1 //===-- LLVMSymbolize.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 // Implementation for LLVM symbolization library.
10 //
11 //===----------------------------------------------------------------------===//
12
13 #include "llvm/DebugInfo/Symbolize/Symbolize.h"
14
15 #include "SymbolizableObjectFile.h"
16
17 #include "llvm/ADT/STLExtras.h"
18 #include "llvm/BinaryFormat/COFF.h"
19 #include "llvm/Config/config.h"
20 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
21 #include "llvm/DebugInfo/PDB/PDB.h"
22 #include "llvm/DebugInfo/PDB/PDBContext.h"
23 #include "llvm/Demangle/Demangle.h"
24 #include "llvm/Object/COFF.h"
25 #include "llvm/Object/MachO.h"
26 #include "llvm/Object/MachOUniversal.h"
27 #include "llvm/Support/CRC.h"
28 #include "llvm/Support/Casting.h"
29 #include "llvm/Support/Compression.h"
30 #include "llvm/Support/DataExtractor.h"
31 #include "llvm/Support/Errc.h"
32 #include "llvm/Support/FileSystem.h"
33 #include "llvm/Support/MemoryBuffer.h"
34 #include "llvm/Support/Path.h"
35 #include <algorithm>
36 #include <cassert>
37 #include <cstring>
38
39 namespace llvm {
40 namespace symbolize {
41
42 template <typename T>
43 Expected<DILineInfo>
symbolizeCodeCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)44 LLVMSymbolizer::symbolizeCodeCommon(const T &ModuleSpecifier,
45 object::SectionedAddress ModuleOffset) {
46
47 auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
48 if (!InfoOrErr)
49 return InfoOrErr.takeError();
50
51 SymbolizableModule *Info = *InfoOrErr;
52
53 // A null module means an error has already been reported. Return an empty
54 // result.
55 if (!Info)
56 return DILineInfo();
57
58 // If the user is giving us relative addresses, add the preferred base of the
59 // object to the offset before we do the query. It's what DIContext expects.
60 if (Opts.RelativeAddresses)
61 ModuleOffset.Address += Info->getModulePreferredBase();
62
63 DILineInfo LineInfo = Info->symbolizeCode(
64 ModuleOffset, DILineInfoSpecifier(Opts.PathStyle, Opts.PrintFunctions),
65 Opts.UseSymbolTable);
66 if (Opts.Demangle)
67 LineInfo.FunctionName = DemangleName(LineInfo.FunctionName, Info);
68 return LineInfo;
69 }
70
71 Expected<DILineInfo>
symbolizeCode(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)72 LLVMSymbolizer::symbolizeCode(const ObjectFile &Obj,
73 object::SectionedAddress ModuleOffset) {
74 return symbolizeCodeCommon(Obj, ModuleOffset);
75 }
76
77 Expected<DILineInfo>
symbolizeCode(const std::string & ModuleName,object::SectionedAddress ModuleOffset)78 LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
79 object::SectionedAddress ModuleOffset) {
80 return symbolizeCodeCommon(ModuleName, ModuleOffset);
81 }
82
83 template <typename T>
symbolizeInlinedCodeCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)84 Expected<DIInliningInfo> LLVMSymbolizer::symbolizeInlinedCodeCommon(
85 const T &ModuleSpecifier, object::SectionedAddress ModuleOffset) {
86 auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
87 if (!InfoOrErr)
88 return InfoOrErr.takeError();
89
90 SymbolizableModule *Info = *InfoOrErr;
91
92 // A null module means an error has already been reported. Return an empty
93 // result.
94 if (!Info)
95 return DIInliningInfo();
96
97 // If the user is giving us relative addresses, add the preferred base of the
98 // object to the offset before we do the query. It's what DIContext expects.
99 if (Opts.RelativeAddresses)
100 ModuleOffset.Address += Info->getModulePreferredBase();
101
102 DIInliningInfo InlinedContext = Info->symbolizeInlinedCode(
103 ModuleOffset, DILineInfoSpecifier(Opts.PathStyle, Opts.PrintFunctions),
104 Opts.UseSymbolTable);
105 if (Opts.Demangle) {
106 for (int i = 0, n = InlinedContext.getNumberOfFrames(); i < n; i++) {
107 auto *Frame = InlinedContext.getMutableFrame(i);
108 Frame->FunctionName = DemangleName(Frame->FunctionName, Info);
109 }
110 }
111 return InlinedContext;
112 }
113
114 Expected<DIInliningInfo>
symbolizeInlinedCode(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)115 LLVMSymbolizer::symbolizeInlinedCode(const ObjectFile &Obj,
116 object::SectionedAddress ModuleOffset) {
117 return symbolizeInlinedCodeCommon(Obj, ModuleOffset);
118 }
119
120 Expected<DIInliningInfo>
symbolizeInlinedCode(const std::string & ModuleName,object::SectionedAddress ModuleOffset)121 LLVMSymbolizer::symbolizeInlinedCode(const std::string &ModuleName,
122 object::SectionedAddress ModuleOffset) {
123 return symbolizeInlinedCodeCommon(ModuleName, ModuleOffset);
124 }
125
126 template <typename T>
127 Expected<DIGlobal>
symbolizeDataCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)128 LLVMSymbolizer::symbolizeDataCommon(const T &ModuleSpecifier,
129 object::SectionedAddress ModuleOffset) {
130
131 auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
132 if (!InfoOrErr)
133 return InfoOrErr.takeError();
134
135 SymbolizableModule *Info = *InfoOrErr;
136 // A null module means an error has already been reported. Return an empty
137 // result.
138 if (!Info)
139 return DIGlobal();
140
141 // If the user is giving us relative addresses, add the preferred base of
142 // the object to the offset before we do the query. It's what DIContext
143 // expects.
144 if (Opts.RelativeAddresses)
145 ModuleOffset.Address += Info->getModulePreferredBase();
146
147 DIGlobal Global = Info->symbolizeData(ModuleOffset);
148 if (Opts.Demangle)
149 Global.Name = DemangleName(Global.Name, Info);
150 return Global;
151 }
152
153 Expected<DIGlobal>
symbolizeData(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)154 LLVMSymbolizer::symbolizeData(const ObjectFile &Obj,
155 object::SectionedAddress ModuleOffset) {
156 return symbolizeDataCommon(Obj, ModuleOffset);
157 }
158
159 Expected<DIGlobal>
symbolizeData(const std::string & ModuleName,object::SectionedAddress ModuleOffset)160 LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
161 object::SectionedAddress ModuleOffset) {
162 return symbolizeDataCommon(ModuleName, ModuleOffset);
163 }
164
165 template <typename T>
166 Expected<std::vector<DILocal>>
symbolizeFrameCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)167 LLVMSymbolizer::symbolizeFrameCommon(const T &ModuleSpecifier,
168 object::SectionedAddress ModuleOffset) {
169 auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
170 if (!InfoOrErr)
171 return InfoOrErr.takeError();
172
173 SymbolizableModule *Info = *InfoOrErr;
174 // A null module means an error has already been reported. Return an empty
175 // result.
176 if (!Info)
177 return std::vector<DILocal>();
178
179 // If the user is giving us relative addresses, add the preferred base of
180 // the object to the offset before we do the query. It's what DIContext
181 // expects.
182 if (Opts.RelativeAddresses)
183 ModuleOffset.Address += Info->getModulePreferredBase();
184
185 return Info->symbolizeFrame(ModuleOffset);
186 }
187
188 Expected<std::vector<DILocal>>
symbolizeFrame(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)189 LLVMSymbolizer::symbolizeFrame(const ObjectFile &Obj,
190 object::SectionedAddress ModuleOffset) {
191 return symbolizeFrameCommon(Obj, ModuleOffset);
192 }
193
194 Expected<std::vector<DILocal>>
symbolizeFrame(const std::string & ModuleName,object::SectionedAddress ModuleOffset)195 LLVMSymbolizer::symbolizeFrame(const std::string &ModuleName,
196 object::SectionedAddress ModuleOffset) {
197 return symbolizeFrameCommon(ModuleName, ModuleOffset);
198 }
199
flush()200 void LLVMSymbolizer::flush() {
201 ObjectForUBPathAndArch.clear();
202 BinaryForPath.clear();
203 ObjectPairForPathArch.clear();
204 Modules.clear();
205 }
206
207 namespace {
208
209 // For Path="/path/to/foo" and Basename="foo" assume that debug info is in
210 // /path/to/foo.dSYM/Contents/Resources/DWARF/foo.
211 // For Path="/path/to/bar.dSYM" and Basename="foo" assume that debug info is in
212 // /path/to/bar.dSYM/Contents/Resources/DWARF/foo.
getDarwinDWARFResourceForPath(const std::string & Path,const std::string & Basename)213 std::string getDarwinDWARFResourceForPath(
214 const std::string &Path, const std::string &Basename) {
215 SmallString<16> ResourceName = StringRef(Path);
216 if (sys::path::extension(Path) != ".dSYM") {
217 ResourceName += ".dSYM";
218 }
219 sys::path::append(ResourceName, "Contents", "Resources", "DWARF");
220 sys::path::append(ResourceName, Basename);
221 return std::string(ResourceName.str());
222 }
223
checkFileCRC(StringRef Path,uint32_t CRCHash)224 bool checkFileCRC(StringRef Path, uint32_t CRCHash) {
225 ErrorOr<std::unique_ptr<MemoryBuffer>> MB =
226 MemoryBuffer::getFileOrSTDIN(Path);
227 if (!MB)
228 return false;
229 return CRCHash == llvm::crc32(arrayRefFromStringRef(MB.get()->getBuffer()));
230 }
231
findDebugBinary(const std::string & OrigPath,const std::string & DebuglinkName,uint32_t CRCHash,const std::string & FallbackDebugPath,std::string & Result)232 bool findDebugBinary(const std::string &OrigPath,
233 const std::string &DebuglinkName, uint32_t CRCHash,
234 const std::string &FallbackDebugPath,
235 std::string &Result) {
236 SmallString<16> OrigDir(OrigPath);
237 llvm::sys::path::remove_filename(OrigDir);
238 SmallString<16> DebugPath = OrigDir;
239 // Try relative/path/to/original_binary/debuglink_name
240 llvm::sys::path::append(DebugPath, DebuglinkName);
241 if (checkFileCRC(DebugPath, CRCHash)) {
242 Result = std::string(DebugPath.str());
243 return true;
244 }
245 // Try relative/path/to/original_binary/.debug/debuglink_name
246 DebugPath = OrigDir;
247 llvm::sys::path::append(DebugPath, ".debug", DebuglinkName);
248 if (checkFileCRC(DebugPath, CRCHash)) {
249 Result = std::string(DebugPath.str());
250 return true;
251 }
252 // Make the path absolute so that lookups will go to
253 // "/usr/lib/debug/full/path/to/debug", not
254 // "/usr/lib/debug/to/debug"
255 llvm::sys::fs::make_absolute(OrigDir);
256 if (!FallbackDebugPath.empty()) {
257 // Try <FallbackDebugPath>/absolute/path/to/original_binary/debuglink_name
258 DebugPath = FallbackDebugPath;
259 } else {
260 #if defined(__NetBSD__)
261 // Try /usr/libdata/debug/absolute/path/to/original_binary/debuglink_name
262 DebugPath = "/usr/libdata/debug";
263 #else
264 // Try /usr/lib/debug/absolute/path/to/original_binary/debuglink_name
265 DebugPath = "/usr/lib/debug";
266 #endif
267 }
268 llvm::sys::path::append(DebugPath, llvm::sys::path::relative_path(OrigDir),
269 DebuglinkName);
270 if (checkFileCRC(DebugPath, CRCHash)) {
271 Result = std::string(DebugPath.str());
272 return true;
273 }
274 return false;
275 }
276
getGNUDebuglinkContents(const ObjectFile * Obj,std::string & DebugName,uint32_t & CRCHash)277 bool getGNUDebuglinkContents(const ObjectFile *Obj, std::string &DebugName,
278 uint32_t &CRCHash) {
279 if (!Obj)
280 return false;
281 for (const SectionRef &Section : Obj->sections()) {
282 StringRef Name;
283 if (Expected<StringRef> NameOrErr = Section.getName())
284 Name = *NameOrErr;
285 else
286 consumeError(NameOrErr.takeError());
287
288 Name = Name.substr(Name.find_first_not_of("._"));
289 if (Name == "gnu_debuglink") {
290 Expected<StringRef> ContentsOrErr = Section.getContents();
291 if (!ContentsOrErr) {
292 consumeError(ContentsOrErr.takeError());
293 return false;
294 }
295 DataExtractor DE(*ContentsOrErr, Obj->isLittleEndian(), 0);
296 uint64_t Offset = 0;
297 if (const char *DebugNameStr = DE.getCStr(&Offset)) {
298 // 4-byte align the offset.
299 Offset = (Offset + 3) & ~0x3;
300 if (DE.isValidOffsetForDataOfSize(Offset, 4)) {
301 DebugName = DebugNameStr;
302 CRCHash = DE.getU32(&Offset);
303 return true;
304 }
305 }
306 break;
307 }
308 }
309 return false;
310 }
311
darwinDsymMatchesBinary(const MachOObjectFile * DbgObj,const MachOObjectFile * Obj)312 bool darwinDsymMatchesBinary(const MachOObjectFile *DbgObj,
313 const MachOObjectFile *Obj) {
314 ArrayRef<uint8_t> dbg_uuid = DbgObj->getUuid();
315 ArrayRef<uint8_t> bin_uuid = Obj->getUuid();
316 if (dbg_uuid.empty() || bin_uuid.empty())
317 return false;
318 return !memcmp(dbg_uuid.data(), bin_uuid.data(), dbg_uuid.size());
319 }
320
321 template <typename ELFT>
getBuildID(const ELFFile<ELFT> & Obj)322 Optional<ArrayRef<uint8_t>> getBuildID(const ELFFile<ELFT> &Obj) {
323 auto PhdrsOrErr = Obj.program_headers();
324 if (!PhdrsOrErr) {
325 consumeError(PhdrsOrErr.takeError());
326 return {};
327 }
328 for (const auto &P : *PhdrsOrErr) {
329 if (P.p_type != ELF::PT_NOTE)
330 continue;
331 Error Err = Error::success();
332 for (auto N : Obj.notes(P, Err))
333 if (N.getType() == ELF::NT_GNU_BUILD_ID && N.getName() == ELF::ELF_NOTE_GNU)
334 return N.getDesc();
335 consumeError(std::move(Err));
336 }
337 return {};
338 }
339
getBuildID(const ELFObjectFileBase * Obj)340 Optional<ArrayRef<uint8_t>> getBuildID(const ELFObjectFileBase *Obj) {
341 Optional<ArrayRef<uint8_t>> BuildID;
342 if (auto *O = dyn_cast<ELFObjectFile<ELF32LE>>(Obj))
343 BuildID = getBuildID(O->getELFFile());
344 else if (auto *O = dyn_cast<ELFObjectFile<ELF32BE>>(Obj))
345 BuildID = getBuildID(O->getELFFile());
346 else if (auto *O = dyn_cast<ELFObjectFile<ELF64LE>>(Obj))
347 BuildID = getBuildID(O->getELFFile());
348 else if (auto *O = dyn_cast<ELFObjectFile<ELF64BE>>(Obj))
349 BuildID = getBuildID(O->getELFFile());
350 else
351 llvm_unreachable("unsupported file format");
352 return BuildID;
353 }
354
findDebugBinary(const std::vector<std::string> & DebugFileDirectory,const ArrayRef<uint8_t> BuildID,std::string & Result)355 bool findDebugBinary(const std::vector<std::string> &DebugFileDirectory,
356 const ArrayRef<uint8_t> BuildID,
357 std::string &Result) {
358 auto getDebugPath = [&](StringRef Directory) {
359 SmallString<128> Path{Directory};
360 sys::path::append(Path, ".build-id",
361 llvm::toHex(BuildID[0], /*LowerCase=*/true),
362 llvm::toHex(BuildID.slice(1), /*LowerCase=*/true));
363 Path += ".debug";
364 return Path;
365 };
366 if (DebugFileDirectory.empty()) {
367 SmallString<128> Path = getDebugPath(
368 #if defined(__NetBSD__)
369 // Try /usr/libdata/debug/.build-id/../...
370 "/usr/libdata/debug"
371 #else
372 // Try /usr/lib/debug/.build-id/../...
373 "/usr/lib/debug"
374 #endif
375 );
376 if (llvm::sys::fs::exists(Path)) {
377 Result = std::string(Path.str());
378 return true;
379 }
380 } else {
381 for (const auto &Directory : DebugFileDirectory) {
382 // Try <debug-file-directory>/.build-id/../...
383 SmallString<128> Path = getDebugPath(Directory);
384 if (llvm::sys::fs::exists(Path)) {
385 Result = std::string(Path.str());
386 return true;
387 }
388 }
389 }
390 return false;
391 }
392
393 } // end anonymous namespace
394
lookUpDsymFile(const std::string & ExePath,const MachOObjectFile * MachExeObj,const std::string & ArchName)395 ObjectFile *LLVMSymbolizer::lookUpDsymFile(const std::string &ExePath,
396 const MachOObjectFile *MachExeObj, const std::string &ArchName) {
397 // On Darwin we may find DWARF in separate object file in
398 // resource directory.
399 std::vector<std::string> DsymPaths;
400 StringRef Filename = sys::path::filename(ExePath);
401 DsymPaths.push_back(
402 getDarwinDWARFResourceForPath(ExePath, std::string(Filename)));
403 for (const auto &Path : Opts.DsymHints) {
404 DsymPaths.push_back(
405 getDarwinDWARFResourceForPath(Path, std::string(Filename)));
406 }
407 for (const auto &Path : DsymPaths) {
408 auto DbgObjOrErr = getOrCreateObject(Path, ArchName);
409 if (!DbgObjOrErr) {
410 // Ignore errors, the file might not exist.
411 consumeError(DbgObjOrErr.takeError());
412 continue;
413 }
414 ObjectFile *DbgObj = DbgObjOrErr.get();
415 if (!DbgObj)
416 continue;
417 const MachOObjectFile *MachDbgObj = dyn_cast<const MachOObjectFile>(DbgObj);
418 if (!MachDbgObj)
419 continue;
420 if (darwinDsymMatchesBinary(MachDbgObj, MachExeObj))
421 return DbgObj;
422 }
423 return nullptr;
424 }
425
lookUpDebuglinkObject(const std::string & Path,const ObjectFile * Obj,const std::string & ArchName)426 ObjectFile *LLVMSymbolizer::lookUpDebuglinkObject(const std::string &Path,
427 const ObjectFile *Obj,
428 const std::string &ArchName) {
429 std::string DebuglinkName;
430 uint32_t CRCHash;
431 std::string DebugBinaryPath;
432 if (!getGNUDebuglinkContents(Obj, DebuglinkName, CRCHash))
433 return nullptr;
434 if (!findDebugBinary(Path, DebuglinkName, CRCHash, Opts.FallbackDebugPath,
435 DebugBinaryPath))
436 return nullptr;
437 auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
438 if (!DbgObjOrErr) {
439 // Ignore errors, the file might not exist.
440 consumeError(DbgObjOrErr.takeError());
441 return nullptr;
442 }
443 return DbgObjOrErr.get();
444 }
445
lookUpBuildIDObject(const std::string & Path,const ELFObjectFileBase * Obj,const std::string & ArchName)446 ObjectFile *LLVMSymbolizer::lookUpBuildIDObject(const std::string &Path,
447 const ELFObjectFileBase *Obj,
448 const std::string &ArchName) {
449 auto BuildID = getBuildID(Obj);
450 if (!BuildID)
451 return nullptr;
452 if (BuildID->size() < 2)
453 return nullptr;
454 std::string DebugBinaryPath;
455 if (!findDebugBinary(Opts.DebugFileDirectory, *BuildID, DebugBinaryPath))
456 return nullptr;
457 auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
458 if (!DbgObjOrErr) {
459 consumeError(DbgObjOrErr.takeError());
460 return nullptr;
461 }
462 return DbgObjOrErr.get();
463 }
464
465 Expected<LLVMSymbolizer::ObjectPair>
getOrCreateObjectPair(const std::string & Path,const std::string & ArchName)466 LLVMSymbolizer::getOrCreateObjectPair(const std::string &Path,
467 const std::string &ArchName) {
468 auto I = ObjectPairForPathArch.find(std::make_pair(Path, ArchName));
469 if (I != ObjectPairForPathArch.end())
470 return I->second;
471
472 auto ObjOrErr = getOrCreateObject(Path, ArchName);
473 if (!ObjOrErr) {
474 ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName),
475 ObjectPair(nullptr, nullptr));
476 return ObjOrErr.takeError();
477 }
478
479 ObjectFile *Obj = ObjOrErr.get();
480 assert(Obj != nullptr);
481 ObjectFile *DbgObj = nullptr;
482
483 if (auto MachObj = dyn_cast<const MachOObjectFile>(Obj))
484 DbgObj = lookUpDsymFile(Path, MachObj, ArchName);
485 else if (auto ELFObj = dyn_cast<const ELFObjectFileBase>(Obj))
486 DbgObj = lookUpBuildIDObject(Path, ELFObj, ArchName);
487 if (!DbgObj)
488 DbgObj = lookUpDebuglinkObject(Path, Obj, ArchName);
489 if (!DbgObj)
490 DbgObj = Obj;
491 ObjectPair Res = std::make_pair(Obj, DbgObj);
492 ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName), Res);
493 return Res;
494 }
495
496 Expected<ObjectFile *>
getOrCreateObject(const std::string & Path,const std::string & ArchName)497 LLVMSymbolizer::getOrCreateObject(const std::string &Path,
498 const std::string &ArchName) {
499 Binary *Bin;
500 auto Pair = BinaryForPath.emplace(Path, OwningBinary<Binary>());
501 if (!Pair.second) {
502 Bin = Pair.first->second.getBinary();
503 } else {
504 Expected<OwningBinary<Binary>> BinOrErr = createBinary(Path);
505 if (!BinOrErr)
506 return BinOrErr.takeError();
507 Pair.first->second = std::move(BinOrErr.get());
508 Bin = Pair.first->second.getBinary();
509 }
510
511 if (!Bin)
512 return static_cast<ObjectFile *>(nullptr);
513
514 if (MachOUniversalBinary *UB = dyn_cast_or_null<MachOUniversalBinary>(Bin)) {
515 auto I = ObjectForUBPathAndArch.find(std::make_pair(Path, ArchName));
516 if (I != ObjectForUBPathAndArch.end())
517 return I->second.get();
518
519 Expected<std::unique_ptr<ObjectFile>> ObjOrErr =
520 UB->getMachOObjectForArch(ArchName);
521 if (!ObjOrErr) {
522 ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
523 std::unique_ptr<ObjectFile>());
524 return ObjOrErr.takeError();
525 }
526 ObjectFile *Res = ObjOrErr->get();
527 ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
528 std::move(ObjOrErr.get()));
529 return Res;
530 }
531 if (Bin->isObject()) {
532 return cast<ObjectFile>(Bin);
533 }
534 return errorCodeToError(object_error::arch_not_found);
535 }
536
537 Expected<SymbolizableModule *>
createModuleInfo(const ObjectFile * Obj,std::unique_ptr<DIContext> Context,StringRef ModuleName)538 LLVMSymbolizer::createModuleInfo(const ObjectFile *Obj,
539 std::unique_ptr<DIContext> Context,
540 StringRef ModuleName) {
541 auto InfoOrErr = SymbolizableObjectFile::create(Obj, std::move(Context),
542 Opts.UntagAddresses);
543 std::unique_ptr<SymbolizableModule> SymMod;
544 if (InfoOrErr)
545 SymMod = std::move(*InfoOrErr);
546 auto InsertResult = Modules.insert(
547 std::make_pair(std::string(ModuleName), std::move(SymMod)));
548 assert(InsertResult.second);
549 if (!InfoOrErr)
550 return InfoOrErr.takeError();
551 return InsertResult.first->second.get();
552 }
553
554 Expected<SymbolizableModule *>
getOrCreateModuleInfo(const std::string & ModuleName)555 LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
556 auto I = Modules.find(ModuleName);
557 if (I != Modules.end())
558 return I->second.get();
559
560 std::string BinaryName = ModuleName;
561 std::string ArchName = Opts.DefaultArch;
562 size_t ColonPos = ModuleName.find_last_of(':');
563 // Verify that substring after colon form a valid arch name.
564 if (ColonPos != std::string::npos) {
565 std::string ArchStr = ModuleName.substr(ColonPos + 1);
566 if (Triple(ArchStr).getArch() != Triple::UnknownArch) {
567 BinaryName = ModuleName.substr(0, ColonPos);
568 ArchName = ArchStr;
569 }
570 }
571 auto ObjectsOrErr = getOrCreateObjectPair(BinaryName, ArchName);
572 if (!ObjectsOrErr) {
573 // Failed to find valid object file.
574 Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
575 return ObjectsOrErr.takeError();
576 }
577 ObjectPair Objects = ObjectsOrErr.get();
578
579 std::unique_ptr<DIContext> Context;
580 // If this is a COFF object containing PDB info, use a PDBContext to
581 // symbolize. Otherwise, use DWARF.
582 if (auto CoffObject = dyn_cast<COFFObjectFile>(Objects.first)) {
583 const codeview::DebugInfo *DebugInfo;
584 StringRef PDBFileName;
585 auto EC = CoffObject->getDebugPDBInfo(DebugInfo, PDBFileName);
586 if (!EC && DebugInfo != nullptr && !PDBFileName.empty()) {
587 using namespace pdb;
588 std::unique_ptr<IPDBSession> Session;
589
590 PDB_ReaderType ReaderType =
591 Opts.UseDIA ? PDB_ReaderType::DIA : PDB_ReaderType::Native;
592 if (auto Err = loadDataForEXE(ReaderType, Objects.first->getFileName(),
593 Session)) {
594 Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
595 // Return along the PDB filename to provide more context
596 return createFileError(PDBFileName, std::move(Err));
597 }
598 Context.reset(new PDBContext(*CoffObject, std::move(Session)));
599 }
600 }
601 if (!Context)
602 Context = DWARFContext::create(*Objects.second, nullptr, Opts.DWPName);
603 return createModuleInfo(Objects.first, std::move(Context), ModuleName);
604 }
605
606 Expected<SymbolizableModule *>
getOrCreateModuleInfo(const ObjectFile & Obj)607 LLVMSymbolizer::getOrCreateModuleInfo(const ObjectFile &Obj) {
608 StringRef ObjName = Obj.getFileName();
609 auto I = Modules.find(ObjName);
610 if (I != Modules.end())
611 return I->second.get();
612
613 std::unique_ptr<DIContext> Context = DWARFContext::create(Obj);
614 // FIXME: handle COFF object with PDB info to use PDBContext
615 return createModuleInfo(&Obj, std::move(Context), ObjName);
616 }
617
618 namespace {
619
620 // Undo these various manglings for Win32 extern "C" functions:
621 // cdecl - _foo
622 // stdcall - _foo@12
623 // fastcall - @foo@12
624 // vectorcall - foo@@12
625 // These are all different linkage names for 'foo'.
demanglePE32ExternCFunc(StringRef SymbolName)626 StringRef demanglePE32ExternCFunc(StringRef SymbolName) {
627 // Remove any '_' or '@' prefix.
628 char Front = SymbolName.empty() ? '\0' : SymbolName[0];
629 if (Front == '_' || Front == '@')
630 SymbolName = SymbolName.drop_front();
631
632 // Remove any '@[0-9]+' suffix.
633 if (Front != '?') {
634 size_t AtPos = SymbolName.rfind('@');
635 if (AtPos != StringRef::npos &&
636 all_of(drop_begin(SymbolName, AtPos + 1), isDigit))
637 SymbolName = SymbolName.substr(0, AtPos);
638 }
639
640 // Remove any ending '@' for vectorcall.
641 if (SymbolName.endswith("@"))
642 SymbolName = SymbolName.drop_back();
643
644 return SymbolName;
645 }
646
647 } // end anonymous namespace
648
649 std::string
DemangleName(const std::string & Name,const SymbolizableModule * DbiModuleDescriptor)650 LLVMSymbolizer::DemangleName(const std::string &Name,
651 const SymbolizableModule *DbiModuleDescriptor) {
652 // We can spoil names of symbols with C linkage, so use an heuristic
653 // approach to check if the name should be demangled.
654 if (Name.substr(0, 2) == "_Z") {
655 int status = 0;
656 char *DemangledName = itaniumDemangle(Name.c_str(), nullptr, nullptr, &status);
657 if (status != 0)
658 return Name;
659 std::string Result = DemangledName;
660 free(DemangledName);
661 return Result;
662 }
663
664 if (!Name.empty() && Name.front() == '?') {
665 // Only do MSVC C++ demangling on symbols starting with '?'.
666 int status = 0;
667 char *DemangledName = microsoftDemangle(
668 Name.c_str(), nullptr, nullptr, nullptr, &status,
669 MSDemangleFlags(MSDF_NoAccessSpecifier | MSDF_NoCallingConvention |
670 MSDF_NoMemberType | MSDF_NoReturnType));
671 if (status != 0)
672 return Name;
673 std::string Result = DemangledName;
674 free(DemangledName);
675 return Result;
676 }
677
678 if (DbiModuleDescriptor && DbiModuleDescriptor->isWin32Module())
679 return std::string(demanglePE32ExternCFunc(Name));
680 return Name;
681 }
682
683 } // namespace symbolize
684 } // namespace llvm
685