xref: /freebsd/contrib/llvm-project/llvm/lib/DebugInfo/Symbolize/Symbolize.cpp (revision 59c8e88e72633afbc47a4ace0d2170d00d51f7dc) 
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/BTF/BTFContext.h"
17 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
18 #include "llvm/DebugInfo/PDB/PDB.h"
19 #include "llvm/DebugInfo/PDB/PDBContext.h"
20 #include "llvm/DebugInfo/Symbolize/SymbolizableObjectFile.h"
21 #include "llvm/Demangle/Demangle.h"
22 #include "llvm/Object/BuildID.h"
23 #include "llvm/Object/COFF.h"
24 #include "llvm/Object/ELFObjectFile.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/DataExtractor.h"
30 #include "llvm/Support/Errc.h"
31 #include "llvm/Support/FileSystem.h"
32 #include "llvm/Support/MemoryBuffer.h"
33 #include "llvm/Support/Path.h"
34 #include <algorithm>
35 #include <cassert>
36 #include <cstring>
37 
38 namespace llvm {
39 namespace codeview {
40 union DebugInfo;
41 }
42 namespace symbolize {
43 
44 LLVMSymbolizer::LLVMSymbolizer() = default;
45 
46 LLVMSymbolizer::LLVMSymbolizer(const Options &Opts)
47     : Opts(Opts),
48       BIDFetcher(std::make_unique<BuildIDFetcher>(Opts.DebugFileDirectory)) {}
49 
50 LLVMSymbolizer::~LLVMSymbolizer() = default;
51 
52 template <typename T>
53 Expected<DILineInfo>
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>
82 LLVMSymbolizer::symbolizeCode(const ObjectFile &Obj,
83                               object::SectionedAddress ModuleOffset) {
84   return symbolizeCodeCommon(Obj, ModuleOffset);
85 }
86 
87 Expected<DILineInfo>
88 LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
89                               object::SectionedAddress ModuleOffset) {
90   return symbolizeCodeCommon(ModuleName, ModuleOffset);
91 }
92 
93 Expected<DILineInfo>
94 LLVMSymbolizer::symbolizeCode(ArrayRef<uint8_t> BuildID,
95                               object::SectionedAddress ModuleOffset) {
96   return symbolizeCodeCommon(BuildID, ModuleOffset);
97 }
98 
99 template <typename T>
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>
131 LLVMSymbolizer::symbolizeInlinedCode(const ObjectFile &Obj,
132                                      object::SectionedAddress ModuleOffset) {
133   return symbolizeInlinedCodeCommon(Obj, ModuleOffset);
134 }
135 
136 Expected<DIInliningInfo>
137 LLVMSymbolizer::symbolizeInlinedCode(const std::string &ModuleName,
138                                      object::SectionedAddress ModuleOffset) {
139   return symbolizeInlinedCodeCommon(ModuleName, ModuleOffset);
140 }
141 
142 Expected<DIInliningInfo>
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>
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>
176 LLVMSymbolizer::symbolizeData(const ObjectFile &Obj,
177                               object::SectionedAddress ModuleOffset) {
178   return symbolizeDataCommon(Obj, ModuleOffset);
179 }
180 
181 Expected<DIGlobal>
182 LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
183                               object::SectionedAddress ModuleOffset) {
184   return symbolizeDataCommon(ModuleName, ModuleOffset);
185 }
186 
187 Expected<DIGlobal>
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>>
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>>
217 LLVMSymbolizer::symbolizeFrame(const ObjectFile &Obj,
218                                object::SectionedAddress ModuleOffset) {
219   return symbolizeFrameCommon(Obj, ModuleOffset);
220 }
221 
222 Expected<std::vector<DILocal>>
223 LLVMSymbolizer::symbolizeFrame(const std::string &ModuleName,
224                                object::SectionedAddress ModuleOffset) {
225   return symbolizeFrameCommon(ModuleName, ModuleOffset);
226 }
227 
228 Expected<std::vector<DILocal>>
229 LLVMSymbolizer::symbolizeFrame(ArrayRef<uint8_t> BuildID,
230                                object::SectionedAddress ModuleOffset) {
231   return symbolizeFrameCommon(BuildID, ModuleOffset);
232 }
233 
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.
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 
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 
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 
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 } // end anonymous namespace
311 
312 ObjectFile *LLVMSymbolizer::lookUpDsymFile(const std::string &ExePath,
313                                            const MachOObjectFile *MachExeObj,
314                                            const std::string &ArchName) {
315   // On Darwin we may find DWARF in separate object file in
316   // resource directory.
317   std::vector<std::string> DsymPaths;
318   StringRef Filename = sys::path::filename(ExePath);
319   DsymPaths.push_back(
320       getDarwinDWARFResourceForPath(ExePath, std::string(Filename)));
321   for (const auto &Path : Opts.DsymHints) {
322     DsymPaths.push_back(
323         getDarwinDWARFResourceForPath(Path, std::string(Filename)));
324   }
325   for (const auto &Path : DsymPaths) {
326     auto DbgObjOrErr = getOrCreateObject(Path, ArchName);
327     if (!DbgObjOrErr) {
328       // Ignore errors, the file might not exist.
329       consumeError(DbgObjOrErr.takeError());
330       continue;
331     }
332     ObjectFile *DbgObj = DbgObjOrErr.get();
333     if (!DbgObj)
334       continue;
335     const MachOObjectFile *MachDbgObj = dyn_cast<const MachOObjectFile>(DbgObj);
336     if (!MachDbgObj)
337       continue;
338     if (darwinDsymMatchesBinary(MachDbgObj, MachExeObj))
339       return DbgObj;
340   }
341   return nullptr;
342 }
343 
344 ObjectFile *LLVMSymbolizer::lookUpDebuglinkObject(const std::string &Path,
345                                                   const ObjectFile *Obj,
346                                                   const std::string &ArchName) {
347   std::string DebuglinkName;
348   uint32_t CRCHash;
349   std::string DebugBinaryPath;
350   if (!getGNUDebuglinkContents(Obj, DebuglinkName, CRCHash))
351     return nullptr;
352   if (!findDebugBinary(Path, DebuglinkName, CRCHash, DebugBinaryPath))
353     return nullptr;
354   auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
355   if (!DbgObjOrErr) {
356     // Ignore errors, the file might not exist.
357     consumeError(DbgObjOrErr.takeError());
358     return nullptr;
359   }
360   return DbgObjOrErr.get();
361 }
362 
363 ObjectFile *LLVMSymbolizer::lookUpBuildIDObject(const std::string &Path,
364                                                 const ELFObjectFileBase *Obj,
365                                                 const std::string &ArchName) {
366   auto BuildID = getBuildID(Obj);
367   if (BuildID.size() < 2)
368     return nullptr;
369   std::string DebugBinaryPath;
370   if (!getOrFindDebugBinary(BuildID, DebugBinaryPath))
371     return nullptr;
372   auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
373   if (!DbgObjOrErr) {
374     consumeError(DbgObjOrErr.takeError());
375     return nullptr;
376   }
377   return DbgObjOrErr.get();
378 }
379 
380 bool LLVMSymbolizer::findDebugBinary(const std::string &OrigPath,
381                                      const std::string &DebuglinkName,
382                                      uint32_t CRCHash, std::string &Result) {
383   SmallString<16> OrigDir(OrigPath);
384   llvm::sys::path::remove_filename(OrigDir);
385   SmallString<16> DebugPath = OrigDir;
386   // Try relative/path/to/original_binary/debuglink_name
387   llvm::sys::path::append(DebugPath, DebuglinkName);
388   if (checkFileCRC(DebugPath, CRCHash)) {
389     Result = std::string(DebugPath.str());
390     return true;
391   }
392   // Try relative/path/to/original_binary/.debug/debuglink_name
393   DebugPath = OrigDir;
394   llvm::sys::path::append(DebugPath, ".debug", DebuglinkName);
395   if (checkFileCRC(DebugPath, CRCHash)) {
396     Result = std::string(DebugPath.str());
397     return true;
398   }
399   // Make the path absolute so that lookups will go to
400   // "/usr/lib/debug/full/path/to/debug", not
401   // "/usr/lib/debug/to/debug"
402   llvm::sys::fs::make_absolute(OrigDir);
403   if (!Opts.FallbackDebugPath.empty()) {
404     // Try <FallbackDebugPath>/absolute/path/to/original_binary/debuglink_name
405     DebugPath = Opts.FallbackDebugPath;
406   } else {
407 #if defined(__NetBSD__)
408     // Try /usr/libdata/debug/absolute/path/to/original_binary/debuglink_name
409     DebugPath = "/usr/libdata/debug";
410 #else
411     // Try /usr/lib/debug/absolute/path/to/original_binary/debuglink_name
412     DebugPath = "/usr/lib/debug";
413 #endif
414   }
415   llvm::sys::path::append(DebugPath, llvm::sys::path::relative_path(OrigDir),
416                           DebuglinkName);
417   if (checkFileCRC(DebugPath, CRCHash)) {
418     Result = std::string(DebugPath.str());
419     return true;
420   }
421   return false;
422 }
423 
424 static StringRef getBuildIDStr(ArrayRef<uint8_t> BuildID) {
425   return StringRef(reinterpret_cast<const char *>(BuildID.data()),
426                    BuildID.size());
427 }
428 
429 bool LLVMSymbolizer::getOrFindDebugBinary(const ArrayRef<uint8_t> BuildID,
430                                           std::string &Result) {
431   StringRef BuildIDStr = getBuildIDStr(BuildID);
432   auto I = BuildIDPaths.find(BuildIDStr);
433   if (I != BuildIDPaths.end()) {
434     Result = I->second;
435     return true;
436   }
437   if (!BIDFetcher)
438     return false;
439   if (std::optional<std::string> Path = BIDFetcher->fetch(BuildID)) {
440     Result = *Path;
441     auto InsertResult = BuildIDPaths.insert({BuildIDStr, Result});
442     assert(InsertResult.second);
443     (void)InsertResult;
444     return true;
445   }
446 
447   return false;
448 }
449 
450 Expected<LLVMSymbolizer::ObjectPair>
451 LLVMSymbolizer::getOrCreateObjectPair(const std::string &Path,
452                                       const std::string &ArchName) {
453   auto I = ObjectPairForPathArch.find(std::make_pair(Path, ArchName));
454   if (I != ObjectPairForPathArch.end()) {
455     recordAccess(BinaryForPath.find(Path)->second);
456     return I->second;
457   }
458 
459   auto ObjOrErr = getOrCreateObject(Path, ArchName);
460   if (!ObjOrErr) {
461     ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName),
462                                   ObjectPair(nullptr, nullptr));
463     return ObjOrErr.takeError();
464   }
465 
466   ObjectFile *Obj = ObjOrErr.get();
467   assert(Obj != nullptr);
468   ObjectFile *DbgObj = nullptr;
469 
470   if (auto MachObj = dyn_cast<const MachOObjectFile>(Obj))
471     DbgObj = lookUpDsymFile(Path, MachObj, ArchName);
472   else if (auto ELFObj = dyn_cast<const ELFObjectFileBase>(Obj))
473     DbgObj = lookUpBuildIDObject(Path, ELFObj, ArchName);
474   if (!DbgObj)
475     DbgObj = lookUpDebuglinkObject(Path, Obj, ArchName);
476   if (!DbgObj)
477     DbgObj = Obj;
478   ObjectPair Res = std::make_pair(Obj, DbgObj);
479   std::string DbgObjPath = DbgObj->getFileName().str();
480   auto Pair =
481       ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName), Res);
482   BinaryForPath.find(DbgObjPath)->second.pushEvictor([this, I = Pair.first]() {
483     ObjectPairForPathArch.erase(I);
484   });
485   return Res;
486 }
487 
488 Expected<ObjectFile *>
489 LLVMSymbolizer::getOrCreateObject(const std::string &Path,
490                                   const std::string &ArchName) {
491   Binary *Bin;
492   auto Pair = BinaryForPath.emplace(Path, OwningBinary<Binary>());
493   if (!Pair.second) {
494     Bin = Pair.first->second->getBinary();
495     recordAccess(Pair.first->second);
496   } else {
497     Expected<OwningBinary<Binary>> BinOrErr = createBinary(Path);
498     if (!BinOrErr)
499       return BinOrErr.takeError();
500 
501     CachedBinary &CachedBin = Pair.first->second;
502     CachedBin = std::move(BinOrErr.get());
503     CachedBin.pushEvictor([this, I = Pair.first]() { BinaryForPath.erase(I); });
504     LRUBinaries.push_back(CachedBin);
505     CacheSize += CachedBin.size();
506     Bin = CachedBin->getBinary();
507   }
508 
509   if (!Bin)
510     return static_cast<ObjectFile *>(nullptr);
511 
512   if (MachOUniversalBinary *UB = dyn_cast_or_null<MachOUniversalBinary>(Bin)) {
513     auto I = ObjectForUBPathAndArch.find(std::make_pair(Path, ArchName));
514     if (I != ObjectForUBPathAndArch.end())
515       return I->second.get();
516 
517     Expected<std::unique_ptr<ObjectFile>> ObjOrErr =
518         UB->getMachOObjectForArch(ArchName);
519     if (!ObjOrErr) {
520       ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
521                                      std::unique_ptr<ObjectFile>());
522       return ObjOrErr.takeError();
523     }
524     ObjectFile *Res = ObjOrErr->get();
525     auto Pair = ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
526                                                std::move(ObjOrErr.get()));
527     BinaryForPath.find(Path)->second.pushEvictor(
528         [this, Iter = Pair.first]() { ObjectForUBPathAndArch.erase(Iter); });
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 *>
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 *>
555 LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
556   std::string BinaryName = ModuleName;
557   std::string ArchName = Opts.DefaultArch;
558   size_t ColonPos = ModuleName.find_last_of(':');
559   // Verify that substring after colon form a valid arch name.
560   if (ColonPos != std::string::npos) {
561     std::string ArchStr = ModuleName.substr(ColonPos + 1);
562     if (Triple(ArchStr).getArch() != Triple::UnknownArch) {
563       BinaryName = ModuleName.substr(0, ColonPos);
564       ArchName = ArchStr;
565     }
566   }
567 
568   auto I = Modules.find(ModuleName);
569   if (I != Modules.end()) {
570     recordAccess(BinaryForPath.find(BinaryName)->second);
571     return I->second.get();
572   }
573 
574   auto ObjectsOrErr = getOrCreateObjectPair(BinaryName, ArchName);
575   if (!ObjectsOrErr) {
576     // Failed to find valid object file.
577     Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
578     return ObjectsOrErr.takeError();
579   }
580   ObjectPair Objects = ObjectsOrErr.get();
581 
582   std::unique_ptr<DIContext> Context;
583   // If this is a COFF object containing PDB info, use a PDBContext to
584   // symbolize. Otherwise, use DWARF.
585   if (auto CoffObject = dyn_cast<COFFObjectFile>(Objects.first)) {
586     const codeview::DebugInfo *DebugInfo;
587     StringRef PDBFileName;
588     auto EC = CoffObject->getDebugPDBInfo(DebugInfo, PDBFileName);
589     if (!EC && DebugInfo != nullptr && !PDBFileName.empty()) {
590 #if 0
591       using namespace pdb;
592       std::unique_ptr<IPDBSession> Session;
593 
594       PDB_ReaderType ReaderType =
595           Opts.UseDIA ? PDB_ReaderType::DIA : PDB_ReaderType::Native;
596       if (auto Err = loadDataForEXE(ReaderType, Objects.first->getFileName(),
597                                     Session)) {
598         Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
599         // Return along the PDB filename to provide more context
600         return createFileError(PDBFileName, std::move(Err));
601       }
602       Context.reset(new PDBContext(*CoffObject, std::move(Session)));
603 #else
604       return make_error<StringError>(
605           "PDB support not compiled in",
606           std::make_error_code(std::errc::not_supported));
607 #endif
608     }
609   }
610   if (!Context)
611     Context = DWARFContext::create(
612         *Objects.second, DWARFContext::ProcessDebugRelocations::Process,
613         nullptr, Opts.DWPName);
614   auto ModuleOrErr =
615       createModuleInfo(Objects.first, std::move(Context), ModuleName);
616   if (ModuleOrErr) {
617     auto I = Modules.find(ModuleName);
618     BinaryForPath.find(BinaryName)->second.pushEvictor([this, I]() {
619       Modules.erase(I);
620     });
621   }
622   return ModuleOrErr;
623 }
624 
625 // For BPF programs .BTF.ext section contains line numbers information,
626 // use it if regular DWARF is not available (e.g. for stripped binary).
627 static bool useBTFContext(const ObjectFile &Obj) {
628   return Obj.makeTriple().isBPF() && !Obj.hasDebugInfo() &&
629          BTFParser::hasBTFSections(Obj);
630 }
631 
632 Expected<SymbolizableModule *>
633 LLVMSymbolizer::getOrCreateModuleInfo(const ObjectFile &Obj) {
634   StringRef ObjName = Obj.getFileName();
635   auto I = Modules.find(ObjName);
636   if (I != Modules.end())
637     return I->second.get();
638 
639   std::unique_ptr<DIContext> Context;
640   if (useBTFContext(Obj))
641     Context = BTFContext::create(Obj);
642   else
643     Context = DWARFContext::create(Obj);
644   // FIXME: handle COFF object with PDB info to use PDBContext
645   return createModuleInfo(&Obj, std::move(Context), ObjName);
646 }
647 
648 Expected<SymbolizableModule *>
649 LLVMSymbolizer::getOrCreateModuleInfo(ArrayRef<uint8_t> BuildID) {
650   std::string Path;
651   if (!getOrFindDebugBinary(BuildID, Path)) {
652     return createStringError(errc::no_such_file_or_directory,
653                              "could not find build ID");
654   }
655   return getOrCreateModuleInfo(Path);
656 }
657 
658 namespace {
659 
660 // Undo these various manglings for Win32 extern "C" functions:
661 // cdecl       - _foo
662 // stdcall     - _foo@12
663 // fastcall    - @foo@12
664 // vectorcall  - foo@@12
665 // These are all different linkage names for 'foo'.
666 StringRef demanglePE32ExternCFunc(StringRef SymbolName) {
667   char Front = SymbolName.empty() ? '\0' : SymbolName[0];
668 
669   // Remove any '@[0-9]+' suffix.
670   bool HasAtNumSuffix = false;
671   if (Front != '?') {
672     size_t AtPos = SymbolName.rfind('@');
673     if (AtPos != StringRef::npos &&
674         all_of(drop_begin(SymbolName, AtPos + 1), isDigit)) {
675       SymbolName = SymbolName.substr(0, AtPos);
676       HasAtNumSuffix = true;
677     }
678   }
679 
680   // Remove any ending '@' for vectorcall.
681   bool IsVectorCall = false;
682   if (HasAtNumSuffix && SymbolName.endswith("@")) {
683     SymbolName = SymbolName.drop_back();
684     IsVectorCall = true;
685   }
686 
687   // If not vectorcall, remove any '_' or '@' prefix.
688   if (!IsVectorCall && (Front == '_' || Front == '@'))
689     SymbolName = SymbolName.drop_front();
690 
691   return SymbolName;
692 }
693 
694 } // end anonymous namespace
695 
696 std::string
697 LLVMSymbolizer::DemangleName(const std::string &Name,
698                              const SymbolizableModule *DbiModuleDescriptor) {
699   std::string Result;
700   if (nonMicrosoftDemangle(Name, Result))
701     return Result;
702 
703   if (!Name.empty() && Name.front() == '?') {
704     // Only do MSVC C++ demangling on symbols starting with '?'.
705     int status = 0;
706     char *DemangledName = microsoftDemangle(
707         Name, nullptr, &status,
708         MSDemangleFlags(MSDF_NoAccessSpecifier | MSDF_NoCallingConvention |
709                         MSDF_NoMemberType | MSDF_NoReturnType));
710     if (status != 0)
711       return Name;
712     Result = DemangledName;
713     free(DemangledName);
714     return Result;
715   }
716 
717   if (DbiModuleDescriptor && DbiModuleDescriptor->isWin32Module()) {
718     std::string DemangledCName(demanglePE32ExternCFunc(Name));
719     // On i386 Windows, the C name mangling for different calling conventions
720     // may also be applied on top of the Itanium or Rust name mangling.
721     if (nonMicrosoftDemangle(DemangledCName, Result))
722       return Result;
723     return DemangledCName;
724   }
725   return Name;
726 }
727 
728 void LLVMSymbolizer::recordAccess(CachedBinary &Bin) {
729   if (Bin->getBinary())
730     LRUBinaries.splice(LRUBinaries.end(), LRUBinaries, Bin.getIterator());
731 }
732 
733 void LLVMSymbolizer::pruneCache() {
734   // Evict the LRU binary until the max cache size is reached or there's <= 1
735   // item in the cache. The MRU binary is always kept to avoid thrashing if it's
736   // larger than the cache size.
737   while (CacheSize > Opts.MaxCacheSize && !LRUBinaries.empty() &&
738          std::next(LRUBinaries.begin()) != LRUBinaries.end()) {
739     CachedBinary &Bin = LRUBinaries.front();
740     CacheSize -= Bin.size();
741     LRUBinaries.pop_front();
742     Bin.evict();
743   }
744 }
745 
746 void CachedBinary::pushEvictor(std::function<void()> NewEvictor) {
747   if (Evictor) {
748     this->Evictor = [OldEvictor = std::move(this->Evictor),
749                      NewEvictor = std::move(NewEvictor)]() {
750       NewEvictor();
751       OldEvictor();
752     };
753   } else {
754     this->Evictor = std::move(NewEvictor);
755   }
756 }
757 
758 } // namespace symbolize
759 } // namespace llvm
760