1 //===-- LLVMSymbolize.cpp -------------------------------------------------===//
3 // The LLVM Compiler Infrastructure
5 // This file is distributed under the University of Illinois Open Source
6 // License. See LICENSE.TXT for details.
8 //===----------------------------------------------------------------------===//
10 // Implementation for LLVM symbolization library.
12 //===----------------------------------------------------------------------===//
14 #include "LLVMSymbolize.h"
15 #include "llvm/ADT/STLExtras.h"
16 #include "llvm/Object/MachO.h"
17 #include "llvm/Support/Casting.h"
18 #include "llvm/Support/FileSystem.h"
19 #include "llvm/Support/Path.h"
26 static bool error(error_code ec) {
29 errs() << "LLVMSymbolizer: error reading file: " << ec.message() << ".\n";
34 getDILineInfoSpecifierFlags(const LLVMSymbolizer::Options &Opts) {
35 uint32_t Flags = llvm::DILineInfoSpecifier::FileLineInfo |
36 llvm::DILineInfoSpecifier::AbsoluteFilePath;
37 if (Opts.PrintFunctions)
38 Flags |= llvm::DILineInfoSpecifier::FunctionName;
42 static void patchFunctionNameInDILineInfo(const std::string &NewFunctionName,
43 DILineInfo &LineInfo) {
44 std::string FileName = LineInfo.getFileName();
45 LineInfo = DILineInfo(StringRef(FileName), StringRef(NewFunctionName),
46 LineInfo.getLine(), LineInfo.getColumn());
49 ModuleInfo::ModuleInfo(ObjectFile *Obj, DIContext *DICtx)
50 : Module(Obj), DebugInfoContext(DICtx) {
52 for (symbol_iterator si = Module->begin_symbols(), se = Module->end_symbols();
53 si != se; si.increment(ec)) {
56 SymbolRef::Type SymbolType;
57 if (error(si->getType(SymbolType)))
59 if (SymbolType != SymbolRef::ST_Function &&
60 SymbolType != SymbolRef::ST_Data)
62 uint64_t SymbolAddress;
63 if (error(si->getAddress(SymbolAddress)) ||
64 SymbolAddress == UnknownAddressOrSize)
67 // Getting symbol size is linear for Mach-O files, so assume that symbol
68 // occupies the memory range up to the following symbol.
69 if (isa<MachOObjectFile>(Obj))
71 else if (error(si->getSize(SymbolSize)) ||
72 SymbolSize == UnknownAddressOrSize)
75 if (error(si->getName(SymbolName)))
77 // Mach-O symbol table names have leading underscore, skip it.
78 if (Module->isMachO() && SymbolName.size() > 0 && SymbolName[0] == '_')
79 SymbolName = SymbolName.drop_front();
80 // FIXME: If a function has alias, there are two entries in symbol table
81 // with same address size. Make sure we choose the correct one.
82 SymbolMapTy &M = SymbolType == SymbolRef::ST_Function ? Functions : Objects;
83 SymbolDesc SD = { SymbolAddress, SymbolSize };
84 M.insert(std::make_pair(SD, SymbolName));
88 bool ModuleInfo::getNameFromSymbolTable(SymbolRef::Type Type, uint64_t Address,
89 std::string &Name, uint64_t &Addr,
90 uint64_t &Size) const {
91 const SymbolMapTy &M = Type == SymbolRef::ST_Function ? Functions : Objects;
94 SymbolDesc SD = { Address, Address };
95 SymbolMapTy::const_iterator it = M.upper_bound(SD);
99 if (it->first.Size != 0 && it->first.Addr + it->first.Size <= Address)
101 Name = it->second.str();
102 Addr = it->first.Addr;
103 Size = it->first.Size;
107 DILineInfo ModuleInfo::symbolizeCode(
108 uint64_t ModuleOffset, const LLVMSymbolizer::Options &Opts) const {
110 if (DebugInfoContext) {
111 LineInfo = DebugInfoContext->getLineInfoForAddress(
112 ModuleOffset, getDILineInfoSpecifierFlags(Opts));
114 // Override function name from symbol table if necessary.
115 if (Opts.PrintFunctions && Opts.UseSymbolTable) {
116 std::string FunctionName;
117 uint64_t Start, Size;
118 if (getNameFromSymbolTable(SymbolRef::ST_Function, ModuleOffset,
119 FunctionName, Start, Size)) {
120 patchFunctionNameInDILineInfo(FunctionName, LineInfo);
126 DIInliningInfo ModuleInfo::symbolizeInlinedCode(
127 uint64_t ModuleOffset, const LLVMSymbolizer::Options &Opts) const {
128 DIInliningInfo InlinedContext;
129 if (DebugInfoContext) {
130 InlinedContext = DebugInfoContext->getInliningInfoForAddress(
131 ModuleOffset, getDILineInfoSpecifierFlags(Opts));
133 // Make sure there is at least one frame in context.
134 if (InlinedContext.getNumberOfFrames() == 0) {
135 InlinedContext.addFrame(DILineInfo());
137 // Override the function name in lower frame with name from symbol table.
138 if (Opts.PrintFunctions && Opts.UseSymbolTable) {
139 DIInliningInfo PatchedInlinedContext;
140 for (uint32_t i = 0, n = InlinedContext.getNumberOfFrames(); i < n; i++) {
141 DILineInfo LineInfo = InlinedContext.getFrame(i);
143 std::string FunctionName;
144 uint64_t Start, Size;
145 if (getNameFromSymbolTable(SymbolRef::ST_Function, ModuleOffset,
146 FunctionName, Start, Size)) {
147 patchFunctionNameInDILineInfo(FunctionName, LineInfo);
150 PatchedInlinedContext.addFrame(LineInfo);
152 InlinedContext = PatchedInlinedContext;
154 return InlinedContext;
157 bool ModuleInfo::symbolizeData(uint64_t ModuleOffset, std::string &Name,
158 uint64_t &Start, uint64_t &Size) const {
159 return getNameFromSymbolTable(SymbolRef::ST_Data, ModuleOffset, Name, Start,
163 const char LLVMSymbolizer::kBadString[] = "??";
165 std::string LLVMSymbolizer::symbolizeCode(const std::string &ModuleName,
166 uint64_t ModuleOffset) {
167 ModuleInfo *Info = getOrCreateModuleInfo(ModuleName);
169 return printDILineInfo(DILineInfo());
170 if (Opts.PrintInlining) {
171 DIInliningInfo InlinedContext =
172 Info->symbolizeInlinedCode(ModuleOffset, Opts);
173 uint32_t FramesNum = InlinedContext.getNumberOfFrames();
174 assert(FramesNum > 0);
176 for (uint32_t i = 0; i < FramesNum; i++) {
177 DILineInfo LineInfo = InlinedContext.getFrame(i);
178 Result += printDILineInfo(LineInfo);
182 DILineInfo LineInfo = Info->symbolizeCode(ModuleOffset, Opts);
183 return printDILineInfo(LineInfo);
186 std::string LLVMSymbolizer::symbolizeData(const std::string &ModuleName,
187 uint64_t ModuleOffset) {
188 std::string Name = kBadString;
191 if (Opts.UseSymbolTable) {
192 if (ModuleInfo *Info = getOrCreateModuleInfo(ModuleName)) {
193 if (Info->symbolizeData(ModuleOffset, Name, Start, Size) && Opts.Demangle)
194 Name = DemangleName(Name);
197 std::stringstream ss;
198 ss << Name << "\n" << Start << " " << Size << "\n";
202 void LLVMSymbolizer::flush() {
203 DeleteContainerSeconds(Modules);
204 DeleteContainerPointers(ParsedBinariesAndObjects);
205 BinaryForPath.clear();
206 ObjectFileForArch.clear();
209 static std::string getDarwinDWARFResourceForPath(const std::string &Path) {
210 StringRef Basename = sys::path::filename(Path);
211 const std::string &DSymDirectory = Path + ".dSYM";
212 SmallString<16> ResourceName = StringRef(DSymDirectory);
213 sys::path::append(ResourceName, "Contents", "Resources", "DWARF");
214 sys::path::append(ResourceName, Basename);
215 return ResourceName.str();
218 LLVMSymbolizer::BinaryPair
219 LLVMSymbolizer::getOrCreateBinary(const std::string &Path) {
220 BinaryMapTy::iterator I = BinaryForPath.find(Path);
221 if (I != BinaryForPath.end())
225 OwningPtr<Binary> ParsedBinary;
226 OwningPtr<Binary> ParsedDbgBinary;
227 if (!error(createBinary(Path, ParsedBinary))) {
228 // Check if it's a universal binary.
229 Bin = ParsedBinary.take();
230 ParsedBinariesAndObjects.push_back(Bin);
231 if (Bin->isMachO() || Bin->isMachOUniversalBinary()) {
232 // On Darwin we may find DWARF in separate object file in
233 // resource directory.
234 const std::string &ResourcePath =
235 getDarwinDWARFResourceForPath(Path);
236 bool ResourceFileExists = false;
237 if (!sys::fs::exists(ResourcePath, ResourceFileExists) &&
238 ResourceFileExists &&
239 !error(createBinary(ResourcePath, ParsedDbgBinary))) {
240 DbgBin = ParsedDbgBinary.take();
241 ParsedBinariesAndObjects.push_back(DbgBin);
247 BinaryPair Res = std::make_pair(Bin, DbgBin);
248 BinaryForPath[Path] = Res;
253 LLVMSymbolizer::getObjectFileFromBinary(Binary *Bin, const std::string &ArchName) {
257 if (MachOUniversalBinary *UB = dyn_cast<MachOUniversalBinary>(Bin)) {
258 ObjectFileForArchMapTy::iterator I = ObjectFileForArch.find(
259 std::make_pair(UB, ArchName));
260 if (I != ObjectFileForArch.end())
262 OwningPtr<ObjectFile> ParsedObj;
263 if (!UB->getObjectForArch(Triple(ArchName).getArch(), ParsedObj)) {
264 Res = ParsedObj.take();
265 ParsedBinariesAndObjects.push_back(Res);
267 ObjectFileForArch[std::make_pair(UB, ArchName)] = Res;
268 } else if (Bin->isObject()) {
269 Res = cast<ObjectFile>(Bin);
275 LLVMSymbolizer::getOrCreateModuleInfo(const std::string &ModuleName) {
276 ModuleMapTy::iterator I = Modules.find(ModuleName);
277 if (I != Modules.end())
279 std::string BinaryName = ModuleName;
280 std::string ArchName = Opts.DefaultArch;
281 size_t ColonPos = ModuleName.find(':');
283 // Recognize a drive letter on win32.
284 if (ColonPos == 1 && isalpha(ModuleName[0]))
285 ColonPos = ModuleName.find(':', 2);
287 if (ColonPos != std::string::npos) {
288 BinaryName = ModuleName.substr(0, ColonPos);
289 ArchName = ModuleName.substr(ColonPos + 1);
291 BinaryPair Binaries = getOrCreateBinary(BinaryName);
292 ObjectFile *Obj = getObjectFileFromBinary(Binaries.first, ArchName);
293 ObjectFile *DbgObj = getObjectFileFromBinary(Binaries.second, ArchName);
296 // Failed to find valid object file.
297 Modules.insert(make_pair(ModuleName, (ModuleInfo *)0));
300 DIContext *Context = DIContext::getDWARFContext(DbgObj);
302 ModuleInfo *Info = new ModuleInfo(Obj, Context);
303 Modules.insert(make_pair(ModuleName, Info));
307 std::string LLVMSymbolizer::printDILineInfo(DILineInfo LineInfo) const {
308 // By default, DILineInfo contains "<invalid>" for function/filename it
309 // cannot fetch. We replace it to "??" to make our output closer to addr2line.
310 static const std::string kDILineInfoBadString = "<invalid>";
311 std::stringstream Result;
312 if (Opts.PrintFunctions) {
313 std::string FunctionName = LineInfo.getFunctionName();
314 if (FunctionName == kDILineInfoBadString)
315 FunctionName = kBadString;
316 else if (Opts.Demangle)
317 FunctionName = DemangleName(FunctionName);
318 Result << FunctionName << "\n";
320 std::string Filename = LineInfo.getFileName();
321 if (Filename == kDILineInfoBadString)
322 Filename = kBadString;
323 Result << Filename << ":" << LineInfo.getLine() << ":" << LineInfo.getColumn()
328 #if !defined(_MSC_VER)
329 // Assume that __cxa_demangle is provided by libcxxabi (except for Windows).
330 extern "C" char *__cxa_demangle(const char *mangled_name, char *output_buffer,
331 size_t *length, int *status);
334 std::string LLVMSymbolizer::DemangleName(const std::string &Name) {
335 #if !defined(_MSC_VER)
337 char *DemangledName = __cxa_demangle(Name.c_str(), 0, 0, &status);
340 std::string Result = DemangledName;
348 } // namespace symbolize