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 "llvm/ADT/STLExtras.h"
16 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
17 #include "llvm/DebugInfo/PDB/PDB.h"
18 #include "llvm/DebugInfo/PDB/PDBContext.h"
19 #include "llvm/DebugInfo/Symbolize/DIFetcher.h"
20 #include "llvm/DebugInfo/Symbolize/SymbolizableObjectFile.h"
21 #include "llvm/Demangle/Demangle.h"
22 #include "llvm/Object/COFF.h"
23 #include "llvm/Object/ELFObjectFile.h"
24 #include "llvm/Object/MachO.h"
25 #include "llvm/Object/MachOUniversal.h"
26 #include "llvm/Support/CRC.h"
27 #include "llvm/Support/Casting.h"
28 #include "llvm/Support/DataExtractor.h"
29 #include "llvm/Support/Errc.h"
30 #include "llvm/Support/FileSystem.h"
31 #include "llvm/Support/MemoryBuffer.h"
32 #include "llvm/Support/Path.h"
33 #include <algorithm>
34 #include <cassert>
35 #include <cstring>
36 
37 namespace llvm {
38 namespace codeview {
39 union DebugInfo;
40 }
41 namespace object {
42 template <class ELFT> class ELFFile;
43 }
44 namespace symbolize {
45 
46 LLVMSymbolizer::LLVMSymbolizer() = default;
47 
LLVMSymbolizer(const Options & Opts)48 LLVMSymbolizer::LLVMSymbolizer(const Options &Opts) : Opts(Opts) {}
49 
50 LLVMSymbolizer::~LLVMSymbolizer() = default;
51 
52 template <typename T>
53 Expected<DILineInfo>
symbolizeCodeCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)54 LLVMSymbolizer::symbolizeCodeCommon(const T &ModuleSpecifier,
55                                     object::SectionedAddress ModuleOffset) {
56 
57   auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
58   if (!InfoOrErr)
59     return InfoOrErr.takeError();
60 
61   SymbolizableModule *Info = *InfoOrErr;
62 
63   // A null module means an error has already been reported. Return an empty
64   // result.
65   if (!Info)
66     return DILineInfo();
67 
68   // If the user is giving us relative addresses, add the preferred base of the
69   // object to the offset before we do the query. It's what DIContext expects.
70   if (Opts.RelativeAddresses)
71     ModuleOffset.Address += Info->getModulePreferredBase();
72 
73   DILineInfo LineInfo = Info->symbolizeCode(
74       ModuleOffset, DILineInfoSpecifier(Opts.PathStyle, Opts.PrintFunctions),
75       Opts.UseSymbolTable);
76   if (Opts.Demangle)
77     LineInfo.FunctionName = DemangleName(LineInfo.FunctionName, Info);
78   return LineInfo;
79 }
80 
81 Expected<DILineInfo>
symbolizeCode(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)82 LLVMSymbolizer::symbolizeCode(const ObjectFile &Obj,
83                               object::SectionedAddress ModuleOffset) {
84   return symbolizeCodeCommon(Obj, ModuleOffset);
85 }
86 
87 Expected<DILineInfo>
symbolizeCode(const std::string & ModuleName,object::SectionedAddress ModuleOffset)88 LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
89                               object::SectionedAddress ModuleOffset) {
90   return symbolizeCodeCommon(ModuleName, ModuleOffset);
91 }
92 
93 Expected<DILineInfo>
symbolizeCode(ArrayRef<uint8_t> BuildID,object::SectionedAddress ModuleOffset)94 LLVMSymbolizer::symbolizeCode(ArrayRef<uint8_t> BuildID,
95                               object::SectionedAddress ModuleOffset) {
96   return symbolizeCodeCommon(BuildID, ModuleOffset);
97 }
98 
99 template <typename T>
symbolizeInlinedCodeCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)100 Expected<DIInliningInfo> LLVMSymbolizer::symbolizeInlinedCodeCommon(
101     const T &ModuleSpecifier, object::SectionedAddress ModuleOffset) {
102   auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
103   if (!InfoOrErr)
104     return InfoOrErr.takeError();
105 
106   SymbolizableModule *Info = *InfoOrErr;
107 
108   // A null module means an error has already been reported. Return an empty
109   // result.
110   if (!Info)
111     return DIInliningInfo();
112 
113   // If the user is giving us relative addresses, add the preferred base of the
114   // object to the offset before we do the query. It's what DIContext expects.
115   if (Opts.RelativeAddresses)
116     ModuleOffset.Address += Info->getModulePreferredBase();
117 
118   DIInliningInfo InlinedContext = Info->symbolizeInlinedCode(
119       ModuleOffset, DILineInfoSpecifier(Opts.PathStyle, Opts.PrintFunctions),
120       Opts.UseSymbolTable);
121   if (Opts.Demangle) {
122     for (int i = 0, n = InlinedContext.getNumberOfFrames(); i < n; i++) {
123       auto *Frame = InlinedContext.getMutableFrame(i);
124       Frame->FunctionName = DemangleName(Frame->FunctionName, Info);
125     }
126   }
127   return InlinedContext;
128 }
129 
130 Expected<DIInliningInfo>
symbolizeInlinedCode(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)131 LLVMSymbolizer::symbolizeInlinedCode(const ObjectFile &Obj,
132                                      object::SectionedAddress ModuleOffset) {
133   return symbolizeInlinedCodeCommon(Obj, ModuleOffset);
134 }
135 
136 Expected<DIInliningInfo>
symbolizeInlinedCode(const std::string & ModuleName,object::SectionedAddress ModuleOffset)137 LLVMSymbolizer::symbolizeInlinedCode(const std::string &ModuleName,
138                                      object::SectionedAddress ModuleOffset) {
139   return symbolizeInlinedCodeCommon(ModuleName, ModuleOffset);
140 }
141 
142 Expected<DIInliningInfo>
symbolizeInlinedCode(ArrayRef<uint8_t> BuildID,object::SectionedAddress ModuleOffset)143 LLVMSymbolizer::symbolizeInlinedCode(ArrayRef<uint8_t> BuildID,
144                                      object::SectionedAddress ModuleOffset) {
145   return symbolizeInlinedCodeCommon(BuildID, ModuleOffset);
146 }
147 
148 template <typename T>
149 Expected<DIGlobal>
symbolizeDataCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)150 LLVMSymbolizer::symbolizeDataCommon(const T &ModuleSpecifier,
151                                     object::SectionedAddress ModuleOffset) {
152 
153   auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
154   if (!InfoOrErr)
155     return InfoOrErr.takeError();
156 
157   SymbolizableModule *Info = *InfoOrErr;
158   // A null module means an error has already been reported. Return an empty
159   // result.
160   if (!Info)
161     return DIGlobal();
162 
163   // If the user is giving us relative addresses, add the preferred base of
164   // the object to the offset before we do the query. It's what DIContext
165   // expects.
166   if (Opts.RelativeAddresses)
167     ModuleOffset.Address += Info->getModulePreferredBase();
168 
169   DIGlobal Global = Info->symbolizeData(ModuleOffset);
170   if (Opts.Demangle)
171     Global.Name = DemangleName(Global.Name, Info);
172   return Global;
173 }
174 
175 Expected<DIGlobal>
symbolizeData(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)176 LLVMSymbolizer::symbolizeData(const ObjectFile &Obj,
177                               object::SectionedAddress ModuleOffset) {
178   return symbolizeDataCommon(Obj, ModuleOffset);
179 }
180 
181 Expected<DIGlobal>
symbolizeData(const std::string & ModuleName,object::SectionedAddress ModuleOffset)182 LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
183                               object::SectionedAddress ModuleOffset) {
184   return symbolizeDataCommon(ModuleName, ModuleOffset);
185 }
186 
187 Expected<DIGlobal>
symbolizeData(ArrayRef<uint8_t> BuildID,object::SectionedAddress ModuleOffset)188 LLVMSymbolizer::symbolizeData(ArrayRef<uint8_t> BuildID,
189                               object::SectionedAddress ModuleOffset) {
190   return symbolizeDataCommon(BuildID, ModuleOffset);
191 }
192 
193 template <typename T>
194 Expected<std::vector<DILocal>>
symbolizeFrameCommon(const T & ModuleSpecifier,object::SectionedAddress ModuleOffset)195 LLVMSymbolizer::symbolizeFrameCommon(const T &ModuleSpecifier,
196                                      object::SectionedAddress ModuleOffset) {
197   auto InfoOrErr = getOrCreateModuleInfo(ModuleSpecifier);
198   if (!InfoOrErr)
199     return InfoOrErr.takeError();
200 
201   SymbolizableModule *Info = *InfoOrErr;
202   // A null module means an error has already been reported. Return an empty
203   // result.
204   if (!Info)
205     return std::vector<DILocal>();
206 
207   // If the user is giving us relative addresses, add the preferred base of
208   // the object to the offset before we do the query. It's what DIContext
209   // expects.
210   if (Opts.RelativeAddresses)
211     ModuleOffset.Address += Info->getModulePreferredBase();
212 
213   return Info->symbolizeFrame(ModuleOffset);
214 }
215 
216 Expected<std::vector<DILocal>>
symbolizeFrame(const ObjectFile & Obj,object::SectionedAddress ModuleOffset)217 LLVMSymbolizer::symbolizeFrame(const ObjectFile &Obj,
218                                object::SectionedAddress ModuleOffset) {
219   return symbolizeFrameCommon(Obj, ModuleOffset);
220 }
221 
222 Expected<std::vector<DILocal>>
symbolizeFrame(const std::string & ModuleName,object::SectionedAddress ModuleOffset)223 LLVMSymbolizer::symbolizeFrame(const std::string &ModuleName,
224                                object::SectionedAddress ModuleOffset) {
225   return symbolizeFrameCommon(ModuleName, ModuleOffset);
226 }
227 
228 Expected<std::vector<DILocal>>
symbolizeFrame(ArrayRef<uint8_t> BuildID,object::SectionedAddress ModuleOffset)229 LLVMSymbolizer::symbolizeFrame(ArrayRef<uint8_t> BuildID,
230                                object::SectionedAddress ModuleOffset) {
231   return symbolizeFrameCommon(BuildID, ModuleOffset);
232 }
233 
flush()234 void LLVMSymbolizer::flush() {
235   ObjectForUBPathAndArch.clear();
236   LRUBinaries.clear();
237   CacheSize = 0;
238   BinaryForPath.clear();
239   ObjectPairForPathArch.clear();
240   Modules.clear();
241   BuildIDPaths.clear();
242 }
243 
244 namespace {
245 
246 // For Path="/path/to/foo" and Basename="foo" assume that debug info is in
247 // /path/to/foo.dSYM/Contents/Resources/DWARF/foo.
248 // For Path="/path/to/bar.dSYM" and Basename="foo" assume that debug info is in
249 // /path/to/bar.dSYM/Contents/Resources/DWARF/foo.
getDarwinDWARFResourceForPath(const std::string & Path,const std::string & Basename)250 std::string getDarwinDWARFResourceForPath(const std::string &Path,
251                                           const std::string &Basename) {
252   SmallString<16> ResourceName = StringRef(Path);
253   if (sys::path::extension(Path) != ".dSYM") {
254     ResourceName += ".dSYM";
255   }
256   sys::path::append(ResourceName, "Contents", "Resources", "DWARF");
257   sys::path::append(ResourceName, Basename);
258   return std::string(ResourceName.str());
259 }
260 
checkFileCRC(StringRef Path,uint32_t CRCHash)261 bool checkFileCRC(StringRef Path, uint32_t CRCHash) {
262   ErrorOr<std::unique_ptr<MemoryBuffer>> MB =
263       MemoryBuffer::getFileOrSTDIN(Path);
264   if (!MB)
265     return false;
266   return CRCHash == llvm::crc32(arrayRefFromStringRef(MB.get()->getBuffer()));
267 }
268 
getGNUDebuglinkContents(const ObjectFile * Obj,std::string & DebugName,uint32_t & CRCHash)269 bool getGNUDebuglinkContents(const ObjectFile *Obj, std::string &DebugName,
270                              uint32_t &CRCHash) {
271   if (!Obj)
272     return false;
273   for (const SectionRef &Section : Obj->sections()) {
274     StringRef Name;
275     consumeError(Section.getName().moveInto(Name));
276 
277     Name = Name.substr(Name.find_first_not_of("._"));
278     if (Name == "gnu_debuglink") {
279       Expected<StringRef> ContentsOrErr = Section.getContents();
280       if (!ContentsOrErr) {
281         consumeError(ContentsOrErr.takeError());
282         return false;
283       }
284       DataExtractor DE(*ContentsOrErr, Obj->isLittleEndian(), 0);
285       uint64_t Offset = 0;
286       if (const char *DebugNameStr = DE.getCStr(&Offset)) {
287         // 4-byte align the offset.
288         Offset = (Offset + 3) & ~0x3;
289         if (DE.isValidOffsetForDataOfSize(Offset, 4)) {
290           DebugName = DebugNameStr;
291           CRCHash = DE.getU32(&Offset);
292           return true;
293         }
294       }
295       break;
296     }
297   }
298   return false;
299 }
300 
darwinDsymMatchesBinary(const MachOObjectFile * DbgObj,const MachOObjectFile * Obj)301 bool darwinDsymMatchesBinary(const MachOObjectFile *DbgObj,
302                              const MachOObjectFile *Obj) {
303   ArrayRef<uint8_t> dbg_uuid = DbgObj->getUuid();
304   ArrayRef<uint8_t> bin_uuid = Obj->getUuid();
305   if (dbg_uuid.empty() || bin_uuid.empty())
306     return false;
307   return !memcmp(dbg_uuid.data(), bin_uuid.data(), dbg_uuid.size());
308 }
309 
310 template <typename ELFT>
getBuildID(const ELFFile<ELFT> & Obj)311 Optional<ArrayRef<uint8_t>> getBuildID(const ELFFile<ELFT> &Obj) {
312   auto PhdrsOrErr = Obj.program_headers();
313   if (!PhdrsOrErr) {
314     consumeError(PhdrsOrErr.takeError());
315     return {};
316   }
317   for (const auto &P : *PhdrsOrErr) {
318     if (P.p_type != ELF::PT_NOTE)
319       continue;
320     Error Err = Error::success();
321     for (auto N : Obj.notes(P, Err))
322       if (N.getType() == ELF::NT_GNU_BUILD_ID &&
323           N.getName() == ELF::ELF_NOTE_GNU)
324         return N.getDesc();
325     consumeError(std::move(Err));
326   }
327   return {};
328 }
329 
330 } // end anonymous namespace
331 
getBuildID(const ELFObjectFileBase * Obj)332 Optional<ArrayRef<uint8_t>> getBuildID(const ELFObjectFileBase *Obj) {
333   Optional<ArrayRef<uint8_t>> BuildID;
334   if (auto *O = dyn_cast<ELFObjectFile<ELF32LE>>(Obj))
335     BuildID = getBuildID(O->getELFFile());
336   else if (auto *O = dyn_cast<ELFObjectFile<ELF32BE>>(Obj))
337     BuildID = getBuildID(O->getELFFile());
338   else if (auto *O = dyn_cast<ELFObjectFile<ELF64LE>>(Obj))
339     BuildID = getBuildID(O->getELFFile());
340   else if (auto *O = dyn_cast<ELFObjectFile<ELF64BE>>(Obj))
341     BuildID = getBuildID(O->getELFFile());
342   else
343     llvm_unreachable("unsupported file format");
344   return BuildID;
345 }
346 
lookUpDsymFile(const std::string & ExePath,const MachOObjectFile * MachExeObj,const std::string & ArchName)347 ObjectFile *LLVMSymbolizer::lookUpDsymFile(const std::string &ExePath,
348                                            const MachOObjectFile *MachExeObj,
349                                            const std::string &ArchName) {
350   // On Darwin we may find DWARF in separate object file in
351   // resource directory.
352   std::vector<std::string> DsymPaths;
353   StringRef Filename = sys::path::filename(ExePath);
354   DsymPaths.push_back(
355       getDarwinDWARFResourceForPath(ExePath, std::string(Filename)));
356   for (const auto &Path : Opts.DsymHints) {
357     DsymPaths.push_back(
358         getDarwinDWARFResourceForPath(Path, std::string(Filename)));
359   }
360   for (const auto &Path : DsymPaths) {
361     auto DbgObjOrErr = getOrCreateObject(Path, ArchName);
362     if (!DbgObjOrErr) {
363       // Ignore errors, the file might not exist.
364       consumeError(DbgObjOrErr.takeError());
365       continue;
366     }
367     ObjectFile *DbgObj = DbgObjOrErr.get();
368     if (!DbgObj)
369       continue;
370     const MachOObjectFile *MachDbgObj = dyn_cast<const MachOObjectFile>(DbgObj);
371     if (!MachDbgObj)
372       continue;
373     if (darwinDsymMatchesBinary(MachDbgObj, MachExeObj))
374       return DbgObj;
375   }
376   return nullptr;
377 }
378 
lookUpDebuglinkObject(const std::string & Path,const ObjectFile * Obj,const std::string & ArchName)379 ObjectFile *LLVMSymbolizer::lookUpDebuglinkObject(const std::string &Path,
380                                                   const ObjectFile *Obj,
381                                                   const std::string &ArchName) {
382   std::string DebuglinkName;
383   uint32_t CRCHash;
384   std::string DebugBinaryPath;
385   if (!getGNUDebuglinkContents(Obj, DebuglinkName, CRCHash))
386     return nullptr;
387   if (!findDebugBinary(Path, DebuglinkName, CRCHash, DebugBinaryPath))
388     return nullptr;
389   auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
390   if (!DbgObjOrErr) {
391     // Ignore errors, the file might not exist.
392     consumeError(DbgObjOrErr.takeError());
393     return nullptr;
394   }
395   return DbgObjOrErr.get();
396 }
397 
lookUpBuildIDObject(const std::string & Path,const ELFObjectFileBase * Obj,const std::string & ArchName)398 ObjectFile *LLVMSymbolizer::lookUpBuildIDObject(const std::string &Path,
399                                                 const ELFObjectFileBase *Obj,
400                                                 const std::string &ArchName) {
401   auto BuildID = getBuildID(Obj);
402   if (!BuildID)
403     return nullptr;
404   if (BuildID->size() < 2)
405     return nullptr;
406   std::string DebugBinaryPath;
407   if (!getOrFindDebugBinary(*BuildID, DebugBinaryPath))
408     return nullptr;
409   auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
410   if (!DbgObjOrErr) {
411     consumeError(DbgObjOrErr.takeError());
412     return nullptr;
413   }
414   return DbgObjOrErr.get();
415 }
416 
findDebugBinary(const std::string & OrigPath,const std::string & DebuglinkName,uint32_t CRCHash,std::string & Result)417 bool LLVMSymbolizer::findDebugBinary(const std::string &OrigPath,
418                                      const std::string &DebuglinkName,
419                                      uint32_t CRCHash, std::string &Result) {
420   SmallString<16> OrigDir(OrigPath);
421   llvm::sys::path::remove_filename(OrigDir);
422   SmallString<16> DebugPath = OrigDir;
423   // Try relative/path/to/original_binary/debuglink_name
424   llvm::sys::path::append(DebugPath, DebuglinkName);
425   if (checkFileCRC(DebugPath, CRCHash)) {
426     Result = std::string(DebugPath.str());
427     return true;
428   }
429   // Try relative/path/to/original_binary/.debug/debuglink_name
430   DebugPath = OrigDir;
431   llvm::sys::path::append(DebugPath, ".debug", DebuglinkName);
432   if (checkFileCRC(DebugPath, CRCHash)) {
433     Result = std::string(DebugPath.str());
434     return true;
435   }
436   // Make the path absolute so that lookups will go to
437   // "/usr/lib/debug/full/path/to/debug", not
438   // "/usr/lib/debug/to/debug"
439   llvm::sys::fs::make_absolute(OrigDir);
440   if (!Opts.FallbackDebugPath.empty()) {
441     // Try <FallbackDebugPath>/absolute/path/to/original_binary/debuglink_name
442     DebugPath = Opts.FallbackDebugPath;
443   } else {
444 #if defined(__NetBSD__)
445     // Try /usr/libdata/debug/absolute/path/to/original_binary/debuglink_name
446     DebugPath = "/usr/libdata/debug";
447 #else
448     // Try /usr/lib/debug/absolute/path/to/original_binary/debuglink_name
449     DebugPath = "/usr/lib/debug";
450 #endif
451   }
452   llvm::sys::path::append(DebugPath, llvm::sys::path::relative_path(OrigDir),
453                           DebuglinkName);
454   if (checkFileCRC(DebugPath, CRCHash)) {
455     Result = std::string(DebugPath.str());
456     return true;
457   }
458   return false;
459 }
460 
getBuildIDStr(ArrayRef<uint8_t> BuildID)461 static StringRef getBuildIDStr(ArrayRef<uint8_t> BuildID) {
462   return StringRef(reinterpret_cast<const char *>(BuildID.data()),
463                    BuildID.size());
464 }
465 
getOrFindDebugBinary(const ArrayRef<uint8_t> BuildID,std::string & Result)466 bool LLVMSymbolizer::getOrFindDebugBinary(const ArrayRef<uint8_t> BuildID,
467                                           std::string &Result) {
468   StringRef BuildIDStr = getBuildIDStr(BuildID);
469   auto I = BuildIDPaths.find(BuildIDStr);
470   if (I != BuildIDPaths.end()) {
471     Result = I->second;
472     return true;
473   }
474   auto recordPath = [&](StringRef Path) {
475     Result = Path.str();
476     auto InsertResult = BuildIDPaths.insert({BuildIDStr, Result});
477     assert(InsertResult.second);
478     (void)InsertResult;
479   };
480 
481   Optional<std::string> Path;
482   Path = LocalDIFetcher(Opts.DebugFileDirectory).fetchBuildID(BuildID);
483   if (Path) {
484     recordPath(*Path);
485     return true;
486   }
487 
488   // Try caller-provided debug info fetchers.
489   for (const std::unique_ptr<DIFetcher> &Fetcher : DIFetchers) {
490     Path = Fetcher->fetchBuildID(BuildID);
491     if (Path) {
492       recordPath(*Path);
493       return true;
494     }
495   }
496 
497   return false;
498 }
499 
500 Expected<LLVMSymbolizer::ObjectPair>
getOrCreateObjectPair(const std::string & Path,const std::string & ArchName)501 LLVMSymbolizer::getOrCreateObjectPair(const std::string &Path,
502                                       const std::string &ArchName) {
503   auto I = ObjectPairForPathArch.find(std::make_pair(Path, ArchName));
504   if (I != ObjectPairForPathArch.end()) {
505     recordAccess(BinaryForPath.find(Path)->second);
506     return I->second;
507   }
508 
509   auto ObjOrErr = getOrCreateObject(Path, ArchName);
510   if (!ObjOrErr) {
511     ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName),
512                                   ObjectPair(nullptr, nullptr));
513     return ObjOrErr.takeError();
514   }
515 
516   ObjectFile *Obj = ObjOrErr.get();
517   assert(Obj != nullptr);
518   ObjectFile *DbgObj = nullptr;
519 
520   if (auto MachObj = dyn_cast<const MachOObjectFile>(Obj))
521     DbgObj = lookUpDsymFile(Path, MachObj, ArchName);
522   else if (auto ELFObj = dyn_cast<const ELFObjectFileBase>(Obj))
523     DbgObj = lookUpBuildIDObject(Path, ELFObj, ArchName);
524   if (!DbgObj)
525     DbgObj = lookUpDebuglinkObject(Path, Obj, ArchName);
526   if (!DbgObj)
527     DbgObj = Obj;
528   ObjectPair Res = std::make_pair(Obj, DbgObj);
529   std::string DbgObjPath = DbgObj->getFileName().str();
530   auto Pair =
531       ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName), Res);
532   BinaryForPath.find(DbgObjPath)->second.pushEvictor([this, I = Pair.first]() {
533     ObjectPairForPathArch.erase(I);
534   });
535   return Res;
536 }
537 
538 Expected<ObjectFile *>
getOrCreateObject(const std::string & Path,const std::string & ArchName)539 LLVMSymbolizer::getOrCreateObject(const std::string &Path,
540                                   const std::string &ArchName) {
541   Binary *Bin;
542   auto Pair = BinaryForPath.emplace(Path, OwningBinary<Binary>());
543   if (!Pair.second) {
544     Bin = Pair.first->second->getBinary();
545     recordAccess(Pair.first->second);
546   } else {
547     Expected<OwningBinary<Binary>> BinOrErr = createBinary(Path);
548     if (!BinOrErr)
549       return BinOrErr.takeError();
550 
551     CachedBinary &CachedBin = Pair.first->second;
552     CachedBin = std::move(BinOrErr.get());
553     CachedBin.pushEvictor([this, I = Pair.first]() { BinaryForPath.erase(I); });
554     LRUBinaries.push_back(CachedBin);
555     CacheSize += CachedBin.size();
556     Bin = CachedBin->getBinary();
557   }
558 
559   if (!Bin)
560     return static_cast<ObjectFile *>(nullptr);
561 
562   if (MachOUniversalBinary *UB = dyn_cast_or_null<MachOUniversalBinary>(Bin)) {
563     auto I = ObjectForUBPathAndArch.find(std::make_pair(Path, ArchName));
564     if (I != ObjectForUBPathAndArch.end())
565       return I->second.get();
566 
567     Expected<std::unique_ptr<ObjectFile>> ObjOrErr =
568         UB->getMachOObjectForArch(ArchName);
569     if (!ObjOrErr) {
570       ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
571                                      std::unique_ptr<ObjectFile>());
572       return ObjOrErr.takeError();
573     }
574     ObjectFile *Res = ObjOrErr->get();
575     auto Pair = ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
576                                                std::move(ObjOrErr.get()));
577     BinaryForPath.find(Path)->second.pushEvictor(
578         [this, Iter = Pair.first]() { ObjectForUBPathAndArch.erase(Iter); });
579     return Res;
580   }
581   if (Bin->isObject()) {
582     return cast<ObjectFile>(Bin);
583   }
584   return errorCodeToError(object_error::arch_not_found);
585 }
586 
587 Expected<SymbolizableModule *>
createModuleInfo(const ObjectFile * Obj,std::unique_ptr<DIContext> Context,StringRef ModuleName)588 LLVMSymbolizer::createModuleInfo(const ObjectFile *Obj,
589                                  std::unique_ptr<DIContext> Context,
590                                  StringRef ModuleName) {
591   auto InfoOrErr = SymbolizableObjectFile::create(Obj, std::move(Context),
592                                                   Opts.UntagAddresses);
593   std::unique_ptr<SymbolizableModule> SymMod;
594   if (InfoOrErr)
595     SymMod = std::move(*InfoOrErr);
596   auto InsertResult = Modules.insert(
597       std::make_pair(std::string(ModuleName), std::move(SymMod)));
598   assert(InsertResult.second);
599   if (!InfoOrErr)
600     return InfoOrErr.takeError();
601   return InsertResult.first->second.get();
602 }
603 
604 Expected<SymbolizableModule *>
getOrCreateModuleInfo(const std::string & ModuleName)605 LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
606   std::string BinaryName = ModuleName;
607   std::string ArchName = Opts.DefaultArch;
608   size_t ColonPos = ModuleName.find_last_of(':');
609   // Verify that substring after colon form a valid arch name.
610   if (ColonPos != std::string::npos) {
611     std::string ArchStr = ModuleName.substr(ColonPos + 1);
612     if (Triple(ArchStr).getArch() != Triple::UnknownArch) {
613       BinaryName = ModuleName.substr(0, ColonPos);
614       ArchName = ArchStr;
615     }
616   }
617 
618   auto I = Modules.find(ModuleName);
619   if (I != Modules.end()) {
620     recordAccess(BinaryForPath.find(BinaryName)->second);
621     return I->second.get();
622   }
623 
624   auto ObjectsOrErr = getOrCreateObjectPair(BinaryName, ArchName);
625   if (!ObjectsOrErr) {
626     // Failed to find valid object file.
627     Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
628     return ObjectsOrErr.takeError();
629   }
630   ObjectPair Objects = ObjectsOrErr.get();
631 
632   std::unique_ptr<DIContext> Context;
633   // If this is a COFF object containing PDB info, use a PDBContext to
634   // symbolize. Otherwise, use DWARF.
635   if (auto CoffObject = dyn_cast<COFFObjectFile>(Objects.first)) {
636     const codeview::DebugInfo *DebugInfo;
637     StringRef PDBFileName;
638     auto EC = CoffObject->getDebugPDBInfo(DebugInfo, PDBFileName);
639     if (!EC && DebugInfo != nullptr && !PDBFileName.empty()) {
640       using namespace pdb;
641       std::unique_ptr<IPDBSession> Session;
642 
643       PDB_ReaderType ReaderType =
644           Opts.UseDIA ? PDB_ReaderType::DIA : PDB_ReaderType::Native;
645       if (auto Err = loadDataForEXE(ReaderType, Objects.first->getFileName(),
646                                     Session)) {
647         Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
648         // Return along the PDB filename to provide more context
649         return createFileError(PDBFileName, std::move(Err));
650       }
651       Context.reset(new PDBContext(*CoffObject, std::move(Session)));
652     }
653   }
654   if (!Context)
655     Context = DWARFContext::create(
656         *Objects.second, DWARFContext::ProcessDebugRelocations::Process,
657         nullptr, Opts.DWPName);
658   auto ModuleOrErr =
659       createModuleInfo(Objects.first, std::move(Context), ModuleName);
660   if (ModuleOrErr) {
661     auto I = Modules.find(ModuleName);
662     BinaryForPath.find(BinaryName)->second.pushEvictor([this, I]() {
663       Modules.erase(I);
664     });
665   }
666   return ModuleOrErr;
667 }
668 
669 Expected<SymbolizableModule *>
getOrCreateModuleInfo(const ObjectFile & Obj)670 LLVMSymbolizer::getOrCreateModuleInfo(const ObjectFile &Obj) {
671   StringRef ObjName = Obj.getFileName();
672   auto I = Modules.find(ObjName);
673   if (I != Modules.end())
674     return I->second.get();
675 
676   std::unique_ptr<DIContext> Context = DWARFContext::create(Obj);
677   // FIXME: handle COFF object with PDB info to use PDBContext
678   return createModuleInfo(&Obj, std::move(Context), ObjName);
679 }
680 
681 Expected<SymbolizableModule *>
getOrCreateModuleInfo(ArrayRef<uint8_t> BuildID)682 LLVMSymbolizer::getOrCreateModuleInfo(ArrayRef<uint8_t> BuildID) {
683   std::string Path;
684   if (!getOrFindDebugBinary(BuildID, Path)) {
685     return createStringError(errc::no_such_file_or_directory,
686                              Twine("could not find build ID '") +
687                                  toHex(BuildID) + "'");
688   }
689   return getOrCreateModuleInfo(Path);
690 }
691 
692 namespace {
693 
694 // Undo these various manglings for Win32 extern "C" functions:
695 // cdecl       - _foo
696 // stdcall     - _foo@12
697 // fastcall    - @foo@12
698 // vectorcall  - foo@@12
699 // These are all different linkage names for 'foo'.
demanglePE32ExternCFunc(StringRef SymbolName)700 StringRef demanglePE32ExternCFunc(StringRef SymbolName) {
701   // Remove any '_' or '@' prefix.
702   char Front = SymbolName.empty() ? '\0' : SymbolName[0];
703   if (Front == '_' || Front == '@')
704     SymbolName = SymbolName.drop_front();
705 
706   // Remove any '@[0-9]+' suffix.
707   if (Front != '?') {
708     size_t AtPos = SymbolName.rfind('@');
709     if (AtPos != StringRef::npos &&
710         all_of(drop_begin(SymbolName, AtPos + 1), isDigit))
711       SymbolName = SymbolName.substr(0, AtPos);
712   }
713 
714   // Remove any ending '@' for vectorcall.
715   if (SymbolName.endswith("@"))
716     SymbolName = SymbolName.drop_back();
717 
718   return SymbolName;
719 }
720 
721 } // end anonymous namespace
722 
723 std::string
DemangleName(const std::string & Name,const SymbolizableModule * DbiModuleDescriptor)724 LLVMSymbolizer::DemangleName(const std::string &Name,
725                              const SymbolizableModule *DbiModuleDescriptor) {
726   std::string Result;
727   if (nonMicrosoftDemangle(Name.c_str(), Result))
728     return Result;
729 
730   if (!Name.empty() && Name.front() == '?') {
731     // Only do MSVC C++ demangling on symbols starting with '?'.
732     int status = 0;
733     char *DemangledName = microsoftDemangle(
734         Name.c_str(), nullptr, nullptr, nullptr, &status,
735         MSDemangleFlags(MSDF_NoAccessSpecifier | MSDF_NoCallingConvention |
736                         MSDF_NoMemberType | MSDF_NoReturnType));
737     if (status != 0)
738       return Name;
739     Result = DemangledName;
740     free(DemangledName);
741     return Result;
742   }
743 
744   if (DbiModuleDescriptor && DbiModuleDescriptor->isWin32Module())
745     return std::string(demanglePE32ExternCFunc(Name));
746   return Name;
747 }
748 
recordAccess(CachedBinary & Bin)749 void LLVMSymbolizer::recordAccess(CachedBinary &Bin) {
750   if (Bin->getBinary())
751     LRUBinaries.splice(LRUBinaries.end(), LRUBinaries, Bin.getIterator());
752 }
753 
pruneCache()754 void LLVMSymbolizer::pruneCache() {
755   // Evict the LRU binary until the max cache size is reached or there's <= 1
756   // item in the cache. The MRU binary is always kept to avoid thrashing if it's
757   // larger than the cache size.
758   while (CacheSize > Opts.MaxCacheSize && !LRUBinaries.empty() &&
759          std::next(LRUBinaries.begin()) != LRUBinaries.end()) {
760     CachedBinary &Bin = LRUBinaries.front();
761     CacheSize -= Bin.size();
762     LRUBinaries.pop_front();
763     Bin.evict();
764   }
765 }
766 
pushEvictor(std::function<void ()> NewEvictor)767 void CachedBinary::pushEvictor(std::function<void()> NewEvictor) {
768   if (Evictor) {
769     this->Evictor = [OldEvictor = std::move(this->Evictor),
770                      NewEvictor = std::move(NewEvictor)]() {
771       NewEvictor();
772       OldEvictor();
773     };
774   } else {
775     this->Evictor = std::move(NewEvictor);
776   }
777 }
778 
779 } // namespace symbolize
780 } // namespace llvm
781