| 1 | //===-- gsymutil.cpp - GSYM dumping and creation utility for llvm ---------===// |
| 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 | #include "llvm/ADT/STLExtras.h" |
| 10 | #include "llvm/DebugInfo/DIContext.h" |
| 11 | #include "llvm/DebugInfo/DWARF/DWARFContext.h" |
| 12 | #include "llvm/DebugInfo/GSYM/CallSiteInfo.h" |
| 13 | #include "llvm/Object/Archive.h" |
| 14 | #include "llvm/Object/ELFObjectFile.h" |
| 15 | #include "llvm/Object/MachOUniversal.h" |
| 16 | #include "llvm/Object/ObjectFile.h" |
| 17 | #include "llvm/Option/ArgList.h" |
| 18 | #include "llvm/Option/Option.h" |
| 19 | #include "llvm/Support/CommandLine.h" |
| 20 | #include "llvm/Support/Debug.h" |
| 21 | #include "llvm/Support/Format.h" |
| 22 | #include "llvm/Support/JSON.h" |
| 23 | #include "llvm/Support/LLVMDriver.h" |
| 24 | #include "llvm/Support/ManagedStatic.h" |
| 25 | #include "llvm/Support/MemoryBuffer.h" |
| 26 | #include "llvm/Support/PrettyStackTrace.h" |
| 27 | #include "llvm/Support/Regex.h" |
| 28 | #include "llvm/Support/Signals.h" |
| 29 | #include "llvm/Support/TargetSelect.h" |
| 30 | #include "llvm/Support/raw_ostream.h" |
| 31 | #include "llvm/TargetParser/Triple.h" |
| 32 | #include <algorithm> |
| 33 | #include <cstring> |
| 34 | #include <inttypes.h> |
| 35 | #include <iostream> |
| 36 | #include <optional> |
| 37 | #include <string> |
| 38 | #include <system_error> |
| 39 | #include <vector> |
| 40 | |
| 41 | #include "llvm/DebugInfo/GSYM/DwarfTransformer.h" |
| 42 | #include "llvm/DebugInfo/GSYM/FunctionInfo.h" |
| 43 | #include "llvm/DebugInfo/GSYM/GsymCreator.h" |
| 44 | #include "llvm/DebugInfo/GSYM/GsymReader.h" |
| 45 | #include "llvm/DebugInfo/GSYM/InlineInfo.h" |
| 46 | #include "llvm/DebugInfo/GSYM/LookupResult.h" |
| 47 | #include "llvm/DebugInfo/GSYM/ObjectFileTransformer.h" |
| 48 | #include "llvm/DebugInfo/GSYM/OutputAggregator.h" |
| 49 | |
| 50 | using namespace llvm; |
| 51 | using namespace gsym; |
| 52 | using namespace object; |
| 53 | |
| 54 | /// @} |
| 55 | /// Command line options. |
| 56 | /// @{ |
| 57 | |
| 58 | using namespace llvm::opt; |
| 59 | enum ID { |
| 60 | OPT_INVALID = 0, // This is not an option ID. |
| 61 | #define OPTION(...) LLVM_MAKE_OPT_ID(__VA_ARGS__), |
| 62 | #include "Opts.inc" |
| 63 | #undef OPTION |
| 64 | }; |
| 65 | |
| 66 | #define OPTTABLE_STR_TABLE_CODE |
| 67 | #include "Opts.inc" |
| 68 | #undef OPTTABLE_STR_TABLE_CODE |
| 69 | |
| 70 | #define OPTTABLE_PREFIXES_TABLE_CODE |
| 71 | #include "Opts.inc" |
| 72 | #undef OPTTABLE_PREFIXES_TABLE_CODE |
| 73 | |
| 74 | const opt::OptTable::Info InfoTable[] = { |
| 75 | #define OPTION(...) LLVM_CONSTRUCT_OPT_INFO(__VA_ARGS__), |
| 76 | #include "Opts.inc" |
| 77 | #undef OPTION |
| 78 | }; |
| 79 | |
| 80 | class GSYMUtilOptTable : public llvm::opt::GenericOptTable { |
| 81 | public: |
| 82 | GSYMUtilOptTable() |
| 83 | : GenericOptTable(OptionStrTable, OptionPrefixesTable, InfoTable) { |
| 84 | setGroupedShortOptions(true); |
| 85 | } |
| 86 | }; |
| 87 | |
| 88 | static bool Verbose; |
| 89 | static std::vector<std::string> InputFilenames; |
| 90 | static std::string ConvertFilename; |
| 91 | static std::vector<std::string> ArchFilters; |
| 92 | static std::string OutputFilename; |
| 93 | static std::string JsonSummaryFile; |
| 94 | static bool Verify; |
| 95 | static unsigned NumThreads; |
| 96 | static uint64_t SegmentSize; |
| 97 | static bool Quiet; |
| 98 | static std::vector<uint64_t> LookupAddresses; |
| 99 | static bool LookupAddressesFromStdin; |
| 100 | static bool UseMergedFunctions = false; |
| 101 | static bool LoadDwarfCallSites = false; |
| 102 | static std::string CallSiteYamlPath; |
| 103 | static std::vector<std::string> MergedFunctionsFilters; |
| 104 | |
| 105 | static void parseArgs(int argc, char **argv) { |
| 106 | GSYMUtilOptTable Tbl; |
| 107 | llvm::StringRef ToolName = argv[0]; |
| 108 | llvm::BumpPtrAllocator A; |
| 109 | llvm::StringSaver Saver{A}; |
| 110 | llvm::opt::InputArgList Args = |
| 111 | Tbl.parseArgs(Argc: argc, Argv: argv, Unknown: OPT_UNKNOWN, Saver, ErrorFn: [&](StringRef Msg) { |
| 112 | llvm::errs() << Msg << '\n'; |
| 113 | std::exit(status: 1); |
| 114 | }); |
| 115 | if (Args.hasArg(Ids: OPT_help)) { |
| 116 | const char *Overview = |
| 117 | "A tool for dumping, searching and creating GSYM files.\n\n" |
| 118 | "Specify one or more GSYM paths as arguments to dump all of the " |
| 119 | "information in each GSYM file.\n" |
| 120 | "Specify a single GSYM file along with one or more --lookup options to " |
| 121 | "lookup addresses within that GSYM file.\n" |
| 122 | "Use the --convert option to specify a file with option --out-file " |
| 123 | "option to convert to GSYM format.\n" ; |
| 124 | |
| 125 | Tbl.printHelp(OS&: llvm::outs(), Usage: "llvm-gsymutil [options] <input GSYM files>" , |
| 126 | Title: Overview); |
| 127 | std::exit(status: 0); |
| 128 | } |
| 129 | if (Args.hasArg(Ids: OPT_version)) { |
| 130 | llvm::outs() << ToolName << '\n'; |
| 131 | cl::PrintVersionMessage(); |
| 132 | std::exit(status: 0); |
| 133 | } |
| 134 | |
| 135 | Verbose = Args.hasArg(Ids: OPT_verbose); |
| 136 | |
| 137 | for (const llvm::opt::Arg *A : Args.filtered(Ids: OPT_INPUT)) |
| 138 | InputFilenames.emplace_back(args: A->getValue()); |
| 139 | |
| 140 | if (const llvm::opt::Arg *A = Args.getLastArg(Ids: OPT_convert_EQ)) |
| 141 | ConvertFilename = A->getValue(); |
| 142 | |
| 143 | for (const llvm::opt::Arg *A : Args.filtered(Ids: OPT_arch_EQ)) |
| 144 | ArchFilters.emplace_back(args: A->getValue()); |
| 145 | |
| 146 | if (const llvm::opt::Arg *A = Args.getLastArg(Ids: OPT_out_file_EQ)) |
| 147 | OutputFilename = A->getValue(); |
| 148 | |
| 149 | if (const llvm::opt::Arg *A = Args.getLastArg(Ids: OPT_json_summary_file_EQ)) |
| 150 | JsonSummaryFile = A->getValue(); |
| 151 | |
| 152 | Verify = Args.hasArg(Ids: OPT_verify); |
| 153 | |
| 154 | if (const llvm::opt::Arg *A = Args.getLastArg(Ids: OPT_num_threads_EQ)) { |
| 155 | StringRef S{A->getValue()}; |
| 156 | if (!llvm::to_integer(S, Num&: NumThreads, Base: 0)) { |
| 157 | llvm::errs() << ToolName << ": for the --num-threads option: '" << S |
| 158 | << "' value invalid for uint argument!\n" ; |
| 159 | std::exit(status: 1); |
| 160 | } |
| 161 | } |
| 162 | |
| 163 | if (const llvm::opt::Arg *A = Args.getLastArg(Ids: OPT_segment_size_EQ)) { |
| 164 | StringRef S{A->getValue()}; |
| 165 | if (!llvm::to_integer(S, Num&: SegmentSize, Base: 0)) { |
| 166 | llvm::errs() << ToolName << ": for the --segment-size option: '" << S |
| 167 | << "' value invalid for uint argument!\n" ; |
| 168 | std::exit(status: 1); |
| 169 | } |
| 170 | } |
| 171 | |
| 172 | Quiet = Args.hasArg(Ids: OPT_quiet); |
| 173 | |
| 174 | for (const llvm::opt::Arg *A : Args.filtered(Ids: OPT_address_EQ)) { |
| 175 | StringRef S{A->getValue()}; |
| 176 | if (!llvm::to_integer(S, Num&: LookupAddresses.emplace_back(), Base: 0)) { |
| 177 | llvm::errs() << ToolName << ": for the --address option: '" << S |
| 178 | << "' value invalid for uint argument!\n" ; |
| 179 | std::exit(status: 1); |
| 180 | } |
| 181 | } |
| 182 | |
| 183 | LookupAddressesFromStdin = Args.hasArg(Ids: OPT_addresses_from_stdin); |
| 184 | UseMergedFunctions = Args.hasArg(Ids: OPT_merged_functions); |
| 185 | |
| 186 | if (Args.hasArg(Ids: OPT_callsites_yaml_file_EQ)) { |
| 187 | CallSiteYamlPath = Args.getLastArgValue(Id: OPT_callsites_yaml_file_EQ); |
| 188 | if (CallSiteYamlPath.empty()) { |
| 189 | llvm::errs() |
| 190 | << ToolName |
| 191 | << ": --callsites-yaml-file option requires a non-empty argument.\n" ; |
| 192 | std::exit(status: 1); |
| 193 | } |
| 194 | } |
| 195 | |
| 196 | LoadDwarfCallSites = Args.hasArg(Ids: OPT_dwarf_callsites); |
| 197 | |
| 198 | for (const llvm::opt::Arg *A : |
| 199 | Args.filtered(Ids: OPT_merged_functions_filter_EQ)) { |
| 200 | MergedFunctionsFilters.push_back(x: A->getValue()); |
| 201 | // Validate the filter is only used with correct flags |
| 202 | if (LookupAddresses.empty() && !LookupAddressesFromStdin) { |
| 203 | llvm::errs() << ToolName |
| 204 | << ": --merged-functions-filter can only be used with " |
| 205 | "--address/--addresses-from-stdin\n" ; |
| 206 | std::exit(status: 1); |
| 207 | } |
| 208 | if (!UseMergedFunctions) { |
| 209 | llvm::errs() |
| 210 | << ToolName |
| 211 | << ": --merged-functions-filter requires --merged-functions\n" ; |
| 212 | std::exit(status: 1); |
| 213 | } |
| 214 | } |
| 215 | } |
| 216 | |
| 217 | /// @} |
| 218 | //===----------------------------------------------------------------------===// |
| 219 | |
| 220 | static void error(Error Err) { |
| 221 | if (!Err) |
| 222 | return; |
| 223 | WithColor::error() << toString(E: std::move(Err)) << "\n" ; |
| 224 | exit(status: 1); |
| 225 | } |
| 226 | |
| 227 | static void error(StringRef Prefix, llvm::Error Err) { |
| 228 | if (!Err) |
| 229 | return; |
| 230 | errs() << Prefix << ": " << Err << "\n" ; |
| 231 | consumeError(Err: std::move(Err)); |
| 232 | exit(status: 1); |
| 233 | } |
| 234 | |
| 235 | static void error(StringRef Prefix, std::error_code EC) { |
| 236 | if (!EC) |
| 237 | return; |
| 238 | errs() << Prefix << ": " << EC.message() << "\n" ; |
| 239 | exit(status: 1); |
| 240 | } |
| 241 | |
| 242 | static uint32_t getCPUType(MachOObjectFile &MachO) { |
| 243 | if (MachO.is64Bit()) |
| 244 | return MachO.getHeader64().cputype; |
| 245 | else |
| 246 | return MachO.getHeader().cputype; |
| 247 | } |
| 248 | |
| 249 | /// Return true if the object file has not been filtered by an --arch option. |
| 250 | static bool filterArch(MachOObjectFile &Obj) { |
| 251 | if (ArchFilters.empty()) |
| 252 | return true; |
| 253 | |
| 254 | Triple ObjTriple(Obj.getArchTriple()); |
| 255 | StringRef ObjArch = ObjTriple.getArchName(); |
| 256 | |
| 257 | for (StringRef Arch : ArchFilters) { |
| 258 | // Match name. |
| 259 | if (Arch == ObjArch) |
| 260 | return true; |
| 261 | |
| 262 | // Match architecture number. |
| 263 | unsigned Value; |
| 264 | if (!Arch.getAsInteger(Radix: 0, Result&: Value)) |
| 265 | if (Value == getCPUType(MachO&: Obj)) |
| 266 | return true; |
| 267 | } |
| 268 | return false; |
| 269 | } |
| 270 | |
| 271 | /// Determine the virtual address that is considered the base address of an ELF |
| 272 | /// object file. |
| 273 | /// |
| 274 | /// The base address of an ELF file is the "p_vaddr" of the first program |
| 275 | /// header whose "p_type" is PT_LOAD. |
| 276 | /// |
| 277 | /// \param ELFFile An ELF object file we will search. |
| 278 | /// |
| 279 | /// \returns A valid image base address if we are able to extract one. |
| 280 | template <class ELFT> |
| 281 | static std::optional<uint64_t> |
| 282 | getImageBaseAddress(const object::ELFFile<ELFT> &ELFFile) { |
| 283 | auto PhdrRangeOrErr = ELFFile.program_headers(); |
| 284 | if (!PhdrRangeOrErr) { |
| 285 | consumeError(PhdrRangeOrErr.takeError()); |
| 286 | return std::nullopt; |
| 287 | } |
| 288 | for (const typename ELFT::Phdr &Phdr : *PhdrRangeOrErr) |
| 289 | if (Phdr.p_type == ELF::PT_LOAD) |
| 290 | return (uint64_t)Phdr.p_vaddr; |
| 291 | return std::nullopt; |
| 292 | } |
| 293 | |
| 294 | /// Determine the virtual address that is considered the base address of mach-o |
| 295 | /// object file. |
| 296 | /// |
| 297 | /// The base address of a mach-o file is the vmaddr of the "__TEXT" segment. |
| 298 | /// |
| 299 | /// \param MachO A mach-o object file we will search. |
| 300 | /// |
| 301 | /// \returns A valid image base address if we are able to extract one. |
| 302 | static std::optional<uint64_t> |
| 303 | getImageBaseAddress(const object::MachOObjectFile *MachO) { |
| 304 | for (const auto &Command : MachO->load_commands()) { |
| 305 | if (Command.C.cmd == MachO::LC_SEGMENT) { |
| 306 | MachO::segment_command SLC = MachO->getSegmentLoadCommand(L: Command); |
| 307 | StringRef SegName = SLC.segname; |
| 308 | if (SegName == "__TEXT" ) |
| 309 | return SLC.vmaddr; |
| 310 | } else if (Command.C.cmd == MachO::LC_SEGMENT_64) { |
| 311 | MachO::segment_command_64 SLC = MachO->getSegment64LoadCommand(L: Command); |
| 312 | StringRef SegName = SLC.segname; |
| 313 | if (SegName == "__TEXT" ) |
| 314 | return SLC.vmaddr; |
| 315 | } |
| 316 | } |
| 317 | return std::nullopt; |
| 318 | } |
| 319 | |
| 320 | /// Determine the virtual address that is considered the base address of an |
| 321 | /// object file. |
| 322 | /// |
| 323 | /// Since GSYM files are used for symbolication, many clients will need to |
| 324 | /// easily adjust addresses they find in stack traces so the lookups happen |
| 325 | /// on unslid addresses from the original object file. If the base address of |
| 326 | /// a GSYM file is set to the base address of the image, then this address |
| 327 | /// adjusting is much easier. |
| 328 | /// |
| 329 | /// \param Obj An object file we will search. |
| 330 | /// |
| 331 | /// \returns A valid image base address if we are able to extract one. |
| 332 | static std::optional<uint64_t> getImageBaseAddress(object::ObjectFile &Obj) { |
| 333 | if (const auto *MachO = dyn_cast<object::MachOObjectFile>(Val: &Obj)) |
| 334 | return getImageBaseAddress(MachO); |
| 335 | else if (const auto *ELFObj = dyn_cast<object::ELF32LEObjectFile>(Val: &Obj)) |
| 336 | return getImageBaseAddress(ELFFile: ELFObj->getELFFile()); |
| 337 | else if (const auto *ELFObj = dyn_cast<object::ELF32BEObjectFile>(Val: &Obj)) |
| 338 | return getImageBaseAddress(ELFFile: ELFObj->getELFFile()); |
| 339 | else if (const auto *ELFObj = dyn_cast<object::ELF64LEObjectFile>(Val: &Obj)) |
| 340 | return getImageBaseAddress(ELFFile: ELFObj->getELFFile()); |
| 341 | else if (const auto *ELFObj = dyn_cast<object::ELF64BEObjectFile>(Val: &Obj)) |
| 342 | return getImageBaseAddress(ELFFile: ELFObj->getELFFile()); |
| 343 | return std::nullopt; |
| 344 | } |
| 345 | |
| 346 | static llvm::Error handleObjectFile(ObjectFile &Obj, const std::string &OutFile, |
| 347 | OutputAggregator &Out) { |
| 348 | auto ThreadCount = |
| 349 | NumThreads > 0 ? NumThreads : std::thread::hardware_concurrency(); |
| 350 | |
| 351 | GsymCreator Gsym(Quiet); |
| 352 | |
| 353 | // See if we can figure out the base address for a given object file, and if |
| 354 | // we can, then set the base address to use to this value. This will ease |
| 355 | // symbolication since clients can slide the GSYM lookup addresses by using |
| 356 | // the load bias of the shared library. |
| 357 | if (auto ImageBaseAddr = getImageBaseAddress(Obj)) |
| 358 | Gsym.setBaseAddress(*ImageBaseAddr); |
| 359 | |
| 360 | // We need to know where the valid sections are that contain instructions. |
| 361 | // See header documentation for DWARFTransformer::SetValidTextRanges() for |
| 362 | // defails. |
| 363 | AddressRanges ; |
| 364 | for (const object::SectionRef &Sect : Obj.sections()) { |
| 365 | if (!Sect.isText()) |
| 366 | continue; |
| 367 | const uint64_t Size = Sect.getSize(); |
| 368 | if (Size == 0) |
| 369 | continue; |
| 370 | const uint64_t StartAddr = Sect.getAddress(); |
| 371 | TextRanges.insert(Range: AddressRange(StartAddr, StartAddr + Size)); |
| 372 | } |
| 373 | |
| 374 | // Make sure there is DWARF to convert first. |
| 375 | std::unique_ptr<DWARFContext> DICtx = DWARFContext::create( |
| 376 | Obj, |
| 377 | /*RelocAction=*/DWARFContext::ProcessDebugRelocations::Process, |
| 378 | L: nullptr, |
| 379 | /*DWPName=*/"" , |
| 380 | /*RecoverableErrorHandler=*/WithColor::defaultErrorHandler, |
| 381 | /*WarningHandler=*/WithColor::defaultWarningHandler, |
| 382 | /*ThreadSafe*/true); |
| 383 | if (!DICtx) |
| 384 | return createStringError(EC: std::errc::invalid_argument, |
| 385 | Fmt: "unable to create DWARF context" ); |
| 386 | |
| 387 | // Make a DWARF transformer object and populate the ranges of the code |
| 388 | // so we don't end up adding invalid functions to GSYM data. |
| 389 | DwarfTransformer DT(*DICtx, Gsym, LoadDwarfCallSites); |
| 390 | if (!TextRanges.empty()) |
| 391 | Gsym.SetValidTextRanges(TextRanges); |
| 392 | |
| 393 | // Convert all DWARF to GSYM. |
| 394 | if (auto Err = DT.convert(NumThreads: ThreadCount, OS&: Out)) |
| 395 | return Err; |
| 396 | |
| 397 | // If enabled, merge functions with identical address ranges as merged |
| 398 | // functions in the first FunctionInfo with that address range. Do this right |
| 399 | // after loading the DWARF data so we don't have to deal with functions from |
| 400 | // the symbol table. |
| 401 | if (UseMergedFunctions) |
| 402 | Gsym.prepareMergedFunctions(Out); |
| 403 | |
| 404 | // Get the UUID and convert symbol table to GSYM. |
| 405 | if (auto Err = ObjectFileTransformer::convert(Obj, Output&: Out, Gsym)) |
| 406 | return Err; |
| 407 | |
| 408 | // If any call site YAML files were specified, load them now. |
| 409 | if (!CallSiteYamlPath.empty()) |
| 410 | if (auto Err = Gsym.loadCallSitesFromYAML(YAMLFile: CallSiteYamlPath)) |
| 411 | return Err; |
| 412 | |
| 413 | // Finalize the GSYM to make it ready to save to disk. This will remove |
| 414 | // duplicate FunctionInfo entries where we might have found an entry from |
| 415 | // debug info and also a symbol table entry from the object file. |
| 416 | if (auto Err = Gsym.finalize(OS&: Out)) |
| 417 | return Err; |
| 418 | |
| 419 | // Save the GSYM file to disk. |
| 420 | llvm::endianness Endian = Obj.makeTriple().isLittleEndian() |
| 421 | ? llvm::endianness::little |
| 422 | : llvm::endianness::big; |
| 423 | |
| 424 | std::optional<uint64_t> OptSegmentSize; |
| 425 | if (SegmentSize > 0) |
| 426 | OptSegmentSize = SegmentSize; |
| 427 | if (auto Err = Gsym.save(Path: OutFile, ByteOrder: Endian, SegmentSize: OptSegmentSize)) |
| 428 | return Err; |
| 429 | |
| 430 | // Verify the DWARF if requested. This will ensure all the info in the DWARF |
| 431 | // can be looked up in the GSYM and that all lookups get matching data. |
| 432 | if (Verify) { |
| 433 | if (auto Err = DT.verify(GsymPath: OutFile, OS&: Out)) |
| 434 | return Err; |
| 435 | } |
| 436 | |
| 437 | return Error::success(); |
| 438 | } |
| 439 | |
| 440 | static llvm::Error handleBuffer(StringRef Filename, MemoryBufferRef Buffer, |
| 441 | const std::string &OutFile, |
| 442 | OutputAggregator &Out) { |
| 443 | Expected<std::unique_ptr<Binary>> BinOrErr = object::createBinary(Source: Buffer); |
| 444 | error(Prefix: Filename, EC: errorToErrorCode(Err: BinOrErr.takeError())); |
| 445 | |
| 446 | if (auto *Obj = dyn_cast<ObjectFile>(Val: BinOrErr->get())) { |
| 447 | Triple ObjTriple(Obj->makeTriple()); |
| 448 | auto ArchName = ObjTriple.getArchName(); |
| 449 | outs() << "Output file (" << ArchName << "): " << OutFile << "\n" ; |
| 450 | if (auto Err = handleObjectFile(Obj&: *Obj, OutFile, Out)) |
| 451 | return Err; |
| 452 | } else if (auto *Fat = dyn_cast<MachOUniversalBinary>(Val: BinOrErr->get())) { |
| 453 | // Iterate over all contained architectures and filter out any that were |
| 454 | // not specified with the "--arch <arch>" option. If the --arch option was |
| 455 | // not specified on the command line, we will process all architectures. |
| 456 | std::vector<std::unique_ptr<MachOObjectFile>> FilterObjs; |
| 457 | for (auto &ObjForArch : Fat->objects()) { |
| 458 | if (auto MachOOrErr = ObjForArch.getAsObjectFile()) { |
| 459 | auto &Obj = **MachOOrErr; |
| 460 | if (filterArch(Obj)) |
| 461 | FilterObjs.emplace_back(args: MachOOrErr->release()); |
| 462 | } else { |
| 463 | error(Prefix: Filename, Err: MachOOrErr.takeError()); |
| 464 | } |
| 465 | } |
| 466 | if (FilterObjs.empty()) |
| 467 | error(Prefix: Filename, Err: createStringError(EC: std::errc::invalid_argument, |
| 468 | Fmt: "no matching architectures found" )); |
| 469 | |
| 470 | // Now handle each architecture we need to convert. |
| 471 | for (auto &Obj : FilterObjs) { |
| 472 | Triple ObjTriple(Obj->getArchTriple()); |
| 473 | auto ArchName = ObjTriple.getArchName(); |
| 474 | std::string ArchOutFile(OutFile); |
| 475 | // If we are only handling a single architecture, then we will use the |
| 476 | // normal output file. If we are handling multiple architectures append |
| 477 | // the architecture name to the end of the out file path so that we |
| 478 | // don't overwrite the previous architecture's gsym file. |
| 479 | if (FilterObjs.size() > 1) { |
| 480 | ArchOutFile.append(n: 1, c: '.'); |
| 481 | ArchOutFile.append(str: ArchName.str()); |
| 482 | } |
| 483 | outs() << "Output file (" << ArchName << "): " << ArchOutFile << "\n" ; |
| 484 | if (auto Err = handleObjectFile(Obj&: *Obj, OutFile: ArchOutFile, Out)) |
| 485 | return Err; |
| 486 | } |
| 487 | } |
| 488 | return Error::success(); |
| 489 | } |
| 490 | |
| 491 | static llvm::Error handleFileConversionToGSYM(StringRef Filename, |
| 492 | const std::string &OutFile, |
| 493 | OutputAggregator &Out) { |
| 494 | ErrorOr<std::unique_ptr<MemoryBuffer>> BuffOrErr = |
| 495 | MemoryBuffer::getFileOrSTDIN(Filename); |
| 496 | error(Prefix: Filename, EC: BuffOrErr.getError()); |
| 497 | std::unique_ptr<MemoryBuffer> Buffer = std::move(BuffOrErr.get()); |
| 498 | return handleBuffer(Filename, Buffer: *Buffer, OutFile, Out); |
| 499 | } |
| 500 | |
| 501 | static llvm::Error convertFileToGSYM(OutputAggregator &Out) { |
| 502 | // Expand any .dSYM bundles to the individual object files contained therein. |
| 503 | std::vector<std::string> Objects; |
| 504 | std::string OutFile = OutputFilename; |
| 505 | if (OutFile.empty()) { |
| 506 | OutFile = ConvertFilename; |
| 507 | OutFile += ".gsym" ; |
| 508 | } |
| 509 | |
| 510 | Out << "Input file: " << ConvertFilename << "\n" ; |
| 511 | |
| 512 | if (auto DsymObjectsOrErr = |
| 513 | MachOObjectFile::findDsymObjectMembers(Path: ConvertFilename)) { |
| 514 | if (DsymObjectsOrErr->empty()) |
| 515 | Objects.push_back(x: ConvertFilename); |
| 516 | else |
| 517 | llvm::append_range(C&: Objects, R&: *DsymObjectsOrErr); |
| 518 | } else { |
| 519 | error(Err: DsymObjectsOrErr.takeError()); |
| 520 | } |
| 521 | |
| 522 | for (StringRef Object : Objects) |
| 523 | if (Error Err = handleFileConversionToGSYM(Filename: Object, OutFile, Out)) |
| 524 | return Err; |
| 525 | return Error::success(); |
| 526 | } |
| 527 | |
| 528 | static void doLookup(GsymReader &Gsym, uint64_t Addr, raw_ostream &OS) { |
| 529 | if (UseMergedFunctions) { |
| 530 | if (auto Results = Gsym.lookupAll(Addr)) { |
| 531 | // If we have filters, count matching results first |
| 532 | size_t NumMatching = Results->size(); |
| 533 | if (!MergedFunctionsFilters.empty()) { |
| 534 | NumMatching = 0; |
| 535 | for (const auto &Result : *Results) { |
| 536 | bool Matches = false; |
| 537 | for (const auto &Filter : MergedFunctionsFilters) { |
| 538 | Regex Pattern(Filter); |
| 539 | if (Pattern.match(String: Result.FuncName)) { |
| 540 | Matches = true; |
| 541 | break; |
| 542 | } |
| 543 | } |
| 544 | if (Matches) |
| 545 | NumMatching++; |
| 546 | } |
| 547 | } |
| 548 | |
| 549 | OS << "Found " << NumMatching << " function" |
| 550 | << (NumMatching != 1 ? "s" : "" ) << " at address " << HEX64(Addr) |
| 551 | << ":\n" ; |
| 552 | |
| 553 | for (size_t i = 0; i < Results->size(); ++i) { |
| 554 | // Skip if doesn't match any filter |
| 555 | if (!MergedFunctionsFilters.empty()) { |
| 556 | bool Matches = false; |
| 557 | for (const auto &Filter : MergedFunctionsFilters) { |
| 558 | Regex Pattern(Filter); |
| 559 | if (Pattern.match(String: Results->at(n: i).FuncName)) { |
| 560 | Matches = true; |
| 561 | break; |
| 562 | } |
| 563 | } |
| 564 | if (!Matches) |
| 565 | continue; |
| 566 | } |
| 567 | |
| 568 | OS << " " << Results->at(n: i); |
| 569 | |
| 570 | if (i != Results->size() - 1) |
| 571 | OS << "\n" ; |
| 572 | } |
| 573 | } |
| 574 | } else { /* UseMergedFunctions == false */ |
| 575 | if (auto Result = Gsym.lookup(Addr)) { |
| 576 | // If verbose is enabled dump the full function info for the address. |
| 577 | if (Verbose) { |
| 578 | if (auto FI = Gsym.getFunctionInfo(Addr)) { |
| 579 | OS << "FunctionInfo for " << HEX64(Addr) << ":\n" ; |
| 580 | Gsym.dump(OS, FI: *FI); |
| 581 | OS << "\nLookupResult for " << HEX64(Addr) << ":\n" ; |
| 582 | } |
| 583 | } |
| 584 | // Don't print call site info if --merged-functions is not specified. |
| 585 | Result->CallSiteFuncRegex.clear(); |
| 586 | OS << Result.get(); |
| 587 | } else { |
| 588 | if (Verbose) |
| 589 | OS << "\nLookupResult for " << HEX64(Addr) << ":\n" ; |
| 590 | OS << HEX64(Addr) << ": " ; |
| 591 | logAllUnhandledErrors(E: Result.takeError(), OS, ErrorBanner: "error: " ); |
| 592 | } |
| 593 | if (Verbose) |
| 594 | OS << "\n" ; |
| 595 | } |
| 596 | } |
| 597 | |
| 598 | int llvm_gsymutil_main(int argc, char **argv, const llvm::ToolContext &) { |
| 599 | // Print a stack trace if we signal out. |
| 600 | sys::PrintStackTraceOnErrorSignal(Argv0: argv[0]); |
| 601 | PrettyStackTraceProgram X(argc, argv); |
| 602 | llvm_shutdown_obj Y; // Call llvm_shutdown() on exit. |
| 603 | |
| 604 | llvm::InitializeAllTargets(); |
| 605 | |
| 606 | parseArgs(argc, argv); |
| 607 | |
| 608 | raw_ostream &OS = outs(); |
| 609 | |
| 610 | OutputAggregator Aggregation(&OS); |
| 611 | if (!ConvertFilename.empty()) { |
| 612 | // Convert DWARF to GSYM |
| 613 | if (!InputFilenames.empty()) { |
| 614 | OS << "error: no input files can be specified when using the --convert " |
| 615 | "option.\n" ; |
| 616 | return 1; |
| 617 | } |
| 618 | // Call error() if we have an error and it will exit with a status of 1 |
| 619 | if (auto Err = convertFileToGSYM(Out&: Aggregation)) |
| 620 | error(Prefix: "DWARF conversion failed: " , Err: std::move(Err)); |
| 621 | |
| 622 | // Report the errors from aggregator: |
| 623 | Aggregation.EnumerateResults(handleCounts: [&](StringRef category, unsigned count) { |
| 624 | OS << category << " occurred " << count << " time(s)\n" ; |
| 625 | }); |
| 626 | if (!JsonSummaryFile.empty()) { |
| 627 | std::error_code EC; |
| 628 | raw_fd_ostream JsonStream(JsonSummaryFile, EC, sys::fs::OF_Text); |
| 629 | if (EC) { |
| 630 | OS << "error opening aggregate error json file '" << JsonSummaryFile |
| 631 | << "' for writing: " << EC.message() << '\n'; |
| 632 | return 1; |
| 633 | } |
| 634 | |
| 635 | llvm::json::Object Categories; |
| 636 | uint64_t ErrorCount = 0; |
| 637 | Aggregation.EnumerateResults(handleCounts: [&](StringRef Category, unsigned Count) { |
| 638 | llvm::json::Object Val; |
| 639 | Val.try_emplace(K: "count" , Args&: Count); |
| 640 | Categories.try_emplace(K: Category, Args: std::move(Val)); |
| 641 | ErrorCount += Count; |
| 642 | }); |
| 643 | llvm::json::Object RootNode; |
| 644 | RootNode.try_emplace(K: "error-categories" , Args: std::move(Categories)); |
| 645 | RootNode.try_emplace(K: "error-count" , Args&: ErrorCount); |
| 646 | |
| 647 | JsonStream << llvm::json::Value(std::move(RootNode)); |
| 648 | } |
| 649 | return 0; |
| 650 | } |
| 651 | |
| 652 | if (LookupAddressesFromStdin) { |
| 653 | if (!LookupAddresses.empty() || !InputFilenames.empty()) { |
| 654 | OS << "error: no input files or addresses can be specified when using " |
| 655 | "the --addresses-from-stdin " |
| 656 | "option.\n" ; |
| 657 | return 1; |
| 658 | } |
| 659 | |
| 660 | std::string InputLine; |
| 661 | std::string CurrentGSYMPath; |
| 662 | std::optional<Expected<GsymReader>> CurrentGsym; |
| 663 | |
| 664 | while (std::getline(is&: std::cin, str&: InputLine)) { |
| 665 | // Strip newline characters. |
| 666 | std::string StrippedInputLine(InputLine); |
| 667 | llvm::erase_if(C&: StrippedInputLine, |
| 668 | P: [](char c) { return c == '\r' || c == '\n'; }); |
| 669 | |
| 670 | StringRef AddrStr, GSYMPath; |
| 671 | std::tie(args&: AddrStr, args&: GSYMPath) = |
| 672 | llvm::StringRef{StrippedInputLine}.split(Separator: ' '); |
| 673 | |
| 674 | if (GSYMPath != CurrentGSYMPath) { |
| 675 | CurrentGsym = GsymReader::openFile(Path: GSYMPath); |
| 676 | if (!*CurrentGsym) |
| 677 | error(Prefix: GSYMPath, Err: CurrentGsym->takeError()); |
| 678 | CurrentGSYMPath = GSYMPath; |
| 679 | } |
| 680 | |
| 681 | uint64_t Addr; |
| 682 | if (AddrStr.getAsInteger(Radix: 0, Result&: Addr)) { |
| 683 | OS << "error: invalid address " << AddrStr |
| 684 | << ", expected: Address GsymFile.\n" ; |
| 685 | return 1; |
| 686 | } |
| 687 | |
| 688 | doLookup(Gsym&: **CurrentGsym, Addr, OS); |
| 689 | |
| 690 | OS << "\n" ; |
| 691 | OS.flush(); |
| 692 | } |
| 693 | |
| 694 | return EXIT_SUCCESS; |
| 695 | } |
| 696 | |
| 697 | // Dump or access data inside GSYM files |
| 698 | for (const auto &GSYMPath : InputFilenames) { |
| 699 | auto Gsym = GsymReader::openFile(Path: GSYMPath); |
| 700 | if (!Gsym) |
| 701 | error(Prefix: GSYMPath, Err: Gsym.takeError()); |
| 702 | |
| 703 | if (LookupAddresses.empty()) { |
| 704 | Gsym->dump(OS&: outs()); |
| 705 | continue; |
| 706 | } |
| 707 | |
| 708 | // Lookup an address in a GSYM file and print any matches. |
| 709 | OS << "Looking up addresses in \"" << GSYMPath << "\":\n" ; |
| 710 | for (auto Addr : LookupAddresses) { |
| 711 | doLookup(Gsym&: *Gsym, Addr, OS); |
| 712 | } |
| 713 | } |
| 714 | return EXIT_SUCCESS; |
| 715 | } |
| 716 | |