1 //===- ArchiveWriter.cpp - ar File Format implementation --------*- C++ -*-===// 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 // This file defines the writeArchive function. 11 // 12 //===----------------------------------------------------------------------===// 13 14 #include "llvm/Object/ArchiveWriter.h" 15 #include "llvm/ADT/ArrayRef.h" 16 #include "llvm/ADT/StringRef.h" 17 #include "llvm/BinaryFormat/Magic.h" 18 #include "llvm/IR/LLVMContext.h" 19 #include "llvm/Object/Archive.h" 20 #include "llvm/Object/ObjectFile.h" 21 #include "llvm/Object/SymbolicFile.h" 22 #include "llvm/Support/EndianStream.h" 23 #include "llvm/Support/Errc.h" 24 #include "llvm/Support/ErrorHandling.h" 25 #include "llvm/Support/Format.h" 26 #include "llvm/Support/Path.h" 27 #include "llvm/Support/ToolOutputFile.h" 28 #include "llvm/Support/raw_ostream.h" 29 30 #if !defined(_MSC_VER) && !defined(__MINGW32__) 31 #include <unistd.h> 32 #else 33 #include <io.h> 34 #endif 35 36 using namespace llvm; 37 38 NewArchiveMember::NewArchiveMember(MemoryBufferRef BufRef) 39 : Buf(MemoryBuffer::getMemBuffer(BufRef, false)), 40 MemberName(BufRef.getBufferIdentifier()) {} 41 42 Expected<NewArchiveMember> 43 NewArchiveMember::getOldMember(const object::Archive::Child &OldMember, 44 bool Deterministic) { 45 Expected<llvm::MemoryBufferRef> BufOrErr = OldMember.getMemoryBufferRef(); 46 if (!BufOrErr) 47 return BufOrErr.takeError(); 48 49 NewArchiveMember M; 50 assert(M.IsNew == false); 51 M.Buf = MemoryBuffer::getMemBuffer(*BufOrErr, false); 52 M.MemberName = M.Buf->getBufferIdentifier(); 53 if (!Deterministic) { 54 auto ModTimeOrErr = OldMember.getLastModified(); 55 if (!ModTimeOrErr) 56 return ModTimeOrErr.takeError(); 57 M.ModTime = ModTimeOrErr.get(); 58 Expected<unsigned> UIDOrErr = OldMember.getUID(); 59 if (!UIDOrErr) 60 return UIDOrErr.takeError(); 61 M.UID = UIDOrErr.get(); 62 Expected<unsigned> GIDOrErr = OldMember.getGID(); 63 if (!GIDOrErr) 64 return GIDOrErr.takeError(); 65 M.GID = GIDOrErr.get(); 66 Expected<sys::fs::perms> AccessModeOrErr = OldMember.getAccessMode(); 67 if (!AccessModeOrErr) 68 return AccessModeOrErr.takeError(); 69 M.Perms = AccessModeOrErr.get(); 70 } 71 return std::move(M); 72 } 73 74 Expected<NewArchiveMember> NewArchiveMember::getFile(StringRef FileName, 75 bool Deterministic) { 76 sys::fs::file_status Status; 77 int FD; 78 if (auto EC = sys::fs::openFileForRead(FileName, FD)) 79 return errorCodeToError(EC); 80 assert(FD != -1); 81 82 if (auto EC = sys::fs::status(FD, Status)) 83 return errorCodeToError(EC); 84 85 // Opening a directory doesn't make sense. Let it fail. 86 // Linux cannot open directories with open(2), although 87 // cygwin and *bsd can. 88 if (Status.type() == sys::fs::file_type::directory_file) 89 return errorCodeToError(make_error_code(errc::is_a_directory)); 90 91 ErrorOr<std::unique_ptr<MemoryBuffer>> MemberBufferOrErr = 92 MemoryBuffer::getOpenFile(FD, FileName, Status.getSize(), false); 93 if (!MemberBufferOrErr) 94 return errorCodeToError(MemberBufferOrErr.getError()); 95 96 if (close(FD) != 0) 97 return errorCodeToError(std::error_code(errno, std::generic_category())); 98 99 NewArchiveMember M; 100 M.IsNew = true; 101 M.Buf = std::move(*MemberBufferOrErr); 102 M.MemberName = M.Buf->getBufferIdentifier(); 103 if (!Deterministic) { 104 M.ModTime = std::chrono::time_point_cast<std::chrono::seconds>( 105 Status.getLastModificationTime()); 106 M.UID = Status.getUser(); 107 M.GID = Status.getGroup(); 108 M.Perms = Status.permissions(); 109 } 110 return std::move(M); 111 } 112 113 template <typename T> 114 static void printWithSpacePadding(raw_ostream &OS, T Data, unsigned Size) { 115 uint64_t OldPos = OS.tell(); 116 OS << Data; 117 unsigned SizeSoFar = OS.tell() - OldPos; 118 assert(SizeSoFar <= Size && "Data doesn't fit in Size"); 119 OS.indent(Size - SizeSoFar); 120 } 121 122 static bool isBSDLike(object::Archive::Kind Kind) { 123 switch (Kind) { 124 case object::Archive::K_GNU: 125 case object::Archive::K_GNU64: 126 return false; 127 case object::Archive::K_BSD: 128 case object::Archive::K_DARWIN: 129 return true; 130 case object::Archive::K_DARWIN64: 131 case object::Archive::K_COFF: 132 break; 133 } 134 llvm_unreachable("not supported for writting"); 135 } 136 137 template <class T> 138 static void print(raw_ostream &Out, object::Archive::Kind Kind, T Val) { 139 support::endian::write(Out, Val, 140 isBSDLike(Kind) ? support::little : support::big); 141 } 142 143 static void printRestOfMemberHeader( 144 raw_ostream &Out, const sys::TimePoint<std::chrono::seconds> &ModTime, 145 unsigned UID, unsigned GID, unsigned Perms, unsigned Size) { 146 printWithSpacePadding(Out, sys::toTimeT(ModTime), 12); 147 148 // The format has only 6 chars for uid and gid. Truncate if the provided 149 // values don't fit. 150 printWithSpacePadding(Out, UID % 1000000, 6); 151 printWithSpacePadding(Out, GID % 1000000, 6); 152 153 printWithSpacePadding(Out, format("%o", Perms), 8); 154 printWithSpacePadding(Out, Size, 10); 155 Out << "`\n"; 156 } 157 158 static void 159 printGNUSmallMemberHeader(raw_ostream &Out, StringRef Name, 160 const sys::TimePoint<std::chrono::seconds> &ModTime, 161 unsigned UID, unsigned GID, unsigned Perms, 162 unsigned Size) { 163 printWithSpacePadding(Out, Twine(Name) + "/", 16); 164 printRestOfMemberHeader(Out, ModTime, UID, GID, Perms, Size); 165 } 166 167 static void 168 printBSDMemberHeader(raw_ostream &Out, uint64_t Pos, StringRef Name, 169 const sys::TimePoint<std::chrono::seconds> &ModTime, 170 unsigned UID, unsigned GID, unsigned Perms, 171 unsigned Size) { 172 uint64_t PosAfterHeader = Pos + 60 + Name.size(); 173 // Pad so that even 64 bit object files are aligned. 174 unsigned Pad = OffsetToAlignment(PosAfterHeader, 8); 175 unsigned NameWithPadding = Name.size() + Pad; 176 printWithSpacePadding(Out, Twine("#1/") + Twine(NameWithPadding), 16); 177 printRestOfMemberHeader(Out, ModTime, UID, GID, Perms, 178 NameWithPadding + Size); 179 Out << Name; 180 while (Pad--) 181 Out.write(uint8_t(0)); 182 } 183 184 static bool useStringTable(bool Thin, StringRef Name) { 185 return Thin || Name.size() >= 16 || Name.contains('/'); 186 } 187 188 // Compute the relative path from From to To. 189 static std::string computeRelativePath(StringRef From, StringRef To) { 190 if (sys::path::is_absolute(From) || sys::path::is_absolute(To)) 191 return To; 192 193 StringRef DirFrom = sys::path::parent_path(From); 194 auto FromI = sys::path::begin(DirFrom); 195 auto ToI = sys::path::begin(To); 196 while (*FromI == *ToI) { 197 ++FromI; 198 ++ToI; 199 } 200 201 SmallString<128> Relative; 202 for (auto FromE = sys::path::end(DirFrom); FromI != FromE; ++FromI) 203 sys::path::append(Relative, ".."); 204 205 for (auto ToE = sys::path::end(To); ToI != ToE; ++ToI) 206 sys::path::append(Relative, *ToI); 207 208 #ifdef _WIN32 209 // Replace backslashes with slashes so that the path is portable between *nix 210 // and Windows. 211 std::replace(Relative.begin(), Relative.end(), '\\', '/'); 212 #endif 213 214 return Relative.str(); 215 } 216 217 static bool is64BitKind(object::Archive::Kind Kind) { 218 switch (Kind) { 219 case object::Archive::K_GNU: 220 case object::Archive::K_BSD: 221 case object::Archive::K_DARWIN: 222 case object::Archive::K_COFF: 223 return false; 224 case object::Archive::K_DARWIN64: 225 case object::Archive::K_GNU64: 226 return true; 227 } 228 llvm_unreachable("not supported for writting"); 229 } 230 231 static void addToStringTable(raw_ostream &Out, StringRef ArcName, 232 const NewArchiveMember &M, bool Thin) { 233 StringRef ID = M.Buf->getBufferIdentifier(); 234 if (Thin) { 235 if (M.IsNew) 236 Out << computeRelativePath(ArcName, ID); 237 else 238 Out << ID; 239 } else 240 Out << M.MemberName; 241 Out << "/\n"; 242 } 243 244 static void printMemberHeader(raw_ostream &Out, uint64_t Pos, 245 raw_ostream &StringTable, 246 object::Archive::Kind Kind, bool Thin, 247 StringRef ArcName, const NewArchiveMember &M, 248 unsigned Size) { 249 if (isBSDLike(Kind)) 250 return printBSDMemberHeader(Out, Pos, M.MemberName, M.ModTime, M.UID, M.GID, 251 M.Perms, Size); 252 if (!useStringTable(Thin, M.MemberName)) 253 return printGNUSmallMemberHeader(Out, M.MemberName, M.ModTime, M.UID, M.GID, 254 M.Perms, Size); 255 Out << '/'; 256 uint64_t NamePos = StringTable.tell(); 257 addToStringTable(StringTable, ArcName, M, Thin); 258 printWithSpacePadding(Out, NamePos, 15); 259 printRestOfMemberHeader(Out, M.ModTime, M.UID, M.GID, M.Perms, Size); 260 } 261 262 namespace { 263 struct MemberData { 264 std::vector<unsigned> Symbols; 265 std::string Header; 266 StringRef Data; 267 StringRef Padding; 268 }; 269 } // namespace 270 271 static MemberData computeStringTable(StringRef Names) { 272 unsigned Size = Names.size(); 273 unsigned Pad = OffsetToAlignment(Size, 2); 274 std::string Header; 275 raw_string_ostream Out(Header); 276 printWithSpacePadding(Out, "//", 48); 277 printWithSpacePadding(Out, Size + Pad, 10); 278 Out << "`\n"; 279 Out.flush(); 280 return {{}, std::move(Header), Names, Pad ? "\n" : ""}; 281 } 282 283 static sys::TimePoint<std::chrono::seconds> now(bool Deterministic) { 284 using namespace std::chrono; 285 286 if (!Deterministic) 287 return time_point_cast<seconds>(system_clock::now()); 288 return sys::TimePoint<seconds>(); 289 } 290 291 static bool isArchiveSymbol(const object::BasicSymbolRef &S) { 292 uint32_t Symflags = S.getFlags(); 293 if (Symflags & object::SymbolRef::SF_FormatSpecific) 294 return false; 295 if (!(Symflags & object::SymbolRef::SF_Global)) 296 return false; 297 if (Symflags & object::SymbolRef::SF_Undefined) 298 return false; 299 return true; 300 } 301 302 static void printNBits(raw_ostream &Out, object::Archive::Kind Kind, 303 uint64_t Val) { 304 if (is64BitKind(Kind)) 305 print<uint64_t>(Out, Kind, Val); 306 else 307 print<uint32_t>(Out, Kind, Val); 308 } 309 310 static void writeSymbolTable(raw_ostream &Out, object::Archive::Kind Kind, 311 bool Deterministic, ArrayRef<MemberData> Members, 312 StringRef StringTable) { 313 if (StringTable.empty()) 314 return; 315 316 unsigned NumSyms = 0; 317 for (const MemberData &M : Members) 318 NumSyms += M.Symbols.size(); 319 320 unsigned Size = 0; 321 Size += is64BitKind(Kind) ? 8 : 4; // Number of entries 322 if (isBSDLike(Kind)) 323 Size += NumSyms * 8; // Table 324 else if (is64BitKind(Kind)) 325 Size += NumSyms * 8; // Table 326 else 327 Size += NumSyms * 4; // Table 328 if (isBSDLike(Kind)) 329 Size += 4; // byte count 330 Size += StringTable.size(); 331 // ld64 expects the members to be 8-byte aligned for 64-bit content and at 332 // least 4-byte aligned for 32-bit content. Opt for the larger encoding 333 // uniformly. 334 // We do this for all bsd formats because it simplifies aligning members. 335 unsigned Alignment = isBSDLike(Kind) ? 8 : 2; 336 unsigned Pad = OffsetToAlignment(Size, Alignment); 337 Size += Pad; 338 339 if (isBSDLike(Kind)) 340 printBSDMemberHeader(Out, Out.tell(), "__.SYMDEF", now(Deterministic), 0, 0, 341 0, Size); 342 else if (is64BitKind(Kind)) 343 printGNUSmallMemberHeader(Out, "/SYM64", now(Deterministic), 0, 0, 0, Size); 344 else 345 printGNUSmallMemberHeader(Out, "", now(Deterministic), 0, 0, 0, Size); 346 347 uint64_t Pos = Out.tell() + Size; 348 349 if (isBSDLike(Kind)) 350 print<uint32_t>(Out, Kind, NumSyms * 8); 351 else 352 printNBits(Out, Kind, NumSyms); 353 354 for (const MemberData &M : Members) { 355 for (unsigned StringOffset : M.Symbols) { 356 if (isBSDLike(Kind)) 357 print<uint32_t>(Out, Kind, StringOffset); 358 printNBits(Out, Kind, Pos); // member offset 359 } 360 Pos += M.Header.size() + M.Data.size() + M.Padding.size(); 361 } 362 363 if (isBSDLike(Kind)) 364 // byte count of the string table 365 print<uint32_t>(Out, Kind, StringTable.size()); 366 Out << StringTable; 367 368 while (Pad--) 369 Out.write(uint8_t(0)); 370 } 371 372 static Expected<std::vector<unsigned>> 373 getSymbols(MemoryBufferRef Buf, raw_ostream &SymNames, bool &HasObject) { 374 std::vector<unsigned> Ret; 375 LLVMContext Context; 376 377 Expected<std::unique_ptr<object::SymbolicFile>> ObjOrErr = 378 object::SymbolicFile::createSymbolicFile(Buf, llvm::file_magic::unknown, 379 &Context); 380 if (!ObjOrErr) { 381 // FIXME: check only for "not an object file" errors. 382 consumeError(ObjOrErr.takeError()); 383 return Ret; 384 } 385 386 HasObject = true; 387 object::SymbolicFile &Obj = *ObjOrErr.get(); 388 for (const object::BasicSymbolRef &S : Obj.symbols()) { 389 if (!isArchiveSymbol(S)) 390 continue; 391 Ret.push_back(SymNames.tell()); 392 if (auto EC = S.printName(SymNames)) 393 return errorCodeToError(EC); 394 SymNames << '\0'; 395 } 396 return Ret; 397 } 398 399 static Expected<std::vector<MemberData>> 400 computeMemberData(raw_ostream &StringTable, raw_ostream &SymNames, 401 object::Archive::Kind Kind, bool Thin, StringRef ArcName, 402 ArrayRef<NewArchiveMember> NewMembers) { 403 static char PaddingData[8] = {'\n', '\n', '\n', '\n', '\n', '\n', '\n', '\n'}; 404 405 // This ignores the symbol table, but we only need the value mod 8 and the 406 // symbol table is aligned to be a multiple of 8 bytes 407 uint64_t Pos = 0; 408 409 std::vector<MemberData> Ret; 410 bool HasObject = false; 411 for (const NewArchiveMember &M : NewMembers) { 412 std::string Header; 413 raw_string_ostream Out(Header); 414 415 MemoryBufferRef Buf = M.Buf->getMemBufferRef(); 416 StringRef Data = Thin ? "" : Buf.getBuffer(); 417 418 // ld64 expects the members to be 8-byte aligned for 64-bit content and at 419 // least 4-byte aligned for 32-bit content. Opt for the larger encoding 420 // uniformly. This matches the behaviour with cctools and ensures that ld64 421 // is happy with archives that we generate. 422 unsigned MemberPadding = Kind == object::Archive::K_DARWIN 423 ? OffsetToAlignment(Data.size(), 8) 424 : 0; 425 unsigned TailPadding = OffsetToAlignment(Data.size() + MemberPadding, 2); 426 StringRef Padding = StringRef(PaddingData, MemberPadding + TailPadding); 427 428 printMemberHeader(Out, Pos, StringTable, Kind, Thin, ArcName, M, 429 Buf.getBufferSize() + MemberPadding); 430 Out.flush(); 431 432 Expected<std::vector<unsigned>> Symbols = 433 getSymbols(Buf, SymNames, HasObject); 434 if (auto E = Symbols.takeError()) 435 return std::move(E); 436 437 Pos += Header.size() + Data.size() + Padding.size(); 438 Ret.push_back({std::move(*Symbols), std::move(Header), Data, Padding}); 439 } 440 // If there are no symbols, emit an empty symbol table, to satisfy Solaris 441 // tools, older versions of which expect a symbol table in a non-empty 442 // archive, regardless of whether there are any symbols in it. 443 if (HasObject && SymNames.tell() == 0) 444 SymNames << '\0' << '\0' << '\0'; 445 return Ret; 446 } 447 448 Error llvm::writeArchive(StringRef ArcName, 449 ArrayRef<NewArchiveMember> NewMembers, 450 bool WriteSymtab, object::Archive::Kind Kind, 451 bool Deterministic, bool Thin, 452 std::unique_ptr<MemoryBuffer> OldArchiveBuf) { 453 assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode"); 454 455 SmallString<0> SymNamesBuf; 456 raw_svector_ostream SymNames(SymNamesBuf); 457 SmallString<0> StringTableBuf; 458 raw_svector_ostream StringTable(StringTableBuf); 459 460 Expected<std::vector<MemberData>> DataOrErr = 461 computeMemberData(StringTable, SymNames, Kind, Thin, ArcName, NewMembers); 462 if (Error E = DataOrErr.takeError()) 463 return E; 464 std::vector<MemberData> &Data = *DataOrErr; 465 466 if (!StringTableBuf.empty()) 467 Data.insert(Data.begin(), computeStringTable(StringTableBuf)); 468 469 // We would like to detect if we need to switch to a 64-bit symbol table. 470 if (WriteSymtab) { 471 uint64_t MaxOffset = 0; 472 uint64_t LastOffset = MaxOffset; 473 for (const auto& M : Data) { 474 // Record the start of the member's offset 475 LastOffset = MaxOffset; 476 // Account for the size of each part associated with the member. 477 MaxOffset += M.Header.size() + M.Data.size() + M.Padding.size(); 478 // We assume 32-bit symbols to see if 32-bit symbols are possible or not. 479 MaxOffset += M.Symbols.size() * 4; 480 } 481 482 // The SYM64 format is used when an archive's member offsets are larger than 483 // 32-bits can hold. The need for this shift in format is detected by 484 // writeArchive. To test this we need to generate a file with a member that 485 // has an offset larger than 32-bits but this demands a very slow test. To 486 // speed the test up we use this environment variable to pretend like the 487 // cutoff happens before 32-bits and instead happens at some much smaller 488 // value. 489 const char *Sym64Env = std::getenv("SYM64_THRESHOLD"); 490 int Sym64Threshold = 32; 491 if (Sym64Env) 492 StringRef(Sym64Env).getAsInteger(10, Sym64Threshold); 493 494 // If LastOffset isn't going to fit in a 32-bit varible we need to switch 495 // to 64-bit. Note that the file can be larger than 4GB as long as the last 496 // member starts before the 4GB offset. 497 if (LastOffset >= (1ULL << Sym64Threshold)) 498 Kind = object::Archive::K_GNU64; 499 } 500 501 Expected<sys::fs::TempFile> Temp = 502 sys::fs::TempFile::create(ArcName + ".temp-archive-%%%%%%%.a"); 503 if (!Temp) 504 return Temp.takeError(); 505 506 raw_fd_ostream Out(Temp->FD, false); 507 if (Thin) 508 Out << "!<thin>\n"; 509 else 510 Out << "!<arch>\n"; 511 512 if (WriteSymtab) 513 writeSymbolTable(Out, Kind, Deterministic, Data, SymNamesBuf); 514 515 for (const MemberData &M : Data) 516 Out << M.Header << M.Data << M.Padding; 517 518 Out.flush(); 519 520 // At this point, we no longer need whatever backing memory 521 // was used to generate the NewMembers. On Windows, this buffer 522 // could be a mapped view of the file we want to replace (if 523 // we're updating an existing archive, say). In that case, the 524 // rename would still succeed, but it would leave behind a 525 // temporary file (actually the original file renamed) because 526 // a file cannot be deleted while there's a handle open on it, 527 // only renamed. So by freeing this buffer, this ensures that 528 // the last open handle on the destination file, if any, is 529 // closed before we attempt to rename. 530 OldArchiveBuf.reset(); 531 532 return Temp->keep(ArcName); 533 } 534