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