISABLE_SYMBOLIZER_PATH_SEARCH
476}
477
478static void ChooseSymbolizerTools(IntrusiveList<SymbolizerTool> *list,
479 LowLevelAllocator *allocator) {
480 if (!common_flags()->symbolize) {
481 VReport(2, "Symbolizer is disabled.\n");
482 return;
483 }
484 if (common_flags()->enable_symbolizer_markup) {
485 VReport(2, "Using symbolizer markup");
486 SymbolizerTool *tool = new (*allocator) MarkupSymbolizerTool();
487 CHECK(tool);
488 list->push_back(x: tool);
489 }
490 if (IsAllocatorOutOfMemory()) {
491 VReport(2, "Cannot use internal symbolizer: out of memory\n");
492 } else if (SymbolizerTool *tool = InternalSymbolizer::get(alloc: allocator)) {
493 VReport(2, "Using internal symbolizer.\n");
494 list->push_back(x: tool);
495 return;
496 }
497 if (SymbolizerTool *tool = LibbacktraceSymbolizer::get(alloc: allocator)) {
498 VReport(2, "Using libbacktrace symbolizer.\n");
499 list->push_back(x: tool);
500 return;
501 }
502
503 if (SymbolizerTool *tool = ChooseExternalSymbolizer(allocator)) {
504 list->push_back(x: tool);
505 }
506
507# if SANITIZER_APPLE
508 VReport(2, "Using dladdr symbolizer.\n");
509 list->push_back(new (*allocator) DlAddrSymbolizer());
510# endif // SANITIZER_APPLE
511}
512
513Symbolizer *Symbolizer::PlatformInit() {
514 IntrusiveList<SymbolizerTool> list;
515 list.clear();
516 ChooseSymbolizerTools(list: &list, allocator: &symbolizer_allocator_);
517 return new (symbolizer_allocator_) Symbolizer(list);
518}
519
520void Symbolizer::LateInitialize() {
521 Symbolizer::GetOrInit();
522 InitializeSwiftDemangler();
523}
524
525} // namespace __sanitizer
526
527#endif // SANITIZER_POSIX
528