Symbolize.cpp revision 321369 
1//===-- LLVMSymbolize.cpp -------------------------------------------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// Implementation for LLVM symbolization library.
11//
12//===----------------------------------------------------------------------===//
13
14#include "llvm/DebugInfo/Symbolize/Symbolize.h"
15
16#include "SymbolizableObjectFile.h"
17
18#include "llvm/ADT/STLExtras.h"
19#include "llvm/BinaryFormat/COFF.h"
20#include "llvm/Config/config.h"
21#include "llvm/DebugInfo/DWARF/DWARFContext.h"
22#include "llvm/DebugInfo/PDB/PDB.h"
23#include "llvm/DebugInfo/PDB/PDBContext.h"
24#include "llvm/Object/COFF.h"
25#include "llvm/Object/ELFObjectFile.h"
26#include "llvm/Object/MachO.h"
27#include "llvm/Object/MachOUniversal.h"
28#include "llvm/Support/Casting.h"
29#include "llvm/Support/Compression.h"
30#include "llvm/Support/DataExtractor.h"
31#include "llvm/Support/Errc.h"
32#include "llvm/Support/FileSystem.h"
33#include "llvm/Support/MemoryBuffer.h"
34#include "llvm/Support/Path.h"
35#include <algorithm>
36#include <cassert>
37#include <cstdlib>
38#include <cstring>
39
40#if defined(_MSC_VER)
41#include <Windows.h>
42
43// This must be included after windows.h.
44#include <DbgHelp.h>
45#pragma comment(lib, "dbghelp.lib")
46
47// Windows.h conflicts with our COFF header definitions.
48#ifdef IMAGE_FILE_MACHINE_I386
49#undef IMAGE_FILE_MACHINE_I386
50#endif
51#endif
52
53namespace llvm {
54namespace symbolize {
55
56Expected<DILineInfo> LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
57                                                  uint64_t ModuleOffset) {
58  SymbolizableModule *Info;
59  if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
60    Info = InfoOrErr.get();
61  else
62    return InfoOrErr.takeError();
63
64  // A null module means an error has already been reported. Return an empty
65  // result.
66  if (!Info)
67    return DILineInfo();
68
69  // If the user is giving us relative addresses, add the preferred base of the
70  // object to the offset before we do the query. It's what DIContext expects.
71  if (Opts.RelativeAddresses)
72    ModuleOffset += Info->getModulePreferredBase();
73
74  DILineInfo LineInfo = Info->symbolizeCode(ModuleOffset, Opts.PrintFunctions,
75                                            Opts.UseSymbolTable);
76  if (Opts.Demangle)
77    LineInfo.FunctionName = DemangleName(LineInfo.FunctionName, Info);
78  return LineInfo;
79}
80
81Expected<DIInliningInfo>
82LLVMSymbolizer::symbolizeInlinedCode(const std::string &ModuleName,
83                                     uint64_t ModuleOffset) {
84  SymbolizableModule *Info;
85  if (auto InfoOrErr = getOrCreateModuleInfo(ModuleName))
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 += 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
111Expected<DIGlobal> LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
112                                                 uint64_t 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 += 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
136void LLVMSymbolizer::flush() {
137  ObjectForUBPathAndArch.clear();
138  BinaryForPath.clear();
139  ObjectPairForPathArch.clear();
140  Modules.clear();
141}
142
143namespace {
144
145// For Path="/path/to/foo" and Basename="foo" assume that debug info is in
146// /path/to/foo.dSYM/Contents/Resources/DWARF/foo.
147// For Path="/path/to/bar.dSYM" and Basename="foo" assume that debug info is in
148// /path/to/bar.dSYM/Contents/Resources/DWARF/foo.
149std::string getDarwinDWARFResourceForPath(
150    const std::string &Path, const std::string &Basename) {
151  SmallString<16> ResourceName = StringRef(Path);
152  if (sys::path::extension(Path) != ".dSYM") {
153    ResourceName += ".dSYM";
154  }
155  sys::path::append(ResourceName, "Contents", "Resources", "DWARF");
156  sys::path::append(ResourceName, Basename);
157  return ResourceName.str();
158}
159
160bool checkFileCRC(StringRef Path, uint32_t CRCHash) {
161  ErrorOr<std::unique_ptr<MemoryBuffer>> MB =
162      MemoryBuffer::getFileOrSTDIN(Path);
163  if (!MB)
164    return false;
165  return !zlib::isAvailable() || CRCHash == zlib::crc32(MB.get()->getBuffer());
166}
167
168bool findDebugBinary(const std::string &OrigPath,
169                     const std::string &DebuglinkName, uint32_t CRCHash,
170                     std::string &Result) {
171  std::string OrigRealPath = OrigPath;
172#if defined(HAVE_REALPATH)
173  if (char *RP = realpath(OrigPath.c_str(), nullptr)) {
174    OrigRealPath = RP;
175    free(RP);
176  }
177#endif
178  SmallString<16> OrigDir(OrigRealPath);
179  llvm::sys::path::remove_filename(OrigDir);
180  SmallString<16> DebugPath = OrigDir;
181  // Try /path/to/original_binary/debuglink_name
182  llvm::sys::path::append(DebugPath, DebuglinkName);
183  if (checkFileCRC(DebugPath, CRCHash)) {
184    Result = DebugPath.str();
185    return true;
186  }
187  // Try /path/to/original_binary/.debug/debuglink_name
188  DebugPath = OrigRealPath;
189  llvm::sys::path::append(DebugPath, ".debug", DebuglinkName);
190  if (checkFileCRC(DebugPath, CRCHash)) {
191    Result = DebugPath.str();
192    return true;
193  }
194  // Try /usr/lib/debug/path/to/original_binary/debuglink_name
195  DebugPath = "/usr/lib/debug";
196  llvm::sys::path::append(DebugPath, llvm::sys::path::relative_path(OrigDir),
197                          DebuglinkName);
198  if (checkFileCRC(DebugPath, CRCHash)) {
199    Result = DebugPath.str();
200    return true;
201  }
202  return false;
203}
204
205bool getGNUDebuglinkContents(const ObjectFile *Obj, std::string &DebugName,
206                             uint32_t &CRCHash) {
207  if (!Obj)
208    return false;
209  for (const SectionRef &Section : Obj->sections()) {
210    StringRef Name;
211    Section.getName(Name);
212    Name = Name.substr(Name.find_first_not_of("._"));
213    if (Name == "gnu_debuglink") {
214      StringRef Data;
215      Section.getContents(Data);
216      DataExtractor DE(Data, Obj->isLittleEndian(), 0);
217      uint32_t Offset = 0;
218      if (const char *DebugNameStr = DE.getCStr(&Offset)) {
219        // 4-byte align the offset.
220        Offset = (Offset + 3) & ~0x3;
221        if (DE.isValidOffsetForDataOfSize(Offset, 4)) {
222          DebugName = DebugNameStr;
223          CRCHash = DE.getU32(&Offset);
224          return true;
225        }
226      }
227      break;
228    }
229  }
230  return false;
231}
232
233bool darwinDsymMatchesBinary(const MachOObjectFile *DbgObj,
234                             const MachOObjectFile *Obj) {
235  ArrayRef<uint8_t> dbg_uuid = DbgObj->getUuid();
236  ArrayRef<uint8_t> bin_uuid = Obj->getUuid();
237  if (dbg_uuid.empty() || bin_uuid.empty())
238    return false;
239  return !memcmp(dbg_uuid.data(), bin_uuid.data(), dbg_uuid.size());
240}
241
242} // end anonymous namespace
243
244ObjectFile *LLVMSymbolizer::lookUpDsymFile(const std::string &ExePath,
245    const MachOObjectFile *MachExeObj, const std::string &ArchName) {
246  // On Darwin we may find DWARF in separate object file in
247  // resource directory.
248  std::vector<std::string> DsymPaths;
249  StringRef Filename = sys::path::filename(ExePath);
250  DsymPaths.push_back(getDarwinDWARFResourceForPath(ExePath, Filename));
251  for (const auto &Path : Opts.DsymHints) {
252    DsymPaths.push_back(getDarwinDWARFResourceForPath(Path, Filename));
253  }
254  for (const auto &Path : DsymPaths) {
255    auto DbgObjOrErr = getOrCreateObject(Path, ArchName);
256    if (!DbgObjOrErr) {
257      // Ignore errors, the file might not exist.
258      consumeError(DbgObjOrErr.takeError());
259      continue;
260    }
261    ObjectFile *DbgObj = DbgObjOrErr.get();
262    if (!DbgObj)
263      continue;
264    const MachOObjectFile *MachDbgObj = dyn_cast<const MachOObjectFile>(DbgObj);
265    if (!MachDbgObj)
266      continue;
267    if (darwinDsymMatchesBinary(MachDbgObj, MachExeObj))
268      return DbgObj;
269  }
270  return nullptr;
271}
272
273ObjectFile *LLVMSymbolizer::lookUpDebuglinkObject(const std::string &Path,
274                                                  const ObjectFile *Obj,
275                                                  const std::string &ArchName) {
276  std::string DebuglinkName;
277  uint32_t CRCHash;
278  std::string DebugBinaryPath;
279  if (!getGNUDebuglinkContents(Obj, DebuglinkName, CRCHash))
280    return nullptr;
281  if (!findDebugBinary(Path, DebuglinkName, CRCHash, DebugBinaryPath))
282    return nullptr;
283  auto DbgObjOrErr = getOrCreateObject(DebugBinaryPath, ArchName);
284  if (!DbgObjOrErr) {
285    // Ignore errors, the file might not exist.
286    consumeError(DbgObjOrErr.takeError());
287    return nullptr;
288  }
289  return DbgObjOrErr.get();
290}
291
292Expected<LLVMSymbolizer::ObjectPair>
293LLVMSymbolizer::getOrCreateObjectPair(const std::string &Path,
294                                      const std::string &ArchName) {
295  const auto &I = ObjectPairForPathArch.find(std::make_pair(Path, ArchName));
296  if (I != ObjectPairForPathArch.end()) {
297    return I->second;
298  }
299
300  auto ObjOrErr = getOrCreateObject(Path, ArchName);
301  if (!ObjOrErr) {
302    ObjectPairForPathArch.insert(std::make_pair(std::make_pair(Path, ArchName),
303                                                ObjectPair(nullptr, nullptr)));
304    return ObjOrErr.takeError();
305  }
306
307  ObjectFile *Obj = ObjOrErr.get();
308  assert(Obj != nullptr);
309  ObjectFile *DbgObj = nullptr;
310
311  if (auto MachObj = dyn_cast<const MachOObjectFile>(Obj))
312    DbgObj = lookUpDsymFile(Path, MachObj, ArchName);
313  if (!DbgObj)
314    DbgObj = lookUpDebuglinkObject(Path, Obj, ArchName);
315  if (!DbgObj)
316    DbgObj = Obj;
317  ObjectPair Res = std::make_pair(Obj, DbgObj);
318  ObjectPairForPathArch.insert(
319      std::make_pair(std::make_pair(Path, ArchName), Res));
320  return Res;
321}
322
323Expected<ObjectFile *>
324LLVMSymbolizer::getOrCreateObject(const std::string &Path,
325                                  const std::string &ArchName) {
326  const auto &I = BinaryForPath.find(Path);
327  Binary *Bin = nullptr;
328  if (I == BinaryForPath.end()) {
329    Expected<OwningBinary<Binary>> BinOrErr = createBinary(Path);
330    if (!BinOrErr) {
331      BinaryForPath.insert(std::make_pair(Path, OwningBinary<Binary>()));
332      return BinOrErr.takeError();
333    }
334    Bin = BinOrErr->getBinary();
335    BinaryForPath.insert(std::make_pair(Path, std::move(BinOrErr.get())));
336  } else {
337    Bin = I->second.getBinary();
338  }
339
340  if (!Bin)
341    return static_cast<ObjectFile *>(nullptr);
342
343  if (MachOUniversalBinary *UB = dyn_cast_or_null<MachOUniversalBinary>(Bin)) {
344    const auto &I = ObjectForUBPathAndArch.find(std::make_pair(Path, ArchName));
345    if (I != ObjectForUBPathAndArch.end()) {
346      return I->second.get();
347    }
348    Expected<std::unique_ptr<ObjectFile>> ObjOrErr =
349        UB->getObjectForArch(ArchName);
350    if (!ObjOrErr) {
351      ObjectForUBPathAndArch.insert(std::make_pair(
352          std::make_pair(Path, ArchName), std::unique_ptr<ObjectFile>()));
353      return ObjOrErr.takeError();
354    }
355    ObjectFile *Res = ObjOrErr->get();
356    ObjectForUBPathAndArch.insert(std::make_pair(std::make_pair(Path, ArchName),
357                                                 std::move(ObjOrErr.get())));
358    return Res;
359  }
360  if (Bin->isObject()) {
361    return cast<ObjectFile>(Bin);
362  }
363  return errorCodeToError(object_error::arch_not_found);
364}
365
366Expected<SymbolizableModule *>
367LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
368  const auto &I = Modules.find(ModuleName);
369  if (I != Modules.end()) {
370    return I->second.get();
371  }
372  std::string BinaryName = ModuleName;
373  std::string ArchName = Opts.DefaultArch;
374  size_t ColonPos = ModuleName.find_last_of(':');
375  // Verify that substring after colon form a valid arch name.
376  if (ColonPos != std::string::npos) {
377    std::string ArchStr = ModuleName.substr(ColonPos + 1);
378    if (Triple(ArchStr).getArch() != Triple::UnknownArch) {
379      BinaryName = ModuleName.substr(0, ColonPos);
380      ArchName = ArchStr;
381    }
382  }
383  auto ObjectsOrErr = getOrCreateObjectPair(BinaryName, ArchName);
384  if (!ObjectsOrErr) {
385    // Failed to find valid object file.
386    Modules.insert(
387        std::make_pair(ModuleName, std::unique_ptr<SymbolizableModule>()));
388    return ObjectsOrErr.takeError();
389  }
390  ObjectPair Objects = ObjectsOrErr.get();
391
392  std::unique_ptr<DIContext> Context;
393  // If this is a COFF object containing PDB info, use a PDBContext to
394  // symbolize. Otherwise, use DWARF.
395  if (auto CoffObject = dyn_cast<COFFObjectFile>(Objects.first)) {
396    const codeview::DebugInfo *DebugInfo;
397    StringRef PDBFileName;
398    auto EC = CoffObject->getDebugPDBInfo(DebugInfo, PDBFileName);
399    if (!EC && DebugInfo != nullptr && !PDBFileName.empty()) {
400#if 0
401      using namespace pdb;
402      std::unique_ptr<IPDBSession> Session;
403      if (auto Err = loadDataForEXE(PDB_ReaderType::DIA,
404                                    Objects.first->getFileName(), Session)) {
405        Modules.insert(
406            std::make_pair(ModuleName, std::unique_ptr<SymbolizableModule>()));
407        return std::move(Err);
408      }
409      Context.reset(new PDBContext(*CoffObject, std::move(Session)));
410#else
411      return make_error<StringError>(
412          "PDB support not compiled in",
413          std::make_error_code(std::errc::not_supported));
414#endif
415    }
416  }
417  if (!Context)
418    Context.reset(new DWARFContextInMemory(*Objects.second));
419  assert(Context);
420  auto InfoOrErr =
421      SymbolizableObjectFile::create(Objects.first, std::move(Context));
422  std::unique_ptr<SymbolizableModule> SymMod;
423  if (InfoOrErr)
424    SymMod = std::move(InfoOrErr.get());
425  auto InsertResult =
426      Modules.insert(std::make_pair(ModuleName, std::move(SymMod)));
427  assert(InsertResult.second);
428  if (auto EC = InfoOrErr.getError())
429    return errorCodeToError(EC);
430  return InsertResult.first->second.get();
431}
432
433namespace {
434
435// Undo these various manglings for Win32 extern "C" functions:
436// cdecl       - _foo
437// stdcall     - _foo@12
438// fastcall    - @foo@12
439// vectorcall  - foo@@12
440// These are all different linkage names for 'foo'.
441StringRef demanglePE32ExternCFunc(StringRef SymbolName) {
442  // Remove any '_' or '@' prefix.
443  char Front = SymbolName.empty() ? '\0' : SymbolName[0];
444  if (Front == '_' || Front == '@')
445    SymbolName = SymbolName.drop_front();
446
447  // Remove any '@[0-9]+' suffix.
448  if (Front != '?') {
449    size_t AtPos = SymbolName.rfind('@');
450    if (AtPos != StringRef::npos &&
451        std::all_of(SymbolName.begin() + AtPos + 1, SymbolName.end(),
452                    [](char C) { return C >= '0' && C <= '9'; })) {
453      SymbolName = SymbolName.substr(0, AtPos);
454    }
455  }
456
457  // Remove any ending '@' for vectorcall.
458  if (SymbolName.endswith("@"))
459    SymbolName = SymbolName.drop_back();
460
461  return SymbolName;
462}
463
464} // end anonymous namespace
465
466#if !defined(_MSC_VER)
467// Assume that __cxa_demangle is provided by libcxxabi (except for Windows).
468extern "C" char *__cxa_demangle(const char *mangled_name, char *output_buffer,
469                                size_t *length, int *status);
470#endif
471
472std::string
473LLVMSymbolizer::DemangleName(const std::string &Name,
474                             const SymbolizableModule *DbiModuleDescriptor) {
475#if !defined(_MSC_VER)
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 = __cxa_demangle(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#else
488  if (!Name.empty() && Name.front() == '?') {
489    // Only do MSVC C++ demangling on symbols starting with '?'.
490    char DemangledName[1024] = {0};
491    DWORD result = ::UnDecorateSymbolName(
492        Name.c_str(), DemangledName, 1023,
493        UNDNAME_NO_ACCESS_SPECIFIERS |       // Strip public, private, protected
494            UNDNAME_NO_ALLOCATION_LANGUAGE | // Strip __thiscall, __stdcall, etc
495            UNDNAME_NO_THROW_SIGNATURES |    // Strip throw() specifications
496            UNDNAME_NO_MEMBER_TYPE | // Strip virtual, static, etc specifiers
497            UNDNAME_NO_MS_KEYWORDS | // Strip all MS extension keywords
498            UNDNAME_NO_FUNCTION_RETURNS); // Strip function return types
499    return (result == 0) ? Name : std::string(DemangledName);
500  }
501#endif
502  if (DbiModuleDescriptor && DbiModuleDescriptor->isWin32Module())
503    return std::string(demanglePE32ExternCFunc(Name));
504  return Name;
505}
506
507} // namespace symbolize
508} // namespace llvm
509