blob: eae9b570e3c1e3610c996ba7acd19b4ffb7ad2e2 [file] [log] [blame]
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +00001//===-- sanitizer_symbolizer_posix_libcdep.cc -----------------------------===//
2//
3// The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file is shared between AddressSanitizer and ThreadSanitizer
11// run-time libraries.
12// POSIX-specific implementation of symbolizer parts.
13//===----------------------------------------------------------------------===//
14
15#include "sanitizer_platform.h"
16#if SANITIZER_POSIX
Alexey Samsonov7847d772013-09-10 14:36:16 +000017#include "sanitizer_allocator_internal.h"
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +000018#include "sanitizer_common.h"
Stephen Hines2d1fdb22014-05-28 23:58:16 -070019#include "sanitizer_flags.h"
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +000020#include "sanitizer_internal_defs.h"
Alexey Samsonov7847d772013-09-10 14:36:16 +000021#include "sanitizer_linux.h"
22#include "sanitizer_placement_new.h"
23#include "sanitizer_procmaps.h"
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -070024#include "sanitizer_symbolizer_internal.h"
Stephen Hines2d1fdb22014-05-28 23:58:16 -070025#include "sanitizer_symbolizer_libbacktrace.h"
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -070026#include "sanitizer_symbolizer_mac.h"
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +000027
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +000028#include <unistd.h>
29
Alexey Samsonov7847d772013-09-10 14:36:16 +000030// C++ demangling function, as required by Itanium C++ ABI. This is weak,
31// because we do not require a C++ ABI library to be linked to a program
32// using sanitizers; if it's not present, we'll just use the mangled name.
33namespace __cxxabiv1 {
34 extern "C" SANITIZER_WEAK_ATTRIBUTE
35 char *__cxa_demangle(const char *mangled, char *buffer,
36 size_t *length, int *status);
37}
38
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +000039namespace __sanitizer {
40
Alexey Samsonov7847d772013-09-10 14:36:16 +000041// Attempts to demangle the name via __cxa_demangle from __cxxabiv1.
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -070042const char *DemangleCXXABI(const char *name) {
Alexey Samsonov7847d772013-09-10 14:36:16 +000043 // FIXME: __cxa_demangle aggressively insists on allocating memory.
44 // There's not much we can do about that, short of providing our
45 // own demangler (libc++abi's implementation could be adapted so that
46 // it does not allocate). For now, we just call it anyway, and we leak
47 // the returned value.
48 if (__cxxabiv1::__cxa_demangle)
49 if (const char *demangled_name =
50 __cxxabiv1::__cxa_demangle(name, 0, 0, 0))
51 return demangled_name;
52
53 return name;
54}
55
Stephen Hines86277eb2015-03-23 12:06:32 -070056// Parses one or more two-line strings in the following format:
57// <function_name>
58// <file_name>:<line_number>[:<column_number>]
59// Used by LLVMSymbolizer, Addr2LinePool and InternalSymbolizer, since all of
60// them use the same output format.
61static void ParseSymbolizePCOutput(const char *str, SymbolizedStack *res) {
62 bool top_frame = true;
63 SymbolizedStack *last = res;
64 while (true) {
65 char *function_name = 0;
66 str = ExtractToken(str, "\n", &function_name);
67 CHECK(function_name);
68 if (function_name[0] == '\0') {
69 // There are no more frames.
Pirama Arumuga Nainarcdce50b2015-07-01 12:26:56 -070070 InternalFree(function_name);
Stephen Hines86277eb2015-03-23 12:06:32 -070071 break;
72 }
73 SymbolizedStack *cur;
74 if (top_frame) {
75 cur = res;
76 top_frame = false;
77 } else {
78 cur = SymbolizedStack::New(res->info.address);
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -070079 cur->info.FillModuleInfo(res->info.module, res->info.module_offset);
Stephen Hines86277eb2015-03-23 12:06:32 -070080 last->next = cur;
81 last = cur;
Stephen Hines2d1fdb22014-05-28 23:58:16 -070082 }
83
Stephen Hines86277eb2015-03-23 12:06:32 -070084 AddressInfo *info = &cur->info;
85 info->function = function_name;
86 // Parse <file>:<line>:<column> buffer.
87 char *file_line_info = 0;
88 str = ExtractToken(str, "\n", &file_line_info);
89 CHECK(file_line_info);
90 const char *line_info = ExtractToken(file_line_info, ":", &info->file);
91 line_info = ExtractInt(line_info, ":", &info->line);
92 line_info = ExtractInt(line_info, "", &info->column);
93 InternalFree(file_line_info);
94
95 // Functions and filenames can be "??", in which case we write 0
96 // to address info to mark that names are unknown.
97 if (0 == internal_strcmp(info->function, "??")) {
98 InternalFree(info->function);
99 info->function = 0;
100 }
101 if (0 == internal_strcmp(info->file, "??")) {
102 InternalFree(info->file);
103 info->file = 0;
104 }
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700105 }
Stephen Hines86277eb2015-03-23 12:06:32 -0700106}
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700107
Stephen Hines86277eb2015-03-23 12:06:32 -0700108// Parses a two-line string in the following format:
109// <symbol_name>
110// <start_address> <size>
111// Used by LLVMSymbolizer and InternalSymbolizer.
112static void ParseSymbolizeDataOutput(const char *str, DataInfo *info) {
113 str = ExtractToken(str, "\n", &info->name);
114 str = ExtractUptr(str, " ", &info->start);
115 str = ExtractUptr(str, "\n", &info->size);
116}
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700117
118// For now we assume the following protocol:
119// For each request of the form
120// <module_name> <module_offset>
121// passed to STDIN, external symbolizer prints to STDOUT response:
122// <function_name>
123// <file_name>:<line_number>:<column_number>
124// <function_name>
125// <file_name>:<line_number>:<column_number>
126// ...
127// <empty line>
128class LLVMSymbolizerProcess : public SymbolizerProcess {
129 public:
130 explicit LLVMSymbolizerProcess(const char *path) : SymbolizerProcess(path) {}
131
132 private:
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700133 bool ReachedEndOfOutput(const char *buffer, uptr length) const override {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700134 // Empty line marks the end of llvm-symbolizer output.
135 return length >= 2 && buffer[length - 1] == '\n' &&
136 buffer[length - 2] == '\n';
137 }
138
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700139 void ExecuteWithDefaultArgs(const char *path_to_binary) const override {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700140#if defined(__x86_64__)
141 const char* const kSymbolizerArch = "--default-arch=x86_64";
142#elif defined(__i386__)
143 const char* const kSymbolizerArch = "--default-arch=i386";
Stephen Hines6d186232014-11-26 17:56:19 -0800144#elif defined(__powerpc64__) && defined(__BIG_ENDIAN__)
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700145 const char* const kSymbolizerArch = "--default-arch=powerpc64";
Stephen Hines6d186232014-11-26 17:56:19 -0800146#elif defined(__powerpc64__) && defined(__LITTLE_ENDIAN__)
147 const char* const kSymbolizerArch = "--default-arch=powerpc64le";
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700148#else
149 const char* const kSymbolizerArch = "--default-arch=unknown";
150#endif
Stephen Hines6d186232014-11-26 17:56:19 -0800151
152 const char *const inline_flag = common_flags()->symbolize_inline_frames
153 ? "--inlining=true"
154 : "--inlining=false";
155 execl(path_to_binary, path_to_binary, inline_flag, kSymbolizerArch,
156 (char *)0);
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700157 }
158};
159
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700160class LLVMSymbolizer : public SymbolizerTool {
161 public:
162 explicit LLVMSymbolizer(const char *path, LowLevelAllocator *allocator)
163 : symbolizer_process_(new(*allocator) LLVMSymbolizerProcess(path)) {}
164
165 bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
166 if (const char *buf = SendCommand(/*is_data*/ false, stack->info.module,
167 stack->info.module_offset)) {
168 ParseSymbolizePCOutput(buf, stack);
169 return true;
170 }
171 return false;
172 }
173
174 bool SymbolizeData(uptr addr, DataInfo *info) override {
175 if (const char *buf =
176 SendCommand(/*is_data*/ true, info->module, info->module_offset)) {
177 ParseSymbolizeDataOutput(buf, info);
178 info->start += (addr - info->module_offset); // Add the base address.
179 return true;
180 }
181 return false;
182 }
183
184 private:
185 const char *SendCommand(bool is_data, const char *module_name,
186 uptr module_offset) {
187 CHECK(module_name);
188 internal_snprintf(buffer_, kBufferSize, "%s\"%s\" 0x%zx\n",
189 is_data ? "DATA " : "", module_name, module_offset);
190 return symbolizer_process_->SendCommand(buffer_);
191 }
192
193 LLVMSymbolizerProcess *symbolizer_process_;
194 static const uptr kBufferSize = 16 * 1024;
195 char buffer_[kBufferSize];
196};
197
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700198class Addr2LineProcess : public SymbolizerProcess {
199 public:
200 Addr2LineProcess(const char *path, const char *module_name)
201 : SymbolizerProcess(path), module_name_(internal_strdup(module_name)) {}
202
203 const char *module_name() const { return module_name_; }
204
205 private:
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700206 bool ReachedEndOfOutput(const char *buffer, uptr length) const override {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700207 // Output should consist of two lines.
208 int num_lines = 0;
209 for (uptr i = 0; i < length; ++i) {
210 if (buffer[i] == '\n')
211 num_lines++;
212 if (num_lines >= 2)
213 return true;
214 }
215 return false;
216 }
217
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700218 void ExecuteWithDefaultArgs(const char *path_to_binary) const override {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700219 execl(path_to_binary, path_to_binary, "-Cfe", module_name_, (char *)0);
220 }
221
222 const char *module_name_; // Owned, leaked.
223};
224
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700225class Addr2LinePool : public SymbolizerTool {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700226 public:
227 explicit Addr2LinePool(const char *addr2line_path,
228 LowLevelAllocator *allocator)
229 : addr2line_path_(addr2line_path), allocator_(allocator),
230 addr2line_pool_(16) {}
231
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700232 bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
233 if (const char *buf =
234 SendCommand(stack->info.module, stack->info.module_offset)) {
235 ParseSymbolizePCOutput(buf, stack);
236 return true;
237 }
238 return false;
239 }
240
241 bool SymbolizeData(uptr addr, DataInfo *info) override {
242 return false;
243 }
244
245 private:
246 const char *SendCommand(const char *module_name, uptr module_offset) {
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700247 Addr2LineProcess *addr2line = 0;
248 for (uptr i = 0; i < addr2line_pool_.size(); ++i) {
249 if (0 ==
250 internal_strcmp(module_name, addr2line_pool_[i]->module_name())) {
251 addr2line = addr2line_pool_[i];
252 break;
253 }
254 }
255 if (!addr2line) {
256 addr2line =
257 new(*allocator_) Addr2LineProcess(addr2line_path_, module_name);
258 addr2line_pool_.push_back(addr2line);
259 }
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700260 CHECK_EQ(0, internal_strcmp(module_name, addr2line->module_name()));
261 char buffer_[kBufferSize];
262 internal_snprintf(buffer_, kBufferSize, "0x%zx\n", module_offset);
263 return addr2line->SendCommand(buffer_);
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700264 }
265
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700266 static const uptr kBufferSize = 32;
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700267 const char *addr2line_path_;
268 LowLevelAllocator *allocator_;
269 InternalMmapVector<Addr2LineProcess*> addr2line_pool_;
Alexey Samsonov7847d772013-09-10 14:36:16 +0000270};
271
Alexey Samsonov7847d772013-09-10 14:36:16 +0000272#if SANITIZER_SUPPORTS_WEAK_HOOKS
273extern "C" {
274SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
275bool __sanitizer_symbolize_code(const char *ModuleName, u64 ModuleOffset,
276 char *Buffer, int MaxLength);
277SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
278bool __sanitizer_symbolize_data(const char *ModuleName, u64 ModuleOffset,
279 char *Buffer, int MaxLength);
280SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
281void __sanitizer_symbolize_flush();
282SANITIZER_INTERFACE_ATTRIBUTE SANITIZER_WEAK_ATTRIBUTE
283int __sanitizer_symbolize_demangle(const char *Name, char *Buffer,
284 int MaxLength);
285} // extern "C"
286
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700287class InternalSymbolizer : public SymbolizerTool {
Alexey Samsonov7847d772013-09-10 14:36:16 +0000288 public:
Peter Collingbournec1a1ed62013-10-25 23:03:29 +0000289 static InternalSymbolizer *get(LowLevelAllocator *alloc) {
Alexey Samsonov7847d772013-09-10 14:36:16 +0000290 if (__sanitizer_symbolize_code != 0 &&
291 __sanitizer_symbolize_data != 0) {
Peter Collingbournec1a1ed62013-10-25 23:03:29 +0000292 return new(*alloc) InternalSymbolizer();
Alexey Samsonov7847d772013-09-10 14:36:16 +0000293 }
294 return 0;
295 }
296
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700297 bool SymbolizePC(uptr addr, SymbolizedStack *stack) override {
298 bool result = __sanitizer_symbolize_code(
299 stack->info.module, stack->info.module_offset, buffer_, kBufferSize);
300 if (result) ParseSymbolizePCOutput(buffer_, stack);
301 return result;
Alexey Samsonov7847d772013-09-10 14:36:16 +0000302 }
303
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700304 bool SymbolizeData(uptr addr, DataInfo *info) override {
305 bool result = __sanitizer_symbolize_data(info->module, info->module_offset,
306 buffer_, kBufferSize);
307 if (result) {
308 ParseSymbolizeDataOutput(buffer_, info);
309 info->start += (addr - info->module_offset); // Add the base address.
310 }
311 return result;
312 }
313
314 void Flush() override {
Alexey Samsonov7847d772013-09-10 14:36:16 +0000315 if (__sanitizer_symbolize_flush)
316 __sanitizer_symbolize_flush();
317 }
318
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700319 const char *Demangle(const char *name) override {
Alexey Samsonov7847d772013-09-10 14:36:16 +0000320 if (__sanitizer_symbolize_demangle) {
321 for (uptr res_length = 1024;
322 res_length <= InternalSizeClassMap::kMaxSize;) {
323 char *res_buff = static_cast<char*>(InternalAlloc(res_length));
324 uptr req_length =
325 __sanitizer_symbolize_demangle(name, res_buff, res_length);
326 if (req_length > res_length) {
327 res_length = req_length + 1;
328 InternalFree(res_buff);
329 continue;
330 }
331 return res_buff;
332 }
333 }
334 return name;
335 }
336
337 private:
338 InternalSymbolizer() { }
339
340 static const int kBufferSize = 16 * 1024;
341 static const int kMaxDemangledNameSize = 1024;
342 char buffer_[kBufferSize];
343};
344#else // SANITIZER_SUPPORTS_WEAK_HOOKS
345
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700346class InternalSymbolizer : public SymbolizerTool {
Alexey Samsonov7847d772013-09-10 14:36:16 +0000347 public:
Peter Collingbournec1a1ed62013-10-25 23:03:29 +0000348 static InternalSymbolizer *get(LowLevelAllocator *alloc) { return 0; }
Alexey Samsonov7847d772013-09-10 14:36:16 +0000349};
350
351#endif // SANITIZER_SUPPORTS_WEAK_HOOKS
352
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700353const char *Symbolizer::PlatformDemangle(const char *name) {
354 return DemangleCXXABI(name);
355}
Peter Collingbournec1a1ed62013-10-25 23:03:29 +0000356
Pirama Arumuga Nainarcdce50b2015-07-01 12:26:56 -0700357void Symbolizer::PlatformPrepareForSandboxing() {}
Alexey Samsonov7847d772013-09-10 14:36:16 +0000358
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700359static SymbolizerTool *ChooseExternalSymbolizer(LowLevelAllocator *allocator) {
360 const char *path = common_flags()->external_symbolizer_path;
361 const char *binary_name = path ? StripModuleName(path) : "";
362 if (path && path[0] == '\0') {
363 VReport(2, "External symbolizer is explicitly disabled.\n");
364 return nullptr;
365 } else if (!internal_strcmp(binary_name, "llvm-symbolizer")) {
366 VReport(2, "Using llvm-symbolizer at user-specified path: %s\n", path);
367 return new(*allocator) LLVMSymbolizer(path, allocator);
368 } else if (!internal_strcmp(binary_name, "atos")) {
369#if SANITIZER_MAC
370 VReport(2, "Using atos at user-specified path: %s\n", path);
371 return new(*allocator) AtosSymbolizer(path, allocator);
372#else // SANITIZER_MAC
373 Report("ERROR: Using `atos` is only supported on Darwin.\n");
374 Die();
375#endif // SANITIZER_MAC
376 } else if (!internal_strcmp(binary_name, "addr2line")) {
377 VReport(2, "Using addr2line at user-specified path: %s\n", path);
378 return new(*allocator) Addr2LinePool(path, allocator);
379 } else if (path) {
380 Report("ERROR: External symbolizer path is set to '%s' which isn't "
381 "a known symbolizer. Please set the path to the llvm-symbolizer "
382 "binary or other known tool.\n", path);
383 Die();
Stephen Hines2d1fdb22014-05-28 23:58:16 -0700384 }
Alexey Samsonov7847d772013-09-10 14:36:16 +0000385
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700386 // Otherwise symbolizer program is unknown, let's search $PATH
387 CHECK(path == nullptr);
388 if (const char *found_path = FindPathToBinary("llvm-symbolizer")) {
389 VReport(2, "Using llvm-symbolizer found at: %s\n", found_path);
390 return new(*allocator) LLVMSymbolizer(found_path, allocator);
391 }
392#if SANITIZER_MAC
393 if (const char *found_path = FindPathToBinary("atos")) {
394 VReport(2, "Using atos found at: %s\n", found_path);
395 return new(*allocator) AtosSymbolizer(found_path, allocator);
396 }
397#endif // SANITIZER_MAC
398 if (common_flags()->allow_addr2line) {
399 if (const char *found_path = FindPathToBinary("addr2line")) {
400 VReport(2, "Using addr2line found at: %s\n", found_path);
401 return new(*allocator) Addr2LinePool(found_path, allocator);
Alexey Samsonove059bd32013-09-10 16:16:27 +0000402 }
403 }
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700404 return nullptr;
405}
Peter Collingbournec1a1ed62013-10-25 23:03:29 +0000406
Pirama Arumuga Nainar7c915052015-04-08 08:58:29 -0700407static void ChooseSymbolizerTools(IntrusiveList<SymbolizerTool> *list,
408 LowLevelAllocator *allocator) {
409 if (!common_flags()->symbolize) {
410 VReport(2, "Symbolizer is disabled.\n");
411 return;
412 }
413 if (SymbolizerTool *tool = InternalSymbolizer::get(allocator)) {
414 VReport(2, "Using internal symbolizer.\n");
415 list->push_back(tool);
416 return;
417 }
418 if (SymbolizerTool *tool = LibbacktraceSymbolizer::get(allocator)) {
419 VReport(2, "Using libbacktrace symbolizer.\n");
420 list->push_back(tool);
421 return;
422 }
423
424 if (SymbolizerTool *tool = ChooseExternalSymbolizer(allocator)) {
425 list->push_back(tool);
426 } else {
427 VReport(2, "No internal or external symbolizer found.\n");
428 }
429
430#if SANITIZER_MAC
431 VReport(2, "Using dladdr symbolizer.\n");
432 list->push_back(new(*allocator) DlAddrSymbolizer());
433#endif // SANITIZER_MAC
434}
435
436Symbolizer *Symbolizer::PlatformInit() {
437 IntrusiveList<SymbolizerTool> list;
438 list.clear();
439 ChooseSymbolizerTools(&list, &symbolizer_allocator_);
Pirama Arumuga Nainar259f7062015-05-06 11:49:53 -0700440 return new(symbolizer_allocator_) Symbolizer(list);
Alexey Samsonov7847d772013-09-10 14:36:16 +0000441}
Alexey Samsonovff0ce7f2013-06-11 08:14:24 +0000442
443} // namespace __sanitizer
444
445#endif // SANITIZER_POSIX