1//===- ArchiveWriter.cpp - ar File Format implementation --------*- C++ -*-===//
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// This file defines the writeArchive function.
10//
11//===----------------------------------------------------------------------===//
12
13#include "llvm/Object/ArchiveWriter.h"
14#include "llvm/ADT/ArrayRef.h"
15#include "llvm/ADT/StringMap.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/COFF.h"
21#include "llvm/Object/COFFImportFile.h"
22#include "llvm/Object/Error.h"
23#include "llvm/Object/IRObjectFile.h"
24#include "llvm/Object/MachO.h"
25#include "llvm/Object/ObjectFile.h"
26#include "llvm/Object/SymbolicFile.h"
27#include "llvm/Object/XCOFFObjectFile.h"
28#include "llvm/Support/Alignment.h"
29#include "llvm/Support/EndianStream.h"
30#include "llvm/Support/Errc.h"
31#include "llvm/Support/ErrorHandling.h"
32#include "llvm/Support/Format.h"
33#include "llvm/Support/MathExtras.h"
34#include "llvm/Support/Path.h"
35#include "llvm/Support/SmallVectorMemoryBuffer.h"
36#include "llvm/Support/raw_ostream.h"
37
38#include <cerrno>
39#include <map>
40
41#if !defined(_MSC_VER) && !defined(__MINGW32__)
42#include <unistd.h>
43#else
44#include <io.h>
45#endif
46
47using namespace llvm;
48using namespace llvm::object;
49
50struct SymMap {
51 bool UseECMap = false;
52 std::map<std::string, uint16_t> Map;
53 std::map<std::string, uint16_t> ECMap;
54};
55
56NewArchiveMember::NewArchiveMember(MemoryBufferRef BufRef)
57 : Buf(MemoryBuffer::getMemBuffer(Ref: BufRef, RequiresNullTerminator: false)),
58 MemberName(BufRef.getBufferIdentifier()) {}
59
60object::Archive::Kind NewArchiveMember::detectKindFromObject() const {
61 auto MemBufferRef = this->Buf->getMemBufferRef();
62 Expected<std::unique_ptr<object::ObjectFile>> OptionalObject =
63 object::ObjectFile::createObjectFile(Object: MemBufferRef);
64
65 if (OptionalObject) {
66 if (isa<object::MachOObjectFile>(Val: **OptionalObject))
67 return object::Archive::K_DARWIN;
68 if (isa<object::XCOFFObjectFile>(Val: **OptionalObject))
69 return object::Archive::K_AIXBIG;
70 if (isa<object::COFFObjectFile>(Val: **OptionalObject) ||
71 isa<object::COFFImportFile>(Val: **OptionalObject))
72 return object::Archive::K_COFF;
73 return object::Archive::K_GNU;
74 }
75
76 // Squelch the error in case we had a non-object file.
77 consumeError(Err: OptionalObject.takeError());
78
79 // If we're adding a bitcode file to the archive, detect the Archive kind
80 // based on the target triple.
81 LLVMContext Context;
82 if (identify_magic(magic: MemBufferRef.getBuffer()) == file_magic::bitcode) {
83 if (auto ObjOrErr = object::SymbolicFile::createSymbolicFile(
84 Object: MemBufferRef, Type: file_magic::bitcode, Context: &Context)) {
85 auto &IRObject = cast<object::IRObjectFile>(Val&: **ObjOrErr);
86 auto TargetTriple = Triple(IRObject.getTargetTriple());
87 return object::Archive::getDefaultKindForTriple(T: TargetTriple);
88 } else {
89 // Squelch the error in case this was not a SymbolicFile.
90 consumeError(Err: ObjOrErr.takeError());
91 }
92 }
93
94 return object::Archive::getDefaultKind();
95}
96
97Expected<NewArchiveMember>
98NewArchiveMember::getOldMember(const object::Archive::Child &OldMember,
99 bool Deterministic) {
100 Expected<llvm::MemoryBufferRef> BufOrErr = OldMember.getMemoryBufferRef();
101 if (!BufOrErr)
102 return BufOrErr.takeError();
103
104 NewArchiveMember M;
105 M.Buf = MemoryBuffer::getMemBuffer(Ref: *BufOrErr, RequiresNullTerminator: false);
106 M.MemberName = M.Buf->getBufferIdentifier();
107 if (!Deterministic) {
108 auto ModTimeOrErr = OldMember.getLastModified();
109 if (!ModTimeOrErr)
110 return ModTimeOrErr.takeError();
111 M.ModTime = ModTimeOrErr.get();
112 Expected<unsigned> UIDOrErr = OldMember.getUID();
113 if (!UIDOrErr)
114 return UIDOrErr.takeError();
115 M.UID = UIDOrErr.get();
116 Expected<unsigned> GIDOrErr = OldMember.getGID();
117 if (!GIDOrErr)
118 return GIDOrErr.takeError();
119 M.GID = GIDOrErr.get();
120 Expected<sys::fs::perms> AccessModeOrErr = OldMember.getAccessMode();
121 if (!AccessModeOrErr)
122 return AccessModeOrErr.takeError();
123 M.Perms = AccessModeOrErr.get();
124 }
125 return std::move(M);
126}
127
128Expected<NewArchiveMember> NewArchiveMember::getFile(StringRef FileName,
129 bool Deterministic) {
130 sys::fs::file_status Status;
131 auto FDOrErr = sys::fs::openNativeFileForRead(Name: FileName);
132 if (!FDOrErr)
133 return FDOrErr.takeError();
134 sys::fs::file_t FD = *FDOrErr;
135 assert(FD != sys::fs::kInvalidFile);
136
137 if (auto EC = sys::fs::status(FD, Result&: Status))
138 return errorCodeToError(EC);
139
140 // Opening a directory doesn't make sense. Let it fail.
141 // Linux cannot open directories with open(2), although
142 // cygwin and *bsd can.
143 if (Status.type() == sys::fs::file_type::directory_file)
144 return errorCodeToError(EC: make_error_code(E: errc::is_a_directory));
145
146 ErrorOr<std::unique_ptr<MemoryBuffer>> MemberBufferOrErr =
147 MemoryBuffer::getOpenFile(FD, Filename: FileName, FileSize: Status.getSize(), RequiresNullTerminator: false);
148 if (!MemberBufferOrErr)
149 return errorCodeToError(EC: MemberBufferOrErr.getError());
150
151 if (auto EC = sys::fs::closeFile(F&: FD))
152 return errorCodeToError(EC);
153
154 NewArchiveMember M;
155 M.Buf = std::move(*MemberBufferOrErr);
156 M.MemberName = M.Buf->getBufferIdentifier();
157 if (!Deterministic) {
158 M.ModTime = std::chrono::time_point_cast<std::chrono::seconds>(
159 t: Status.getLastModificationTime());
160 M.UID = Status.getUser();
161 M.GID = Status.getGroup();
162 M.Perms = Status.permissions();
163 }
164 return std::move(M);
165}
166
167template <typename T>
168static void printWithSpacePadding(raw_ostream &OS, T Data, unsigned Size) {
169 uint64_t OldPos = OS.tell();
170 OS << Data;
171 unsigned SizeSoFar = OS.tell() - OldPos;
172 assert(SizeSoFar <= Size && "Data doesn't fit in Size");
173 OS.indent(NumSpaces: Size - SizeSoFar);
174}
175
176static bool isDarwin(object::Archive::Kind Kind) {
177 return Kind == object::Archive::K_DARWIN ||
178 Kind == object::Archive::K_DARWIN64;
179}
180
181static bool isAIXBigArchive(object::Archive::Kind Kind) {
182 return Kind == object::Archive::K_AIXBIG;
183}
184
185static bool isCOFFArchive(object::Archive::Kind Kind) {
186 return Kind == object::Archive::K_COFF;
187}
188
189static bool isBSDLike(object::Archive::Kind Kind) {
190 switch (Kind) {
191 case object::Archive::K_GNU:
192 case object::Archive::K_GNU64:
193 case object::Archive::K_AIXBIG:
194 case object::Archive::K_COFF:
195 return false;
196 case object::Archive::K_BSD:
197 case object::Archive::K_DARWIN:
198 case object::Archive::K_DARWIN64:
199 return true;
200 }
201 llvm_unreachable("not supported for writting");
202}
203
204template <class T>
205static void print(raw_ostream &Out, object::Archive::Kind Kind, T Val) {
206 support::endian::write(Out, Val,
207 isBSDLike(Kind) ? llvm::endianness::little
208 : llvm::endianness::big);
209}
210
211template <class T> static void printLE(raw_ostream &Out, T Val) {
212 support::endian::write(Out, Val, llvm::endianness::little);
213}
214
215static void printRestOfMemberHeader(
216 raw_ostream &Out, const sys::TimePoint<std::chrono::seconds> &ModTime,
217 unsigned UID, unsigned GID, unsigned Perms, uint64_t Size) {
218 printWithSpacePadding(OS&: Out, Data: sys::toTimeT(TP: ModTime), Size: 12);
219
220 // The format has only 6 chars for uid and gid. Truncate if the provided
221 // values don't fit.
222 printWithSpacePadding(OS&: Out, Data: UID % 1000000, Size: 6);
223 printWithSpacePadding(OS&: Out, Data: GID % 1000000, Size: 6);
224
225 printWithSpacePadding(OS&: Out, Data: format(Fmt: "%o", Vals: Perms), Size: 8);
226 printWithSpacePadding(OS&: Out, Data: Size, Size: 10);
227 Out << "`\n";
228}
229
230static void
231printGNUSmallMemberHeader(raw_ostream &Out, StringRef Name,
232 const sys::TimePoint<std::chrono::seconds> &ModTime,
233 unsigned UID, unsigned GID, unsigned Perms,
234 uint64_t Size) {
235 printWithSpacePadding(OS&: Out, Data: Twine(Name) + "/", Size: 16);
236 printRestOfMemberHeader(Out, ModTime, UID, GID, Perms, Size);
237}
238
239static void
240printBSDMemberHeader(raw_ostream &Out, uint64_t Pos, StringRef Name,
241 const sys::TimePoint<std::chrono::seconds> &ModTime,
242 unsigned UID, unsigned GID, unsigned Perms, uint64_t Size) {
243 uint64_t PosAfterHeader = Pos + 60 + Name.size();
244 // Pad so that even 64 bit object files are aligned.
245 unsigned Pad = offsetToAlignment(Value: PosAfterHeader, Alignment: Align(8));
246 unsigned NameWithPadding = Name.size() + Pad;
247 printWithSpacePadding(OS&: Out, Data: Twine("#1/") + Twine(NameWithPadding), Size: 16);
248 printRestOfMemberHeader(Out, ModTime, UID, GID, Perms,
249 Size: NameWithPadding + Size);
250 Out << Name;
251 while (Pad--)
252 Out.write(C: uint8_t(0));
253}
254
255static void
256printBigArchiveMemberHeader(raw_ostream &Out, StringRef Name,
257 const sys::TimePoint<std::chrono::seconds> &ModTime,
258 unsigned UID, unsigned GID, unsigned Perms,
259 uint64_t Size, uint64_t PrevOffset,
260 uint64_t NextOffset) {
261 unsigned NameLen = Name.size();
262
263 printWithSpacePadding(OS&: Out, Data: Size, Size: 20); // File member size
264 printWithSpacePadding(OS&: Out, Data: NextOffset, Size: 20); // Next member header offset
265 printWithSpacePadding(OS&: Out, Data: PrevOffset, Size: 20); // Previous member header offset
266 printWithSpacePadding(OS&: Out, Data: sys::toTimeT(TP: ModTime), Size: 12); // File member date
267 // The big archive format has 12 chars for uid and gid.
268 printWithSpacePadding(OS&: Out, Data: UID % 1000000000000, Size: 12); // UID
269 printWithSpacePadding(OS&: Out, Data: GID % 1000000000000, Size: 12); // GID
270 printWithSpacePadding(OS&: Out, Data: format(Fmt: "%o", Vals: Perms), Size: 12); // Permission
271 printWithSpacePadding(OS&: Out, Data: NameLen, Size: 4); // Name length
272 if (NameLen) {
273 printWithSpacePadding(OS&: Out, Data: Name, Size: NameLen); // Name
274 if (NameLen % 2)
275 Out.write(C: uint8_t(0)); // Null byte padding
276 }
277 Out << "`\n"; // Terminator
278}
279
280static bool useStringTable(bool Thin, StringRef Name) {
281 return Thin || Name.size() >= 16 || Name.contains(C: '/');
282}
283
284static bool is64BitKind(object::Archive::Kind Kind) {
285 switch (Kind) {
286 case object::Archive::K_GNU:
287 case object::Archive::K_BSD:
288 case object::Archive::K_DARWIN:
289 case object::Archive::K_COFF:
290 return false;
291 case object::Archive::K_AIXBIG:
292 case object::Archive::K_DARWIN64:
293 case object::Archive::K_GNU64:
294 return true;
295 }
296 llvm_unreachable("not supported for writting");
297}
298
299static void
300printMemberHeader(raw_ostream &Out, uint64_t Pos, raw_ostream &StringTable,
301 StringMap<uint64_t> &MemberNames, object::Archive::Kind Kind,
302 bool Thin, const NewArchiveMember &M,
303 sys::TimePoint<std::chrono::seconds> ModTime, uint64_t Size) {
304 if (isBSDLike(Kind))
305 return printBSDMemberHeader(Out, Pos, Name: M.MemberName, ModTime, UID: M.UID, GID: M.GID,
306 Perms: M.Perms, Size);
307 if (!useStringTable(Thin, Name: M.MemberName))
308 return printGNUSmallMemberHeader(Out, Name: M.MemberName, ModTime, UID: M.UID, GID: M.GID,
309 Perms: M.Perms, Size);
310 Out << '/';
311 uint64_t NamePos;
312 if (Thin) {
313 NamePos = StringTable.tell();
314 StringTable << M.MemberName << "/\n";
315 } else {
316 auto Insertion = MemberNames.insert(KV: {M.MemberName, uint64_t(0)});
317 if (Insertion.second) {
318 Insertion.first->second = StringTable.tell();
319 StringTable << M.MemberName;
320 if (isCOFFArchive(Kind))
321 StringTable << '\0';
322 else
323 StringTable << "/\n";
324 }
325 NamePos = Insertion.first->second;
326 }
327 printWithSpacePadding(OS&: Out, Data: NamePos, Size: 15);
328 printRestOfMemberHeader(Out, ModTime, UID: M.UID, GID: M.GID, Perms: M.Perms, Size);
329}
330
331namespace {
332struct MemberData {
333 std::vector<unsigned> Symbols;
334 std::string Header;
335 StringRef Data;
336 StringRef Padding;
337 uint64_t PreHeadPadSize = 0;
338 std::unique_ptr<SymbolicFile> SymFile = nullptr;
339};
340} // namespace
341
342static MemberData computeStringTable(StringRef Names) {
343 unsigned Size = Names.size();
344 unsigned Pad = offsetToAlignment(Value: Size, Alignment: Align(2));
345 std::string Header;
346 raw_string_ostream Out(Header);
347 printWithSpacePadding(OS&: Out, Data: "//", Size: 48);
348 printWithSpacePadding(OS&: Out, Data: Size + Pad, Size: 10);
349 Out << "`\n";
350 Out.flush();
351 return {.Symbols: {}, .Header: std::move(Header), .Data: Names, .Padding: Pad ? "\n" : ""};
352}
353
354static sys::TimePoint<std::chrono::seconds> now(bool Deterministic) {
355 using namespace std::chrono;
356
357 if (!Deterministic)
358 return time_point_cast<seconds>(t: system_clock::now());
359 return sys::TimePoint<seconds>();
360}
361
362static bool isArchiveSymbol(const object::BasicSymbolRef &S) {
363 Expected<uint32_t> SymFlagsOrErr = S.getFlags();
364 if (!SymFlagsOrErr)
365 // TODO: Actually report errors helpfully.
366 report_fatal_error(Err: SymFlagsOrErr.takeError());
367 if (*SymFlagsOrErr & object::SymbolRef::SF_FormatSpecific)
368 return false;
369 if (!(*SymFlagsOrErr & object::SymbolRef::SF_Global))
370 return false;
371 if (*SymFlagsOrErr & object::SymbolRef::SF_Undefined)
372 return false;
373 return true;
374}
375
376static void printNBits(raw_ostream &Out, object::Archive::Kind Kind,
377 uint64_t Val) {
378 if (is64BitKind(Kind))
379 print<uint64_t>(Out, Kind, Val);
380 else
381 print<uint32_t>(Out, Kind, Val);
382}
383
384static uint64_t computeSymbolTableSize(object::Archive::Kind Kind,
385 uint64_t NumSyms, uint64_t OffsetSize,
386 uint64_t StringTableSize,
387 uint32_t *Padding = nullptr) {
388 assert((OffsetSize == 4 || OffsetSize == 8) && "Unsupported OffsetSize");
389 uint64_t Size = OffsetSize; // Number of entries
390 if (isBSDLike(Kind))
391 Size += NumSyms * OffsetSize * 2; // Table
392 else
393 Size += NumSyms * OffsetSize; // Table
394 if (isBSDLike(Kind))
395 Size += OffsetSize; // byte count
396 Size += StringTableSize;
397 // ld64 expects the members to be 8-byte aligned for 64-bit content and at
398 // least 4-byte aligned for 32-bit content. Opt for the larger encoding
399 // uniformly.
400 // We do this for all bsd formats because it simplifies aligning members.
401 // For the big archive format, the symbol table is the last member, so there
402 // is no need to align.
403 uint32_t Pad = isAIXBigArchive(Kind)
404 ? 0
405 : offsetToAlignment(Value: Size, Alignment: Align(isBSDLike(Kind) ? 8 : 2));
406
407 Size += Pad;
408 if (Padding)
409 *Padding = Pad;
410 return Size;
411}
412
413static uint64_t computeSymbolMapSize(uint64_t NumObj, SymMap &SymMap,
414 uint32_t *Padding = nullptr) {
415 uint64_t Size = sizeof(uint32_t) * 2; // Number of symbols and objects entries
416 Size += NumObj * sizeof(uint32_t); // Offset table
417
418 for (auto S : SymMap.Map)
419 Size += sizeof(uint16_t) + S.first.length() + 1;
420
421 uint32_t Pad = offsetToAlignment(Value: Size, Alignment: Align(2));
422 Size += Pad;
423 if (Padding)
424 *Padding = Pad;
425 return Size;
426}
427
428static uint64_t computeECSymbolsSize(SymMap &SymMap,
429 uint32_t *Padding = nullptr) {
430 uint64_t Size = sizeof(uint32_t); // Number of symbols
431
432 for (auto S : SymMap.ECMap)
433 Size += sizeof(uint16_t) + S.first.length() + 1;
434
435 uint32_t Pad = offsetToAlignment(Value: Size, Alignment: Align(2));
436 Size += Pad;
437 if (Padding)
438 *Padding = Pad;
439 return Size;
440}
441
442static void writeSymbolTableHeader(raw_ostream &Out, object::Archive::Kind Kind,
443 bool Deterministic, uint64_t Size,
444 uint64_t PrevMemberOffset = 0,
445 uint64_t NextMemberOffset = 0) {
446 if (isBSDLike(Kind)) {
447 const char *Name = is64BitKind(Kind) ? "__.SYMDEF_64" : "__.SYMDEF";
448 printBSDMemberHeader(Out, Pos: Out.tell(), Name, ModTime: now(Deterministic), UID: 0, GID: 0, Perms: 0,
449 Size);
450 } else if (isAIXBigArchive(Kind)) {
451 printBigArchiveMemberHeader(Out, Name: "", ModTime: now(Deterministic), UID: 0, GID: 0, Perms: 0, Size,
452 PrevOffset: PrevMemberOffset, NextOffset: NextMemberOffset);
453 } else {
454 const char *Name = is64BitKind(Kind) ? "/SYM64" : "";
455 printGNUSmallMemberHeader(Out, Name, ModTime: now(Deterministic), UID: 0, GID: 0, Perms: 0, Size);
456 }
457}
458
459static uint64_t computeHeadersSize(object::Archive::Kind Kind,
460 uint64_t NumMembers,
461 uint64_t StringMemberSize, uint64_t NumSyms,
462 uint64_t SymNamesSize, SymMap *SymMap) {
463 uint32_t OffsetSize = is64BitKind(Kind) ? 8 : 4;
464 uint64_t SymtabSize =
465 computeSymbolTableSize(Kind, NumSyms, OffsetSize, StringTableSize: SymNamesSize);
466 auto computeSymbolTableHeaderSize = [=] {
467 SmallString<0> TmpBuf;
468 raw_svector_ostream Tmp(TmpBuf);
469 writeSymbolTableHeader(Out&: Tmp, Kind, Deterministic: true, Size: SymtabSize);
470 return TmpBuf.size();
471 };
472 uint32_t HeaderSize = computeSymbolTableHeaderSize();
473 uint64_t Size = strlen(s: "!<arch>\n") + HeaderSize + SymtabSize;
474
475 if (SymMap) {
476 Size += HeaderSize + computeSymbolMapSize(NumObj: NumMembers, SymMap&: *SymMap);
477 if (SymMap->ECMap.size())
478 Size += HeaderSize + computeECSymbolsSize(SymMap&: *SymMap);
479 }
480
481 return Size + StringMemberSize;
482}
483
484static Expected<std::unique_ptr<SymbolicFile>>
485getSymbolicFile(MemoryBufferRef Buf, LLVMContext &Context,
486 object::Archive::Kind Kind, function_ref<void(Error)> Warn) {
487 const file_magic Type = identify_magic(magic: Buf.getBuffer());
488 // Don't attempt to read non-symbolic file types.
489 if (!object::SymbolicFile::isSymbolicFile(Type, Context: &Context))
490 return nullptr;
491 if (Type == file_magic::bitcode) {
492 auto ObjOrErr = object::SymbolicFile::createSymbolicFile(
493 Object: Buf, Type: file_magic::bitcode, Context: &Context);
494 // An error reading a bitcode file most likely indicates that the file
495 // was created by a compiler from the future. Normally we don't try to
496 // implement forwards compatibility for bitcode files, but when creating an
497 // archive we can implement best-effort forwards compatibility by treating
498 // the file as a blob and not creating symbol index entries for it. lld and
499 // mold ignore the archive symbol index, so provided that you use one of
500 // these linkers, LTO will work as long as lld or the gold plugin is newer
501 // than the compiler. We only ignore errors if the archive format is one
502 // that is supported by a linker that is known to ignore the index,
503 // otherwise there's no chance of this working so we may as well error out.
504 // We print a warning on read failure so that users of linkers that rely on
505 // the symbol index can diagnose the issue.
506 //
507 // This is the same behavior as GNU ar when the linker plugin returns an
508 // error when reading the input file. If the bitcode file is actually
509 // malformed, it will be diagnosed at link time.
510 if (!ObjOrErr) {
511 switch (Kind) {
512 case object::Archive::K_BSD:
513 case object::Archive::K_GNU:
514 case object::Archive::K_GNU64:
515 Warn(ObjOrErr.takeError());
516 return nullptr;
517 case object::Archive::K_AIXBIG:
518 case object::Archive::K_COFF:
519 case object::Archive::K_DARWIN:
520 case object::Archive::K_DARWIN64:
521 return ObjOrErr.takeError();
522 }
523 }
524 return std::move(*ObjOrErr);
525 } else {
526 auto ObjOrErr = object::SymbolicFile::createSymbolicFile(Object: Buf);
527 if (!ObjOrErr)
528 return ObjOrErr.takeError();
529 return std::move(*ObjOrErr);
530 }
531}
532
533static bool is64BitSymbolicFile(const SymbolicFile *SymObj) {
534 return SymObj != nullptr ? SymObj->is64Bit() : false;
535}
536
537// Log2 of PAGESIZE(4096) on an AIX system.
538static const uint32_t Log2OfAIXPageSize = 12;
539
540// In the AIX big archive format, since the data content follows the member file
541// name, if the name ends on an odd byte, an extra byte will be added for
542// padding. This ensures that the data within the member file starts at an even
543// byte.
544static const uint32_t MinBigArchiveMemDataAlign = 2;
545
546template <typename AuxiliaryHeader>
547uint16_t getAuxMaxAlignment(uint16_t AuxHeaderSize, AuxiliaryHeader *AuxHeader,
548 uint16_t Log2OfMaxAlign) {
549 // If the member doesn't have an auxiliary header, it isn't a loadable object
550 // and so it just needs aligning at the minimum value.
551 if (AuxHeader == nullptr)
552 return MinBigArchiveMemDataAlign;
553
554 // If the auxiliary header does not have both MaxAlignOfData and
555 // MaxAlignOfText field, it is not a loadable shared object file, so align at
556 // the minimum value. The 'ModuleType' member is located right after
557 // 'MaxAlignOfData' in the AuxiliaryHeader.
558 if (AuxHeaderSize < offsetof(AuxiliaryHeader, ModuleType))
559 return MinBigArchiveMemDataAlign;
560
561 // If the XCOFF object file does not have a loader section, it is not
562 // loadable, so align at the minimum value.
563 if (AuxHeader->SecNumOfLoader == 0)
564 return MinBigArchiveMemDataAlign;
565
566 // The content of the loadable member file needs to be aligned at MAX(maximum
567 // alignment of .text, maximum alignment of .data) if there are both fields.
568 // If the desired alignment is > PAGESIZE, 32-bit members are aligned on a
569 // word boundary, while 64-bit members are aligned on a PAGESIZE(2^12=4096)
570 // boundary.
571 uint16_t Log2OfAlign =
572 std::max(AuxHeader->MaxAlignOfText, AuxHeader->MaxAlignOfData);
573 return 1 << (Log2OfAlign > Log2OfAIXPageSize ? Log2OfMaxAlign : Log2OfAlign);
574}
575
576// AIX big archives may contain shared object members. The AIX OS requires these
577// members to be aligned if they are 64-bit and recommends it for 32-bit
578// members. This ensures that when these members are loaded they are aligned in
579// memory.
580static uint32_t getMemberAlignment(SymbolicFile *SymObj) {
581 XCOFFObjectFile *XCOFFObj = dyn_cast_or_null<XCOFFObjectFile>(Val: SymObj);
582 if (!XCOFFObj)
583 return MinBigArchiveMemDataAlign;
584
585 // If the desired alignment is > PAGESIZE, 32-bit members are aligned on a
586 // word boundary, while 64-bit members are aligned on a PAGESIZE boundary.
587 return XCOFFObj->is64Bit()
588 ? getAuxMaxAlignment(AuxHeaderSize: XCOFFObj->fileHeader64()->AuxHeaderSize,
589 AuxHeader: XCOFFObj->auxiliaryHeader64(),
590 Log2OfMaxAlign: Log2OfAIXPageSize)
591 : getAuxMaxAlignment(AuxHeaderSize: XCOFFObj->fileHeader32()->AuxHeaderSize,
592 AuxHeader: XCOFFObj->auxiliaryHeader32(), Log2OfMaxAlign: 2);
593}
594
595static void writeSymbolTable(raw_ostream &Out, object::Archive::Kind Kind,
596 bool Deterministic, ArrayRef<MemberData> Members,
597 StringRef StringTable, uint64_t MembersOffset,
598 unsigned NumSyms, uint64_t PrevMemberOffset = 0,
599 uint64_t NextMemberOffset = 0,
600 bool Is64Bit = false) {
601 // We don't write a symbol table on an archive with no members -- except on
602 // Darwin, where the linker will abort unless the archive has a symbol table.
603 if (StringTable.empty() && !isDarwin(Kind) && !isCOFFArchive(Kind))
604 return;
605
606 uint64_t OffsetSize = is64BitKind(Kind) ? 8 : 4;
607 uint32_t Pad;
608 uint64_t Size = computeSymbolTableSize(Kind, NumSyms, OffsetSize,
609 StringTableSize: StringTable.size(), Padding: &Pad);
610 writeSymbolTableHeader(Out, Kind, Deterministic, Size, PrevMemberOffset,
611 NextMemberOffset);
612
613 if (isBSDLike(Kind))
614 printNBits(Out, Kind, Val: NumSyms * 2 * OffsetSize);
615 else
616 printNBits(Out, Kind, Val: NumSyms);
617
618 uint64_t Pos = MembersOffset;
619 for (const MemberData &M : Members) {
620 if (isAIXBigArchive(Kind)) {
621 Pos += M.PreHeadPadSize;
622 if (is64BitSymbolicFile(SymObj: M.SymFile.get()) != Is64Bit) {
623 Pos += M.Header.size() + M.Data.size() + M.Padding.size();
624 continue;
625 }
626 }
627
628 for (unsigned StringOffset : M.Symbols) {
629 if (isBSDLike(Kind))
630 printNBits(Out, Kind, Val: StringOffset);
631 printNBits(Out, Kind, Val: Pos); // member offset
632 }
633 Pos += M.Header.size() + M.Data.size() + M.Padding.size();
634 }
635
636 if (isBSDLike(Kind))
637 // byte count of the string table
638 printNBits(Out, Kind, Val: StringTable.size());
639 Out << StringTable;
640
641 while (Pad--)
642 Out.write(C: uint8_t(0));
643}
644
645static void writeSymbolMap(raw_ostream &Out, object::Archive::Kind Kind,
646 bool Deterministic, ArrayRef<MemberData> Members,
647 SymMap &SymMap, uint64_t MembersOffset) {
648 uint32_t Pad;
649 uint64_t Size = computeSymbolMapSize(NumObj: Members.size(), SymMap, Padding: &Pad);
650 writeSymbolTableHeader(Out, Kind, Deterministic, Size, PrevMemberOffset: 0);
651
652 uint32_t Pos = MembersOffset;
653
654 printLE<uint32_t>(Out, Val: Members.size());
655 for (const MemberData &M : Members) {
656 printLE(Out, Val: Pos); // member offset
657 Pos += M.Header.size() + M.Data.size() + M.Padding.size();
658 }
659
660 printLE<uint32_t>(Out, Val: SymMap.Map.size());
661
662 for (auto S : SymMap.Map)
663 printLE(Out, Val: S.second);
664 for (auto S : SymMap.Map)
665 Out << S.first << '\0';
666
667 while (Pad--)
668 Out.write(C: uint8_t(0));
669}
670
671static void writeECSymbols(raw_ostream &Out, object::Archive::Kind Kind,
672 bool Deterministic, ArrayRef<MemberData> Members,
673 SymMap &SymMap) {
674 uint32_t Pad;
675 uint64_t Size = computeECSymbolsSize(SymMap, Padding: &Pad);
676 printGNUSmallMemberHeader(Out, Name: "/<ECSYMBOLS>", ModTime: now(Deterministic), UID: 0, GID: 0, Perms: 0,
677 Size);
678
679 printLE<uint32_t>(Out, Val: SymMap.ECMap.size());
680
681 for (auto S : SymMap.ECMap)
682 printLE(Out, Val: S.second);
683 for (auto S : SymMap.ECMap)
684 Out << S.first << '\0';
685 while (Pad--)
686 Out.write(C: uint8_t(0));
687}
688
689static bool isECObject(object::SymbolicFile &Obj) {
690 if (Obj.isCOFF())
691 return cast<llvm::object::COFFObjectFile>(Val: &Obj)->getMachine() !=
692 COFF::IMAGE_FILE_MACHINE_ARM64;
693
694 if (Obj.isCOFFImportFile())
695 return cast<llvm::object::COFFImportFile>(Val: &Obj)->getMachine() !=
696 COFF::IMAGE_FILE_MACHINE_ARM64;
697
698 if (Obj.isIR()) {
699 Expected<std::string> TripleStr =
700 getBitcodeTargetTriple(Buffer: Obj.getMemoryBufferRef());
701 if (!TripleStr)
702 return false;
703 Triple T(std::move(*TripleStr));
704 return T.isWindowsArm64EC() || T.getArch() == Triple::x86_64;
705 }
706
707 return false;
708}
709
710static bool isAnyArm64COFF(object::SymbolicFile &Obj) {
711 if (Obj.isCOFF())
712 return COFF::isAnyArm64(Machine: cast<COFFObjectFile>(Val: &Obj)->getMachine());
713
714 if (Obj.isCOFFImportFile())
715 return COFF::isAnyArm64(Machine: cast<COFFImportFile>(Val: &Obj)->getMachine());
716
717 if (Obj.isIR()) {
718 Expected<std::string> TripleStr =
719 getBitcodeTargetTriple(Buffer: Obj.getMemoryBufferRef());
720 if (!TripleStr)
721 return false;
722 Triple T(std::move(*TripleStr));
723 return T.isOSWindows() && T.getArch() == Triple::aarch64;
724 }
725
726 return false;
727}
728
729bool isImportDescriptor(StringRef Name) {
730 return Name.starts_with(Prefix: ImportDescriptorPrefix) ||
731 Name == StringRef{NullImportDescriptorSymbolName} ||
732 (Name.starts_with(Prefix: NullThunkDataPrefix) &&
733 Name.ends_with(Suffix: NullThunkDataSuffix));
734}
735
736static Expected<std::vector<unsigned>> getSymbols(SymbolicFile *Obj,
737 uint16_t Index,
738 raw_ostream &SymNames,
739 SymMap *SymMap) {
740 std::vector<unsigned> Ret;
741
742 if (Obj == nullptr)
743 return Ret;
744
745 std::map<std::string, uint16_t> *Map = nullptr;
746 if (SymMap)
747 Map = SymMap->UseECMap && isECObject(Obj&: *Obj) ? &SymMap->ECMap : &SymMap->Map;
748
749 for (const object::BasicSymbolRef &S : Obj->symbols()) {
750 if (!isArchiveSymbol(S))
751 continue;
752 if (Map) {
753 std::string Name;
754 raw_string_ostream NameStream(Name);
755 if (Error E = S.printName(OS&: NameStream))
756 return std::move(E);
757 if (!Map->try_emplace(k: Name, args&: Index).second)
758 continue; // ignore duplicated symbol
759 if (Map == &SymMap->Map) {
760 Ret.push_back(x: SymNames.tell());
761 SymNames << Name << '\0';
762 // If EC is enabled, then the import descriptors are NOT put into EC
763 // objects so we need to copy them to the EC map manually.
764 if (SymMap->UseECMap && isImportDescriptor(Name))
765 SymMap->ECMap[Name] = Index;
766 }
767 } else {
768 Ret.push_back(x: SymNames.tell());
769 if (Error E = S.printName(OS&: SymNames))
770 return std::move(E);
771 SymNames << '\0';
772 }
773 }
774 return Ret;
775}
776
777static Expected<std::vector<MemberData>>
778computeMemberData(raw_ostream &StringTable, raw_ostream &SymNames,
779 object::Archive::Kind Kind, bool Thin, bool Deterministic,
780 SymtabWritingMode NeedSymbols, SymMap *SymMap,
781 LLVMContext &Context, ArrayRef<NewArchiveMember> NewMembers,
782 std::optional<bool> IsEC, function_ref<void(Error)> Warn) {
783 static char PaddingData[8] = {'\n', '\n', '\n', '\n', '\n', '\n', '\n', '\n'};
784 uint64_t MemHeadPadSize = 0;
785 uint64_t Pos =
786 isAIXBigArchive(Kind) ? sizeof(object::BigArchive::FixLenHdr) : 0;
787
788 std::vector<MemberData> Ret;
789 bool HasObject = false;
790
791 // Deduplicate long member names in the string table and reuse earlier name
792 // offsets. This especially saves space for COFF Import libraries where all
793 // members have the same name.
794 StringMap<uint64_t> MemberNames;
795
796 // UniqueTimestamps is a special case to improve debugging on Darwin:
797 //
798 // The Darwin linker does not link debug info into the final
799 // binary. Instead, it emits entries of type N_OSO in the output
800 // binary's symbol table, containing references to the linked-in
801 // object files. Using that reference, the debugger can read the
802 // debug data directly from the object files. Alternatively, an
803 // invocation of 'dsymutil' will link the debug data from the object
804 // files into a dSYM bundle, which can be loaded by the debugger,
805 // instead of the object files.
806 //
807 // For an object file, the N_OSO entries contain the absolute path
808 // path to the file, and the file's timestamp. For an object
809 // included in an archive, the path is formatted like
810 // "/absolute/path/to/archive.a(member.o)", and the timestamp is the
811 // archive member's timestamp, rather than the archive's timestamp.
812 //
813 // However, this doesn't always uniquely identify an object within
814 // an archive -- an archive file can have multiple entries with the
815 // same filename. (This will happen commonly if the original object
816 // files started in different directories.) The only way they get
817 // distinguished, then, is via the timestamp. But this process is
818 // unable to find the correct object file in the archive when there
819 // are two files of the same name and timestamp.
820 //
821 // Additionally, timestamp==0 is treated specially, and causes the
822 // timestamp to be ignored as a match criteria.
823 //
824 // That will "usually" work out okay when creating an archive not in
825 // deterministic timestamp mode, because the objects will probably
826 // have been created at different timestamps.
827 //
828 // To ameliorate this problem, in deterministic archive mode (which
829 // is the default), on Darwin we will emit a unique non-zero
830 // timestamp for each entry with a duplicated name. This is still
831 // deterministic: the only thing affecting that timestamp is the
832 // order of the files in the resultant archive.
833 //
834 // See also the functions that handle the lookup:
835 // in lldb: ObjectContainerBSDArchive::Archive::FindObject()
836 // in llvm/tools/dsymutil: BinaryHolder::GetArchiveMemberBuffers().
837 bool UniqueTimestamps = Deterministic && isDarwin(Kind);
838 std::map<StringRef, unsigned> FilenameCount;
839 if (UniqueTimestamps) {
840 for (const NewArchiveMember &M : NewMembers)
841 FilenameCount[M.MemberName]++;
842 for (auto &Entry : FilenameCount)
843 Entry.second = Entry.second > 1 ? 1 : 0;
844 }
845
846 std::vector<std::unique_ptr<SymbolicFile>> SymFiles;
847
848 if (NeedSymbols != SymtabWritingMode::NoSymtab || isAIXBigArchive(Kind)) {
849 for (const NewArchiveMember &M : NewMembers) {
850 Expected<std::unique_ptr<SymbolicFile>> SymFileOrErr = getSymbolicFile(
851 Buf: M.Buf->getMemBufferRef(), Context, Kind, Warn: [&](Error Err) {
852 Warn(createFileError(F: M.MemberName, E: std::move(Err)));
853 });
854 if (!SymFileOrErr)
855 return createFileError(F: M.MemberName, E: SymFileOrErr.takeError());
856 SymFiles.push_back(x: std::move(*SymFileOrErr));
857 }
858 }
859
860 if (SymMap) {
861 if (IsEC) {
862 SymMap->UseECMap = *IsEC;
863 } else {
864 // When IsEC is not specified by the caller, use it when we have both
865 // any ARM64 object (ARM64 or ARM64EC) and any EC object (ARM64EC or
866 // AMD64). This may be a single ARM64EC object, but may also be separate
867 // ARM64 and AMD64 objects.
868 bool HaveArm64 = false, HaveEC = false;
869 for (std::unique_ptr<SymbolicFile> &SymFile : SymFiles) {
870 if (!SymFile)
871 continue;
872 if (!HaveArm64)
873 HaveArm64 = isAnyArm64COFF(Obj&: *SymFile);
874 if (!HaveEC)
875 HaveEC = isECObject(Obj&: *SymFile);
876 if (HaveArm64 && HaveEC) {
877 SymMap->UseECMap = true;
878 break;
879 }
880 }
881 }
882 }
883
884 // The big archive format needs to know the offset of the previous member
885 // header.
886 uint64_t PrevOffset = 0;
887 uint64_t NextMemHeadPadSize = 0;
888
889 for (uint32_t Index = 0; Index < NewMembers.size(); ++Index) {
890 const NewArchiveMember *M = &NewMembers[Index];
891 std::string Header;
892 raw_string_ostream Out(Header);
893
894 MemoryBufferRef Buf = M->Buf->getMemBufferRef();
895 StringRef Data = Thin ? "" : Buf.getBuffer();
896
897 // ld64 expects the members to be 8-byte aligned for 64-bit content and at
898 // least 4-byte aligned for 32-bit content. Opt for the larger encoding
899 // uniformly. This matches the behaviour with cctools and ensures that ld64
900 // is happy with archives that we generate.
901 unsigned MemberPadding =
902 isDarwin(Kind) ? offsetToAlignment(Value: Data.size(), Alignment: Align(8)) : 0;
903 unsigned TailPadding =
904 offsetToAlignment(Value: Data.size() + MemberPadding, Alignment: Align(2));
905 StringRef Padding = StringRef(PaddingData, MemberPadding + TailPadding);
906
907 sys::TimePoint<std::chrono::seconds> ModTime;
908 if (UniqueTimestamps)
909 // Increment timestamp for each file of a given name.
910 ModTime = sys::toTimePoint(T: FilenameCount[M->MemberName]++);
911 else
912 ModTime = M->ModTime;
913
914 uint64_t Size = Buf.getBufferSize() + MemberPadding;
915 if (Size > object::Archive::MaxMemberSize) {
916 std::string StringMsg =
917 "File " + M->MemberName.str() + " exceeds size limit";
918 return make_error<object::GenericBinaryError>(
919 Args: std::move(StringMsg), Args: object::object_error::parse_failed);
920 }
921
922 std::unique_ptr<SymbolicFile> CurSymFile;
923 if (!SymFiles.empty())
924 CurSymFile = std::move(SymFiles[Index]);
925
926 // In the big archive file format, we need to calculate and include the next
927 // member offset and previous member offset in the file member header.
928 if (isAIXBigArchive(Kind)) {
929 uint64_t OffsetToMemData = Pos + sizeof(object::BigArMemHdrType) +
930 alignTo(Value: M->MemberName.size(), Align: 2);
931
932 if (M == NewMembers.begin())
933 NextMemHeadPadSize =
934 alignToPowerOf2(Value: OffsetToMemData,
935 Align: getMemberAlignment(SymObj: CurSymFile.get())) -
936 OffsetToMemData;
937
938 MemHeadPadSize = NextMemHeadPadSize;
939 Pos += MemHeadPadSize;
940 uint64_t NextOffset = Pos + sizeof(object::BigArMemHdrType) +
941 alignTo(Value: M->MemberName.size(), Align: 2) + alignTo(Value: Size, Align: 2);
942
943 // If there is another member file after this, we need to calculate the
944 // padding before the header.
945 if (Index + 1 != SymFiles.size()) {
946 uint64_t OffsetToNextMemData =
947 NextOffset + sizeof(object::BigArMemHdrType) +
948 alignTo(Value: NewMembers[Index + 1].MemberName.size(), Align: 2);
949 NextMemHeadPadSize =
950 alignToPowerOf2(Value: OffsetToNextMemData,
951 Align: getMemberAlignment(SymObj: SymFiles[Index + 1].get())) -
952 OffsetToNextMemData;
953 NextOffset += NextMemHeadPadSize;
954 }
955 printBigArchiveMemberHeader(Out, Name: M->MemberName, ModTime, UID: M->UID, GID: M->GID,
956 Perms: M->Perms, Size, PrevOffset, NextOffset);
957 PrevOffset = Pos;
958 } else {
959 printMemberHeader(Out, Pos, StringTable, MemberNames, Kind, Thin, M: *M,
960 ModTime, Size);
961 }
962 Out.flush();
963
964 std::vector<unsigned> Symbols;
965 if (NeedSymbols != SymtabWritingMode::NoSymtab) {
966 Expected<std::vector<unsigned>> SymbolsOrErr =
967 getSymbols(Obj: CurSymFile.get(), Index: Index + 1, SymNames, SymMap);
968 if (!SymbolsOrErr)
969 return createFileError(F: M->MemberName, E: SymbolsOrErr.takeError());
970 Symbols = std::move(*SymbolsOrErr);
971 if (CurSymFile)
972 HasObject = true;
973 }
974
975 Pos += Header.size() + Data.size() + Padding.size();
976 Ret.push_back(x: {.Symbols: std::move(Symbols), .Header: std::move(Header), .Data: Data, .Padding: Padding,
977 .PreHeadPadSize: MemHeadPadSize, .SymFile: std::move(CurSymFile)});
978 }
979 // If there are no symbols, emit an empty symbol table, to satisfy Solaris
980 // tools, older versions of which expect a symbol table in a non-empty
981 // archive, regardless of whether there are any symbols in it.
982 if (HasObject && SymNames.tell() == 0 && !isCOFFArchive(Kind))
983 SymNames << '\0' << '\0' << '\0';
984 return std::move(Ret);
985}
986
987namespace llvm {
988
989static ErrorOr<SmallString<128>> canonicalizePath(StringRef P) {
990 SmallString<128> Ret = P;
991 std::error_code Err = sys::fs::make_absolute(path&: Ret);
992 if (Err)
993 return Err;
994 sys::path::remove_dots(path&: Ret, /*removedotdot*/ remove_dot_dot: true);
995 return Ret;
996}
997
998// Compute the relative path from From to To.
999Expected<std::string> computeArchiveRelativePath(StringRef From, StringRef To) {
1000 ErrorOr<SmallString<128>> PathToOrErr = canonicalizePath(P: To);
1001 ErrorOr<SmallString<128>> DirFromOrErr = canonicalizePath(P: From);
1002 if (!PathToOrErr || !DirFromOrErr)
1003 return errorCodeToError(EC: errnoAsErrorCode());
1004
1005 const SmallString<128> &PathTo = *PathToOrErr;
1006 const SmallString<128> &DirFrom = sys::path::parent_path(path: *DirFromOrErr);
1007
1008 // Can't construct a relative path between different roots
1009 if (sys::path::root_name(path: PathTo) != sys::path::root_name(path: DirFrom))
1010 return sys::path::convert_to_slash(path: PathTo);
1011
1012 // Skip common prefixes
1013 auto FromTo =
1014 std::mismatch(first1: sys::path::begin(path: DirFrom), last1: sys::path::end(path: DirFrom),
1015 first2: sys::path::begin(path: PathTo));
1016 auto FromI = FromTo.first;
1017 auto ToI = FromTo.second;
1018
1019 // Construct relative path
1020 SmallString<128> Relative;
1021 for (auto FromE = sys::path::end(path: DirFrom); FromI != FromE; ++FromI)
1022 sys::path::append(path&: Relative, style: sys::path::Style::posix, a: "..");
1023
1024 for (auto ToE = sys::path::end(path: PathTo); ToI != ToE; ++ToI)
1025 sys::path::append(path&: Relative, style: sys::path::Style::posix, a: *ToI);
1026
1027 return std::string(Relative);
1028}
1029
1030Error writeArchiveToStream(raw_ostream &Out,
1031 ArrayRef<NewArchiveMember> NewMembers,
1032 SymtabWritingMode WriteSymtab,
1033 object::Archive::Kind Kind, bool Deterministic,
1034 bool Thin, std::optional<bool> IsEC,
1035 function_ref<void(Error)> Warn) {
1036 assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode");
1037
1038 SmallString<0> SymNamesBuf;
1039 raw_svector_ostream SymNames(SymNamesBuf);
1040 SmallString<0> StringTableBuf;
1041 raw_svector_ostream StringTable(StringTableBuf);
1042 SymMap SymMap;
1043 bool ShouldWriteSymtab = WriteSymtab != SymtabWritingMode::NoSymtab;
1044
1045 // COFF symbol map uses 16-bit indexes, so we can't use it if there are too
1046 // many members. COFF format also requires symbol table presence, so use
1047 // GNU format when NoSymtab is requested.
1048 if (isCOFFArchive(Kind) && (NewMembers.size() > 0xfffe || !ShouldWriteSymtab))
1049 Kind = object::Archive::K_GNU;
1050
1051 // In the scenario when LLVMContext is populated SymbolicFile will contain a
1052 // reference to it, thus SymbolicFile should be destroyed first.
1053 LLVMContext Context;
1054
1055 Expected<std::vector<MemberData>> DataOrErr = computeMemberData(
1056 StringTable, SymNames, Kind, Thin, Deterministic, NeedSymbols: WriteSymtab,
1057 SymMap: isCOFFArchive(Kind) ? &SymMap : nullptr, Context, NewMembers, IsEC, Warn);
1058 if (Error E = DataOrErr.takeError())
1059 return E;
1060 std::vector<MemberData> &Data = *DataOrErr;
1061
1062 uint64_t StringTableSize = 0;
1063 MemberData StringTableMember;
1064 if (!StringTableBuf.empty() && !isAIXBigArchive(Kind)) {
1065 StringTableMember = computeStringTable(Names: StringTableBuf);
1066 StringTableSize = StringTableMember.Header.size() +
1067 StringTableMember.Data.size() +
1068 StringTableMember.Padding.size();
1069 }
1070
1071 // We would like to detect if we need to switch to a 64-bit symbol table.
1072 uint64_t LastMemberEndOffset = 0;
1073 uint64_t LastMemberHeaderOffset = 0;
1074 uint64_t NumSyms = 0;
1075 uint64_t NumSyms32 = 0; // Store symbol number of 32-bit member files.
1076
1077 for (const auto &M : Data) {
1078 // Record the start of the member's offset
1079 LastMemberEndOffset += M.PreHeadPadSize;
1080 LastMemberHeaderOffset = LastMemberEndOffset;
1081 // Account for the size of each part associated with the member.
1082 LastMemberEndOffset += M.Header.size() + M.Data.size() + M.Padding.size();
1083 NumSyms += M.Symbols.size();
1084
1085 // AIX big archive files may contain two global symbol tables. The
1086 // first global symbol table locates 32-bit file members that define global
1087 // symbols; the second global symbol table does the same for 64-bit file
1088 // members. As a big archive can have both 32-bit and 64-bit file members,
1089 // we need to know the number of symbols in each symbol table individually.
1090 if (isAIXBigArchive(Kind) && ShouldWriteSymtab) {
1091 if (!is64BitSymbolicFile(SymObj: M.SymFile.get()))
1092 NumSyms32 += M.Symbols.size();
1093 }
1094 }
1095
1096 std::optional<uint64_t> HeadersSize;
1097
1098 // The symbol table is put at the end of the big archive file. The symbol
1099 // table is at the start of the archive file for other archive formats.
1100 if (ShouldWriteSymtab && !is64BitKind(Kind)) {
1101 // We assume 32-bit offsets to see if 32-bit symbols are possible or not.
1102 HeadersSize = computeHeadersSize(Kind, NumMembers: Data.size(), StringMemberSize: StringTableSize,
1103 NumSyms, SymNamesSize: SymNamesBuf.size(),
1104 SymMap: isCOFFArchive(Kind) ? &SymMap : nullptr);
1105
1106 // The SYM64 format is used when an archive's member offsets are larger than
1107 // 32-bits can hold. The need for this shift in format is detected by
1108 // writeArchive. To test this we need to generate a file with a member that
1109 // has an offset larger than 32-bits but this demands a very slow test. To
1110 // speed the test up we use this environment variable to pretend like the
1111 // cutoff happens before 32-bits and instead happens at some much smaller
1112 // value.
1113 uint64_t Sym64Threshold = 1ULL << 32;
1114 const char *Sym64Env = std::getenv(name: "SYM64_THRESHOLD");
1115 if (Sym64Env)
1116 StringRef(Sym64Env).getAsInteger(Radix: 10, Result&: Sym64Threshold);
1117
1118 // If LastMemberHeaderOffset isn't going to fit in a 32-bit varible we need
1119 // to switch to 64-bit. Note that the file can be larger than 4GB as long as
1120 // the last member starts before the 4GB offset.
1121 if (*HeadersSize + LastMemberHeaderOffset >= Sym64Threshold) {
1122 if (Kind == object::Archive::K_DARWIN)
1123 Kind = object::Archive::K_DARWIN64;
1124 else
1125 Kind = object::Archive::K_GNU64;
1126 HeadersSize.reset();
1127 }
1128 }
1129
1130 if (Thin)
1131 Out << "!<thin>\n";
1132 else if (isAIXBigArchive(Kind))
1133 Out << "<bigaf>\n";
1134 else
1135 Out << "!<arch>\n";
1136
1137 if (!isAIXBigArchive(Kind)) {
1138 if (ShouldWriteSymtab) {
1139 if (!HeadersSize)
1140 HeadersSize = computeHeadersSize(
1141 Kind, NumMembers: Data.size(), StringMemberSize: StringTableSize, NumSyms, SymNamesSize: SymNamesBuf.size(),
1142 SymMap: isCOFFArchive(Kind) ? &SymMap : nullptr);
1143 writeSymbolTable(Out, Kind, Deterministic, Members: Data, StringTable: SymNamesBuf,
1144 MembersOffset: *HeadersSize, NumSyms);
1145
1146 if (isCOFFArchive(Kind))
1147 writeSymbolMap(Out, Kind, Deterministic, Members: Data, SymMap, MembersOffset: *HeadersSize);
1148 }
1149
1150 if (StringTableSize)
1151 Out << StringTableMember.Header << StringTableMember.Data
1152 << StringTableMember.Padding;
1153
1154 if (ShouldWriteSymtab && SymMap.ECMap.size())
1155 writeECSymbols(Out, Kind, Deterministic, Members: Data, SymMap);
1156
1157 for (const MemberData &M : Data)
1158 Out << M.Header << M.Data << M.Padding;
1159 } else {
1160 HeadersSize = sizeof(object::BigArchive::FixLenHdr);
1161 LastMemberEndOffset += *HeadersSize;
1162 LastMemberHeaderOffset += *HeadersSize;
1163
1164 // For the big archive (AIX) format, compute a table of member names and
1165 // offsets, used in the member table.
1166 uint64_t MemberTableNameStrTblSize = 0;
1167 std::vector<size_t> MemberOffsets;
1168 std::vector<StringRef> MemberNames;
1169 // Loop across object to find offset and names.
1170 uint64_t MemberEndOffset = sizeof(object::BigArchive::FixLenHdr);
1171 for (size_t I = 0, Size = NewMembers.size(); I != Size; ++I) {
1172 const NewArchiveMember &Member = NewMembers[I];
1173 MemberTableNameStrTblSize += Member.MemberName.size() + 1;
1174 MemberEndOffset += Data[I].PreHeadPadSize;
1175 MemberOffsets.push_back(x: MemberEndOffset);
1176 MemberNames.push_back(x: Member.MemberName);
1177 // File member name ended with "`\n". The length is included in
1178 // BigArMemHdrType.
1179 MemberEndOffset += sizeof(object::BigArMemHdrType) +
1180 alignTo(Value: Data[I].Data.size(), Align: 2) +
1181 alignTo(Value: Member.MemberName.size(), Align: 2);
1182 }
1183
1184 // AIX member table size.
1185 uint64_t MemberTableSize = 20 + // Number of members field
1186 20 * MemberOffsets.size() +
1187 MemberTableNameStrTblSize;
1188
1189 SmallString<0> SymNamesBuf32;
1190 SmallString<0> SymNamesBuf64;
1191 raw_svector_ostream SymNames32(SymNamesBuf32);
1192 raw_svector_ostream SymNames64(SymNamesBuf64);
1193
1194 if (ShouldWriteSymtab && NumSyms)
1195 // Generate the symbol names for the members.
1196 for (const auto &M : Data) {
1197 Expected<std::vector<unsigned>> SymbolsOrErr = getSymbols(
1198 Obj: M.SymFile.get(), Index: 0,
1199 SymNames&: is64BitSymbolicFile(SymObj: M.SymFile.get()) ? SymNames64 : SymNames32,
1200 SymMap: nullptr);
1201 if (!SymbolsOrErr)
1202 return SymbolsOrErr.takeError();
1203 }
1204
1205 uint64_t MemberTableEndOffset =
1206 LastMemberEndOffset +
1207 alignTo(Value: sizeof(object::BigArMemHdrType) + MemberTableSize, Align: 2);
1208
1209 // In AIX OS, The 'GlobSymOffset' field in the fixed-length header contains
1210 // the offset to the 32-bit global symbol table, and the 'GlobSym64Offset'
1211 // contains the offset to the 64-bit global symbol table.
1212 uint64_t GlobalSymbolOffset =
1213 (ShouldWriteSymtab &&
1214 (WriteSymtab != SymtabWritingMode::BigArchive64) && NumSyms32 > 0)
1215 ? MemberTableEndOffset
1216 : 0;
1217
1218 uint64_t GlobalSymbolOffset64 = 0;
1219 uint64_t NumSyms64 = NumSyms - NumSyms32;
1220 if (ShouldWriteSymtab && (WriteSymtab != SymtabWritingMode::BigArchive32) &&
1221 NumSyms64 > 0) {
1222 if (GlobalSymbolOffset == 0)
1223 GlobalSymbolOffset64 = MemberTableEndOffset;
1224 else
1225 // If there is a global symbol table for 32-bit members,
1226 // the 64-bit global symbol table is after the 32-bit one.
1227 GlobalSymbolOffset64 =
1228 GlobalSymbolOffset + sizeof(object::BigArMemHdrType) +
1229 (NumSyms32 + 1) * 8 + alignTo(Value: SymNamesBuf32.size(), Align: 2);
1230 }
1231
1232 // Fixed Sized Header.
1233 printWithSpacePadding(OS&: Out, Data: NewMembers.size() ? LastMemberEndOffset : 0,
1234 Size: 20); // Offset to member table
1235 // If there are no file members in the archive, there will be no global
1236 // symbol table.
1237 printWithSpacePadding(OS&: Out, Data: GlobalSymbolOffset, Size: 20);
1238 printWithSpacePadding(OS&: Out, Data: GlobalSymbolOffset64, Size: 20);
1239 printWithSpacePadding(OS&: Out,
1240 Data: NewMembers.size()
1241 ? sizeof(object::BigArchive::FixLenHdr) +
1242 Data[0].PreHeadPadSize
1243 : 0,
1244 Size: 20); // Offset to first archive member
1245 printWithSpacePadding(OS&: Out, Data: NewMembers.size() ? LastMemberHeaderOffset : 0,
1246 Size: 20); // Offset to last archive member
1247 printWithSpacePadding(
1248 OS&: Out, Data: 0,
1249 Size: 20); // Offset to first member of free list - Not supported yet
1250
1251 for (const MemberData &M : Data) {
1252 Out << std::string(M.PreHeadPadSize, '\0');
1253 Out << M.Header << M.Data;
1254 if (M.Data.size() % 2)
1255 Out << '\0';
1256 }
1257
1258 if (NewMembers.size()) {
1259 // Member table.
1260 printBigArchiveMemberHeader(Out, Name: "", ModTime: sys::toTimePoint(T: 0), UID: 0, GID: 0, Perms: 0,
1261 Size: MemberTableSize, PrevOffset: LastMemberHeaderOffset,
1262 NextOffset: GlobalSymbolOffset ? GlobalSymbolOffset
1263 : GlobalSymbolOffset64);
1264 printWithSpacePadding(OS&: Out, Data: MemberOffsets.size(), Size: 20); // Number of members
1265 for (uint64_t MemberOffset : MemberOffsets)
1266 printWithSpacePadding(OS&: Out, Data: MemberOffset,
1267 Size: 20); // Offset to member file header.
1268 for (StringRef MemberName : MemberNames)
1269 Out << MemberName << '\0'; // Member file name, null byte padding.
1270
1271 if (MemberTableNameStrTblSize % 2)
1272 Out << '\0'; // Name table must be tail padded to an even number of
1273 // bytes.
1274
1275 if (ShouldWriteSymtab) {
1276 // Write global symbol table for 32-bit file members.
1277 if (GlobalSymbolOffset) {
1278 writeSymbolTable(Out, Kind, Deterministic, Members: Data, StringTable: SymNamesBuf32,
1279 MembersOffset: *HeadersSize, NumSyms: NumSyms32, PrevMemberOffset: LastMemberEndOffset,
1280 NextMemberOffset: GlobalSymbolOffset64);
1281 // Add padding between the symbol tables, if needed.
1282 if (GlobalSymbolOffset64 && (SymNamesBuf32.size() % 2))
1283 Out << '\0';
1284 }
1285
1286 // Write global symbol table for 64-bit file members.
1287 if (GlobalSymbolOffset64)
1288 writeSymbolTable(Out, Kind, Deterministic, Members: Data, StringTable: SymNamesBuf64,
1289 MembersOffset: *HeadersSize, NumSyms: NumSyms64,
1290 PrevMemberOffset: GlobalSymbolOffset ? GlobalSymbolOffset
1291 : LastMemberEndOffset,
1292 NextMemberOffset: 0, Is64Bit: true);
1293 }
1294 }
1295 }
1296 Out.flush();
1297 return Error::success();
1298}
1299
1300void warnToStderr(Error Err) {
1301 llvm::logAllUnhandledErrors(E: std::move(Err), OS&: llvm::errs(), ErrorBanner: "warning: ");
1302}
1303
1304Error writeArchive(StringRef ArcName, ArrayRef<NewArchiveMember> NewMembers,
1305 SymtabWritingMode WriteSymtab, object::Archive::Kind Kind,
1306 bool Deterministic, bool Thin,
1307 std::unique_ptr<MemoryBuffer> OldArchiveBuf,
1308 std::optional<bool> IsEC, function_ref<void(Error)> Warn) {
1309 Expected<sys::fs::TempFile> Temp =
1310 sys::fs::TempFile::create(Model: ArcName + ".temp-archive-%%%%%%%.a");
1311 if (!Temp)
1312 return Temp.takeError();
1313 raw_fd_ostream Out(Temp->FD, false);
1314
1315 if (Error E = writeArchiveToStream(Out, NewMembers, WriteSymtab, Kind,
1316 Deterministic, Thin, IsEC, Warn)) {
1317 if (Error DiscardError = Temp->discard())
1318 return joinErrors(E1: std::move(E), E2: std::move(DiscardError));
1319 return E;
1320 }
1321
1322 // At this point, we no longer need whatever backing memory
1323 // was used to generate the NewMembers. On Windows, this buffer
1324 // could be a mapped view of the file we want to replace (if
1325 // we're updating an existing archive, say). In that case, the
1326 // rename would still succeed, but it would leave behind a
1327 // temporary file (actually the original file renamed) because
1328 // a file cannot be deleted while there's a handle open on it,
1329 // only renamed. So by freeing this buffer, this ensures that
1330 // the last open handle on the destination file, if any, is
1331 // closed before we attempt to rename.
1332 OldArchiveBuf.reset();
1333
1334 return Temp->keep(Name: ArcName);
1335}
1336
1337Expected<std::unique_ptr<MemoryBuffer>>
1338writeArchiveToBuffer(ArrayRef<NewArchiveMember> NewMembers,
1339 SymtabWritingMode WriteSymtab, object::Archive::Kind Kind,
1340 bool Deterministic, bool Thin,
1341 function_ref<void(Error)> Warn) {
1342 SmallVector<char, 0> ArchiveBufferVector;
1343 raw_svector_ostream ArchiveStream(ArchiveBufferVector);
1344
1345 if (Error E =
1346 writeArchiveToStream(Out&: ArchiveStream, NewMembers, WriteSymtab, Kind,
1347 Deterministic, Thin, IsEC: std::nullopt, Warn))
1348 return std::move(E);
1349
1350 return std::make_unique<SmallVectorMemoryBuffer>(
1351 args: std::move(ArchiveBufferVector), /*RequiresNullTerminator=*/args: false);
1352}
1353
1354} // namespace llvm
1355