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/DebugInfo/DWARF/DWARFContext.h"
20 #include "llvm/DebugInfo/PDB/PDB.h"
21 #include "llvm/DebugInfo/PDB/PDBContext.h"
22 #include "llvm/Demangle/Demangle.h"
23 #include "llvm/Object/COFF.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/Compression.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 #if defined(_MSC_VER)
39 #include <Windows.h>
40 
41 // This must be included after windows.h.
42 #include <DbgHelp.h>
43 #pragma comment(lib, "dbghelp.lib")
44 
45 // Windows.h conflicts with our COFF header definitions.
46 #ifdef IMAGE_FILE_MACHINE_I386
47 #undef IMAGE_FILE_MACHINE_I386
48 #endif
49 #endif
50 
51 namespace llvm {
52 namespace symbolize {
53 
54 Expected<DILineInfo>
55 LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
56                               object::SectionedAddress ModuleOffset) {
57   SymbolizableModule *Info;
58   if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
59     Info = InfoOrErr.get();
60   else
61     return InfoOrErr.takeError();
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(ModuleOffset, Opts.PrintFunctions,
74                                             Opts.UseSymbolTable);
75   if (Opts.Demangle)
76     LineInfo.FunctionName = DemangleName(LineInfo.FunctionName, Info);
77   return LineInfo;
78 }
79 
80 Expected<DIInliningInfo>
81 LLVMSymbolizer::symbolizeInlinedCode(const std::string &ModuleName,
82                                      object::SectionedAddress ModuleOffset) {
83   SymbolizableModule *Info;
84   if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
85     Info = InfoOrErr.get();
86   else
87     return InfoOrErr.takeError();
88 
89   // A null module means an error has already been reported. Return an empty
90   // result.
91   if (!Info)
92     return DIInliningInfo();
93 
94   // If the user is giving us relative addresses, add the preferred base of the
95   // object to the offset before we do the query. It's what DIContext expects.
96   if (Opts.RelativeAddresses)
97     ModuleOffset.Address += Info->getModulePreferredBase();
98 
99   DIInliningInfo InlinedContext = Info->symbolizeInlinedCode(
100       ModuleOffset, Opts.PrintFunctions, Opts.UseSymbolTable);
101   if (Opts.Demangle) {
102     for (int i = 0, n = InlinedContext.getNumberOfFrames(); i < n; i++) {
103       auto *Frame = InlinedContext.getMutableFrame(i);
104       Frame->FunctionName = DemangleName(Frame->FunctionName, Info);
105     }
106   }
107   return InlinedContext;
108 }
109 
110 Expected<DIGlobal>
111 LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
112                               object::SectionedAddress ModuleOffset) {
113   SymbolizableModule *Info;
114   if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
115     Info = InfoOrErr.get();
116   else
117     return InfoOrErr.takeError();
118 
119   // A null module means an error has already been reported. Return an empty
120   // result.
121   if (!Info)
122     return DIGlobal();
123 
124   // If the user is giving us relative addresses, add the preferred base of
125   // the object to the offset before we do the query. It's what DIContext
126   // expects.
127   if (Opts.RelativeAddresses)
128     ModuleOffset.Address += Info->getModulePreferredBase();
129 
130   DIGlobal Global = Info->symbolizeData(ModuleOffset);
131   if (Opts.Demangle)
132     Global.Name = DemangleName(Global.Name, Info);
133   return Global;
134 }
135 
136 Expected<std::vector<DILocal>>
137 LLVMSymbolizer::symbolizeFrame(const std::string &ModuleName,
138                                object::SectionedAddress ModuleOffset) {
139   SymbolizableModule *Info;
140   if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
141     Info = InfoOrErr.get();
142   else
143     return InfoOrErr.takeError();
144 
145   // A null module means an error has already been reported. Return an empty
146   // result.
147   if (!Info)
148     return std::vector<DILocal>();
149 
150   // If the user is giving us relative addresses, add the preferred base of
151   // the object to the offset before we do the query. It's what DIContext
152   // expects.
153   if (Opts.RelativeAddresses)
154     ModuleOffset.Address += Info->getModulePreferredBase();
155 
156   return Info->symbolizeFrame(ModuleOffset);
157 }
158 
159 void LLVMSymbolizer::flush() {
160   ObjectForUBPathAndArch.clear();
161   BinaryForPath.clear();
162   ObjectPairForPathArch.clear();
163   Modules.clear();
164 }
165 
166 namespace {
167 
168 // For Path="/path/to/foo" and Basename="foo" assume that debug info is in
169 // /path/to/foo.dSYM/Contents/Resources/DWARF/foo.
170 // For Path="/path/to/bar.dSYM" and Basename="foo" assume that debug info is in
171 // /path/to/bar.dSYM/Contents/Resources/DWARF/foo.
172 std::string getDarwinDWARFResourceForPath(
173     const std::string &Path, const std::string &Basename) {
174   SmallString<16> ResourceName = StringRef(Path);
175   if (sys::path::extension(Path) != ".dSYM") {
176     ResourceName += ".dSYM";
177   }
178   sys::path::append(ResourceName, "Contents", "Resources", "DWARF");
179   sys::path::append(ResourceName, Basename);
180   return ResourceName.str();
181 }
182 
183 bool checkFileCRC(StringRef Path, uint32_t CRCHash) {
184   ErrorOr<std::unique_ptr<MemoryBuffer>> MB =
185       MemoryBuffer::getFileOrSTDIN(Path);
186   if (!MB)
187     return false;
188   return CRCHash == llvm::crc32(0, MB.get()->getBuffer());
189 }
190 
191 bool findDebugBinary(const std::string &OrigPath,
192                      const std::string &DebuglinkName, uint32_t CRCHash,
193                      const std::string &FallbackDebugPath,
194                      std::string &Result) {
195   SmallString<16> OrigDir(OrigPath);
196   llvm::sys::path::remove_filename(OrigDir);
197   SmallString<16> DebugPath = OrigDir;
198   // Try relative/path/to/original_binary/debuglink_name
199   llvm::sys::path::append(DebugPath, DebuglinkName);
200   if (checkFileCRC(DebugPath, CRCHash)) {
201     Result = DebugPath.str();
202     return true;
203   }
204   // Try relative/path/to/original_binary/.debug/debuglink_name
205   DebugPath = OrigDir;
206   llvm::sys::path::append(DebugPath, ".debug", DebuglinkName);
207   if (checkFileCRC(DebugPath, CRCHash)) {
208     Result = DebugPath.str();
209     return true;
210   }
211   // Make the path absolute so that lookups will go to
212   // "/usr/lib/debug/full/path/to/debug", not
213   // "/usr/lib/debug/to/debug"
214   llvm::sys::fs::make_absolute(OrigDir);
215   if (!FallbackDebugPath.empty()) {
216     // Try <FallbackDebugPath>/absolute/path/to/original_binary/debuglink_name
217     DebugPath = FallbackDebugPath;
218   } else {
219 #if defined(__NetBSD__)
220     // Try /usr/libdata/debug/absolute/path/to/original_binary/debuglink_name
221     DebugPath = "/usr/libdata/debug";
222 #else
223     // Try /usr/lib/debug/absolute/path/to/original_binary/debuglink_name
224     DebugPath = "/usr/lib/debug";
225 #endif
226   }
227   llvm::sys::path::append(DebugPath, llvm::sys::path::relative_path(OrigDir),
228                           DebuglinkName);
229   if (checkFileCRC(DebugPath, CRCHash)) {
230     Result = DebugPath.str();
231     return true;
232   }
233   return false;
234 }
235 
236 bool getGNUDebuglinkContents(const ObjectFile *Obj, std::string &DebugName,
237                              uint32_t &CRCHash) {
238   if (!Obj)
239     return false;
240   for (const SectionRef &Section : Obj->sections()) {
241     StringRef Name;
242     Section.getName(Name);
243     Name = Name.substr(Name.find_first_not_of("._"));
244     if (Name == "gnu_debuglink") {
245       Expected<StringRef> ContentsOrErr = Section.getContents();
246       if (!ContentsOrErr) {
247         consumeError(ContentsOrErr.takeError());
248         return false;
249       }
250       DataExtractor DE(*ContentsOrErr, Obj->isLittleEndian(), 0);
251       uint32_t Offset = 0;
252       if (const char *DebugNameStr = DE.getCStr(&Offset)) {
253         // 4-byte align the offset.
254         Offset = (Offset + 3) & ~0x3;
255         if (DE.isValidOffsetForDataOfSize(Offset, 4)) {
256           DebugName = DebugNameStr;
257           CRCHash = DE.getU32(&Offset);
258           return true;
259         }
260       }
261       break;
262     }
263   }
264   return false;
265 }
266 
267 bool darwinDsymMatchesBinary(const MachOObjectFile *DbgObj,
268                              const MachOObjectFile *Obj) {
269   ArrayRef<uint8_t> dbg_uuid = DbgObj->getUuid();
270   ArrayRef<uint8_t> bin_uuid = Obj->getUuid();
271   if (dbg_uuid.empty() || bin_uuid.empty())
272     return false;
273   return !memcmp(dbg_uuid.data(), bin_uuid.data(), dbg_uuid.size());
274 }
275 
276 } // end anonymous namespace
277 
278 ObjectFile *LLVMSymbolizer::lookUpDsymFile(const std::string &ExePath,
279     const MachOObjectFile *MachExeObj, const std::string &ArchName) {
280   // On Darwin we may find DWARF in separate object file in
281   // resource directory.
282   std::vector<std::string> DsymPaths;
283   StringRef Filename = sys::path::filename(ExePath);
284   DsymPaths.push_back(getDarwinDWARFResourceForPath(ExePath, Filename));
285   for (const auto &Path : Opts.DsymHints) {
286     DsymPaths.push_back(getDarwinDWARFResourceForPath(Path, Filename));
287   }
288   for (const auto &Path : DsymPaths) {
289     auto DbgObjOrErr = getOrCreateObject(Path, ArchName);
290     if (!DbgObjOrErr) {
291       // Ignore errors, the file might not exist.
292       consumeError(DbgObjOrErr.takeError());
293       continue;
294     }
295     ObjectFile *DbgObj = DbgObjOrErr.get();
296     if (!DbgObj)
297       continue;
298     const MachOObjectFile *MachDbgObj = dyn_cast<const MachOObjectFile>(DbgObj);
299     if (!MachDbgObj)
300       continue;
301     if (darwinDsymMatchesBinary(MachDbgObj, MachExeObj))
302       return DbgObj;
303   }
304   return nullptr;
305 }
306 
307 ObjectFile *LLVMSymbolizer::lookUpDebuglinkObject(const std::string &Path,
308                                                   const ObjectFile *Obj,
309                                                   const std::string &ArchName) {
310   std::string DebuglinkName;
311   uint32_t CRCHash;
312   std::string DebugBinaryPath;
313   if (!getGNUDebuglinkContents(Obj, DebuglinkName, CRCHash))
314     return nullptr;
315   if (!findDebugBinary(Path, DebuglinkName, CRCHash, Opts.FallbackDebugPath,
316                        DebugBinaryPath))
317     return nullptr;
318   auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
319   if (!DbgObjOrErr) {
320     // Ignore errors, the file might not exist.
321     consumeError(DbgObjOrErr.takeError());
322     return nullptr;
323   }
324   return DbgObjOrErr.get();
325 }
326 
327 Expected<LLVMSymbolizer::ObjectPair>
328 LLVMSymbolizer::getOrCreateObjectPair(const std::string &Path,
329                                       const std::string &ArchName) {
330   auto I = ObjectPairForPathArch.find(std::make_pair(Path, ArchName));
331   if (I != ObjectPairForPathArch.end())
332     return I->second;
333 
334   auto ObjOrErr = getOrCreateObject(Path, ArchName);
335   if (!ObjOrErr) {
336     ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName),
337                                   ObjectPair(nullptr, nullptr));
338     return ObjOrErr.takeError();
339   }
340 
341   ObjectFile *Obj = ObjOrErr.get();
342   assert(Obj != nullptr);
343   ObjectFile *DbgObj = nullptr;
344 
345   if (auto MachObj = dyn_cast<const MachOObjectFile>(Obj))
346     DbgObj = lookUpDsymFile(Path, MachObj, ArchName);
347   if (!DbgObj)
348     DbgObj = lookUpDebuglinkObject(Path, Obj, ArchName);
349   if (!DbgObj)
350     DbgObj = Obj;
351   ObjectPair Res = std::make_pair(Obj, DbgObj);
352   ObjectPairForPathArch.emplace(std::make_pair(Path, ArchName), Res);
353   return Res;
354 }
355 
356 Expected<ObjectFile *>
357 LLVMSymbolizer::getOrCreateObject(const std::string &Path,
358                                   const std::string &ArchName) {
359   Binary *Bin;
360   auto Pair = BinaryForPath.emplace(Path, OwningBinary<Binary>());
361   if (!Pair.second) {
362     Bin = Pair.first->second.getBinary();
363   } else {
364     Expected<OwningBinary<Binary>> BinOrErr = createBinary(Path);
365     if (!BinOrErr)
366       return BinOrErr.takeError();
367     Pair.first->second = std::move(BinOrErr.get());
368     Bin = Pair.first->second.getBinary();
369   }
370 
371   if (!Bin)
372     return static_cast<ObjectFile *>(nullptr);
373 
374   if (MachOUniversalBinary *UB = dyn_cast_or_null<MachOUniversalBinary>(Bin)) {
375     auto I = ObjectForUBPathAndArch.find(std::make_pair(Path, ArchName));
376     if (I != ObjectForUBPathAndArch.end())
377       return I->second.get();
378 
379     Expected<std::unique_ptr<ObjectFile>> ObjOrErr =
380         UB->getObjectForArch(ArchName);
381     if (!ObjOrErr) {
382       ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
383                                      std::unique_ptr<ObjectFile>());
384       return ObjOrErr.takeError();
385     }
386     ObjectFile *Res = ObjOrErr->get();
387     ObjectForUBPathAndArch.emplace(std::make_pair(Path, ArchName),
388                                    std::move(ObjOrErr.get()));
389     return Res;
390   }
391   if (Bin->isObject()) {
392     return cast<ObjectFile>(Bin);
393   }
394   return errorCodeToError(object_error::arch_not_found);
395 }
396 
397 Expected<SymbolizableModule *>
398 LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
399   auto I = Modules.find(ModuleName);
400   if (I != Modules.end())
401     return I->second.get();
402 
403   std::string BinaryName = ModuleName;
404   std::string ArchName = Opts.DefaultArch;
405   size_t ColonPos = ModuleName.find_last_of(':');
406   // Verify that substring after colon form a valid arch name.
407   if (ColonPos != std::string::npos) {
408     std::string ArchStr = ModuleName.substr(ColonPos + 1);
409     if (Triple(ArchStr).getArch() != Triple::UnknownArch) {
410       BinaryName = ModuleName.substr(0, ColonPos);
411       ArchName = ArchStr;
412     }
413   }
414   auto ObjectsOrErr = getOrCreateObjectPair(BinaryName, ArchName);
415   if (!ObjectsOrErr) {
416     // Failed to find valid object file.
417     Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
418     return ObjectsOrErr.takeError();
419   }
420   ObjectPair Objects = ObjectsOrErr.get();
421 
422   std::unique_ptr<DIContext> Context;
423   // If this is a COFF object containing PDB info, use a PDBContext to
424   // symbolize. Otherwise, use DWARF.
425   if (auto CoffObject = dyn_cast<COFFObjectFile>(Objects.first)) {
426     const codeview::DebugInfo *DebugInfo;
427     StringRef PDBFileName;
428     auto EC = CoffObject->getDebugPDBInfo(DebugInfo, PDBFileName);
429     if (!EC && DebugInfo != nullptr && !PDBFileName.empty()) {
430       using namespace pdb;
431       std::unique_ptr<IPDBSession> Session;
432       if (auto Err = loadDataForEXE(PDB_ReaderType::DIA,
433                                     Objects.first->getFileName(), Session)) {
434         Modules.emplace(ModuleName, std::unique_ptr<SymbolizableModule>());
435         // Return along the PDB filename to provide more context
436         return createFileError(PDBFileName, std::move(Err));
437       }
438       Context.reset(new PDBContext(*CoffObject, std::move(Session)));
439     }
440   }
441   if (!Context)
442     Context =
443         DWARFContext::create(*Objects.second, nullptr,
444                              DWARFContext::defaultErrorHandler, Opts.DWPName);
445   auto InfoOrErr =
446       SymbolizableObjectFile::create(Objects.first, std::move(Context));
447   std::unique_ptr<SymbolizableModule> SymMod;
448   if (InfoOrErr)
449     SymMod = std::move(InfoOrErr.get());
450   auto InsertResult = Modules.emplace(ModuleName, std::move(SymMod));
451   assert(InsertResult.second);
452   if (auto EC = InfoOrErr.getError())
453     return errorCodeToError(EC);
454   return InsertResult.first->second.get();
455 }
456 
457 namespace {
458 
459 // Undo these various manglings for Win32 extern "C" functions:
460 // cdecl       - _foo
461 // stdcall     - _foo@12
462 // fastcall    - @foo@12
463 // vectorcall  - foo@@12
464 // These are all different linkage names for 'foo'.
465 StringRef demanglePE32ExternCFunc(StringRef SymbolName) {
466   // Remove any '_' or '@' prefix.
467   char Front = SymbolName.empty() ? '\0' : SymbolName[0];
468   if (Front == '_' || Front == '@')
469     SymbolName = SymbolName.drop_front();
470 
471   // Remove any '@[0-9]+' suffix.
472   if (Front != '?') {
473     size_t AtPos = SymbolName.rfind('@');
474     if (AtPos != StringRef::npos &&
475         std::all_of(SymbolName.begin() + AtPos + 1, SymbolName.end(),
476                     [](char C) { return C >= '0' && C <= '9'; })) {
477       SymbolName = SymbolName.substr(0, AtPos);
478     }
479   }
480 
481   // Remove any ending '@' for vectorcall.
482   if (SymbolName.endswith("@"))
483     SymbolName = SymbolName.drop_back();
484 
485   return SymbolName;
486 }
487 
488 } // end anonymous namespace
489 
490 std::string
491 LLVMSymbolizer::DemangleName(const std::string &Name,
492                              const SymbolizableModule *DbiModuleDescriptor) {
493   // We can spoil names of symbols with C linkage, so use an heuristic
494   // approach to check if the name should be demangled.
495   if (Name.substr(0, 2) == "_Z") {
496     int status = 0;
497     char *DemangledName = itaniumDemangle(Name.c_str(), nullptr, nullptr, &status);
498     if (status != 0)
499       return Name;
500     std::string Result = DemangledName;
501     free(DemangledName);
502     return Result;
503   }
504 
505 #if defined(_MSC_VER)
506   if (!Name.empty() && Name.front() == '?') {
507     // Only do MSVC C++ demangling on symbols starting with '?'.
508     char DemangledName[1024] = {0};
509     DWORD result = ::UnDecorateSymbolName(
510         Name.c_str(), DemangledName, 1023,
511         UNDNAME_NO_ACCESS_SPECIFIERS |       // Strip public, private, protected
512             UNDNAME_NO_ALLOCATION_LANGUAGE | // Strip __thiscall, __stdcall, etc
513             UNDNAME_NO_THROW_SIGNATURES |    // Strip throw() specifications
514             UNDNAME_NO_MEMBER_TYPE | // Strip virtual, static, etc specifiers
515             UNDNAME_NO_MS_KEYWORDS | // Strip all MS extension keywords
516             UNDNAME_NO_FUNCTION_RETURNS); // Strip function return types
517     return (result == 0) ? Name : std::string(DemangledName);
518   }
519 #endif
520   if (DbiModuleDescriptor && DbiModuleDescriptor->isWin32Module())
521     return std::string(demanglePE32ExternCFunc(Name));
522   return Name;
523 }
524 
525 } // namespace symbolize
526 } // namespace llvm
527