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     return false;
126   case object::Archive::K_BSD:
127   case object::Archive::K_DARWIN:
128     return true;
129   case object::Archive::K_GNU64:
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 static void print32(raw_ostream &Out, object::Archive::Kind Kind,
138                     uint32_t Val) {
139   if (isBSDLike(Kind))
140     support::endian::Writer<support::little>(Out).write(Val);
141   else
142     support::endian::Writer<support::big>(Out).write(Val);
143 }
144 
145 static void printRestOfMemberHeader(
146     raw_ostream &Out, const sys::TimePoint<std::chrono::seconds> &ModTime,
147     unsigned UID, unsigned GID, unsigned Perms, unsigned Size) {
148   printWithSpacePadding(Out, sys::toTimeT(ModTime), 12);
149 
150   // The format has only 6 chars for uid and gid. Truncate if the provided
151   // values don't fit.
152   printWithSpacePadding(Out, UID % 1000000, 6);
153   printWithSpacePadding(Out, GID % 1000000, 6);
154 
155   printWithSpacePadding(Out, format("%o", Perms), 8);
156   printWithSpacePadding(Out, Size, 10);
157   Out << "`\n";
158 }
159 
160 static void
161 printGNUSmallMemberHeader(raw_ostream &Out, StringRef Name,
162                           const sys::TimePoint<std::chrono::seconds> &ModTime,
163                           unsigned UID, unsigned GID, unsigned Perms,
164                           unsigned Size) {
165   printWithSpacePadding(Out, Twine(Name) + "/", 16);
166   printRestOfMemberHeader(Out, ModTime, UID, GID, Perms, Size);
167 }
168 
169 static void
170 printBSDMemberHeader(raw_ostream &Out, StringRef Name,
171                      const sys::TimePoint<std::chrono::seconds> &ModTime,
172                      unsigned UID, unsigned GID, unsigned Perms,
173                      unsigned Size) {
174   uint64_t PosAfterHeader = Out.tell() + 60 + Name.size();
175   // Pad so that even 64 bit object files are aligned.
176   unsigned Pad = OffsetToAlignment(PosAfterHeader, 8);
177   unsigned NameWithPadding = Name.size() + Pad;
178   printWithSpacePadding(Out, Twine("#1/") + Twine(NameWithPadding), 16);
179   printRestOfMemberHeader(Out, ModTime, UID, GID, Perms,
180                           NameWithPadding + Size);
181   Out << Name;
182   assert(PosAfterHeader == Out.tell());
183   while (Pad--)
184     Out.write(uint8_t(0));
185 }
186 
187 static bool useStringTable(bool Thin, StringRef Name) {
188   return Thin || Name.size() >= 16 || Name.contains('/');
189 }
190 
191 static void
192 printMemberHeader(raw_ostream &Out, object::Archive::Kind Kind, bool Thin,
193                   StringRef Name,
194                   std::vector<unsigned>::iterator &StringMapIndexIter,
195                   const sys::TimePoint<std::chrono::seconds> &ModTime,
196                   unsigned UID, unsigned GID, unsigned Perms, unsigned Size) {
197   if (isBSDLike(Kind))
198     return printBSDMemberHeader(Out, Name, ModTime, UID, GID, Perms, Size);
199   if (!useStringTable(Thin, Name))
200     return printGNUSmallMemberHeader(Out, Name, ModTime, UID, GID, Perms, Size);
201   Out << '/';
202   printWithSpacePadding(Out, *StringMapIndexIter++, 15);
203   printRestOfMemberHeader(Out, ModTime, UID, GID, Perms, Size);
204 }
205 
206 // Compute the relative path from From to To.
207 static std::string computeRelativePath(StringRef From, StringRef To) {
208   if (sys::path::is_absolute(From) || sys::path::is_absolute(To))
209     return To;
210 
211   StringRef DirFrom = sys::path::parent_path(From);
212   auto FromI = sys::path::begin(DirFrom);
213   auto ToI = sys::path::begin(To);
214   while (*FromI == *ToI) {
215     ++FromI;
216     ++ToI;
217   }
218 
219   SmallString<128> Relative;
220   for (auto FromE = sys::path::end(DirFrom); FromI != FromE; ++FromI)
221     sys::path::append(Relative, "..");
222 
223   for (auto ToE = sys::path::end(To); ToI != ToE; ++ToI)
224     sys::path::append(Relative, *ToI);
225 
226 #ifdef LLVM_ON_WIN32
227   // Replace backslashes with slashes so that the path is portable between *nix
228   // and Windows.
229   std::replace(Relative.begin(), Relative.end(), '\\', '/');
230 #endif
231 
232   return Relative.str();
233 }
234 
235 static void writeStringTable(raw_fd_ostream &Out, StringRef ArcName,
236                              ArrayRef<NewArchiveMember> Members,
237                              std::vector<unsigned> &StringMapIndexes,
238                              bool Thin) {
239   unsigned StartOffset = 0;
240   for (const NewArchiveMember &M : Members) {
241     StringRef Path = M.Buf->getBufferIdentifier();
242     StringRef Name = M.MemberName;
243     if (!useStringTable(Thin, Name))
244       continue;
245     if (StartOffset == 0) {
246       printWithSpacePadding(Out, "//", 58);
247       Out << "`\n";
248       StartOffset = Out.tell();
249     }
250     StringMapIndexes.push_back(Out.tell() - StartOffset);
251 
252     if (Thin) {
253       if (M.IsNew)
254         Out << computeRelativePath(ArcName, Path);
255       else
256         Out << M.Buf->getBufferIdentifier();
257     } else
258       Out << Name;
259 
260     Out << "/\n";
261   }
262   if (StartOffset == 0)
263     return;
264   if (Out.tell() % 2)
265     Out << '\n';
266   int Pos = Out.tell();
267   Out.seek(StartOffset - 12);
268   printWithSpacePadding(Out, Pos - StartOffset, 10);
269   Out.seek(Pos);
270 }
271 
272 static sys::TimePoint<std::chrono::seconds> now(bool Deterministic) {
273   using namespace std::chrono;
274 
275   if (!Deterministic)
276     return time_point_cast<seconds>(system_clock::now());
277   return sys::TimePoint<seconds>();
278 }
279 
280 static bool isArchiveSymbol(const object::BasicSymbolRef &S) {
281   uint32_t Symflags = S.getFlags();
282   if (Symflags & object::SymbolRef::SF_FormatSpecific)
283     return false;
284   if (!(Symflags & object::SymbolRef::SF_Global))
285     return false;
286   if (Symflags & object::SymbolRef::SF_Undefined &&
287       !(Symflags & object::SymbolRef::SF_Indirect))
288     return false;
289   return true;
290 }
291 
292 // Returns the offset of the first reference to a member offset.
293 static Expected<unsigned>
294 writeSymbolTable(raw_fd_ostream &Out, object::Archive::Kind Kind,
295                  ArrayRef<NewArchiveMember> Members,
296                  std::vector<unsigned> &MemberOffsetRefs, bool Deterministic) {
297   unsigned HeaderStartOffset = 0;
298   unsigned BodyStartOffset = 0;
299   SmallString<128> NameBuf;
300   raw_svector_ostream NameOS(NameBuf);
301   LLVMContext Context;
302   for (unsigned MemberNum = 0, N = Members.size(); MemberNum < N; ++MemberNum) {
303     MemoryBufferRef MemberBuffer = Members[MemberNum].Buf->getMemBufferRef();
304     Expected<std::unique_ptr<object::SymbolicFile>> ObjOrErr =
305         object::SymbolicFile::createSymbolicFile(
306             MemberBuffer, llvm::file_magic::unknown, &Context);
307     if (!ObjOrErr) {
308       // FIXME: check only for "not an object file" errors.
309       consumeError(ObjOrErr.takeError());
310       continue;
311     }
312     object::SymbolicFile &Obj = *ObjOrErr.get();
313 
314     if (!HeaderStartOffset) {
315       HeaderStartOffset = Out.tell();
316       if (isBSDLike(Kind))
317         printBSDMemberHeader(Out, "__.SYMDEF", now(Deterministic), 0, 0, 0, 0);
318       else
319         printGNUSmallMemberHeader(Out, "", now(Deterministic), 0, 0, 0, 0);
320       BodyStartOffset = Out.tell();
321       print32(Out, Kind, 0); // number of entries or bytes
322     }
323 
324     for (const object::BasicSymbolRef &S : Obj.symbols()) {
325       if (!isArchiveSymbol(S))
326         continue;
327 
328       unsigned NameOffset = NameOS.tell();
329       if (std::error_code EC = S.printName(NameOS))
330         return errorCodeToError(EC);
331       NameOS << '\0';
332       MemberOffsetRefs.push_back(MemberNum);
333       if (isBSDLike(Kind))
334         print32(Out, Kind, NameOffset);
335       print32(Out, Kind, 0); // member offset
336     }
337   }
338 
339   if (HeaderStartOffset == 0)
340     return 0;
341 
342   // ld64 prefers the cctools type archive which pads its string table to a
343   // boundary of sizeof(int32_t).
344   if (isBSDLike(Kind))
345     for (unsigned P = OffsetToAlignment(NameOS.tell(), sizeof(int32_t)); P--;)
346       NameOS << '\0';
347 
348   StringRef StringTable = NameOS.str();
349   if (isBSDLike(Kind))
350     print32(Out, Kind, StringTable.size()); // byte count of the string table
351   Out << StringTable;
352   // If there are no symbols, emit an empty symbol table, to satisfy Solaris
353   // tools, older versions of which expect a symbol table in a non-empty
354   // archive, regardless of whether there are any symbols in it.
355   if (StringTable.size() == 0)
356     print32(Out, Kind, 0);
357 
358   // ld64 expects the members to be 8-byte aligned for 64-bit content and at
359   // least 4-byte aligned for 32-bit content.  Opt for the larger encoding
360   // uniformly.
361   // We do this for all bsd formats because it simplifies aligning members.
362   unsigned Alignment = isBSDLike(Kind) ? 8 : 2;
363   unsigned Pad = OffsetToAlignment(Out.tell(), Alignment);
364   while (Pad--)
365     Out.write(uint8_t(0));
366 
367   // Patch up the size of the symbol table now that we know how big it is.
368   unsigned Pos = Out.tell();
369   const unsigned MemberHeaderSize = 60;
370   Out.seek(HeaderStartOffset + 48); // offset of the size field.
371   printWithSpacePadding(Out, Pos - MemberHeaderSize - HeaderStartOffset, 10);
372 
373   // Patch up the number of symbols.
374   Out.seek(BodyStartOffset);
375   unsigned NumSyms = MemberOffsetRefs.size();
376   if (isBSDLike(Kind))
377     print32(Out, Kind, NumSyms * 8);
378   else
379     print32(Out, Kind, NumSyms);
380 
381   Out.seek(Pos);
382   return BodyStartOffset + 4;
383 }
384 
385 Error llvm::writeArchive(StringRef ArcName,
386                          ArrayRef<NewArchiveMember> NewMembers,
387                          bool WriteSymtab, object::Archive::Kind Kind,
388                          bool Deterministic, bool Thin,
389                          std::unique_ptr<MemoryBuffer> OldArchiveBuf) {
390   assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode");
391   SmallString<128> TmpArchive;
392   int TmpArchiveFD;
393   if (auto EC = sys::fs::createUniqueFile(ArcName + ".temp-archive-%%%%%%%.a",
394                                           TmpArchiveFD, TmpArchive))
395     return errorCodeToError(EC);
396 
397   ToolOutputFile Output(TmpArchive, TmpArchiveFD);
398   raw_fd_ostream &Out = Output.os();
399   if (Thin)
400     Out << "!<thin>\n";
401   else
402     Out << "!<arch>\n";
403 
404   std::vector<unsigned> MemberOffsetRefs;
405 
406   unsigned MemberReferenceOffset = 0;
407   if (WriteSymtab) {
408     Expected<unsigned> MemberReferenceOffsetOrErr = writeSymbolTable(
409         Out, Kind, NewMembers, MemberOffsetRefs, Deterministic);
410     if (auto E = MemberReferenceOffsetOrErr.takeError())
411       return E;
412     MemberReferenceOffset = MemberReferenceOffsetOrErr.get();
413   }
414 
415   std::vector<unsigned> StringMapIndexes;
416   if (!isBSDLike(Kind))
417     writeStringTable(Out, ArcName, NewMembers, StringMapIndexes, Thin);
418 
419   std::vector<unsigned>::iterator StringMapIndexIter = StringMapIndexes.begin();
420   std::vector<unsigned> MemberOffset;
421   for (const NewArchiveMember &M : NewMembers) {
422     MemoryBufferRef File = M.Buf->getMemBufferRef();
423     unsigned Padding = 0;
424 
425     unsigned Pos = Out.tell();
426     MemberOffset.push_back(Pos);
427 
428     // ld64 expects the members to be 8-byte aligned for 64-bit content and at
429     // least 4-byte aligned for 32-bit content.  Opt for the larger encoding
430     // uniformly.  This matches the behaviour with cctools and ensures that ld64
431     // is happy with archives that we generate.
432     if (Kind == object::Archive::K_DARWIN)
433       Padding = OffsetToAlignment(M.Buf->getBufferSize(), 8);
434 
435     printMemberHeader(Out, Kind, Thin, M.MemberName, StringMapIndexIter,
436                       M.ModTime, M.UID, M.GID, M.Perms,
437                       M.Buf->getBufferSize() + Padding);
438 
439     if (!Thin)
440       Out << File.getBuffer();
441 
442     while (Padding--)
443       Out << '\n';
444     if (Out.tell() % 2)
445       Out << '\n';
446   }
447 
448   if (MemberReferenceOffset) {
449     Out.seek(MemberReferenceOffset);
450     for (unsigned MemberNum : MemberOffsetRefs) {
451       if (isBSDLike(Kind))
452         Out.seek(Out.tell() + 4); // skip over the string offset
453       print32(Out, Kind, MemberOffset[MemberNum]);
454     }
455   }
456 
457   Output.keep();
458   Out.close();
459 
460   // At this point, we no longer need whatever backing memory
461   // was used to generate the NewMembers. On Windows, this buffer
462   // could be a mapped view of the file we want to replace (if
463   // we're updating an existing archive, say). In that case, the
464   // rename would still succeed, but it would leave behind a
465   // temporary file (actually the original file renamed) because
466   // a file cannot be deleted while there's a handle open on it,
467   // only renamed. So by freeing this buffer, this ensures that
468   // the last open handle on the destination file, if any, is
469   // closed before we attempt to rename.
470   OldArchiveBuf.reset();
471 
472   sys::fs::rename(TmpArchive, ArcName);
473   return Error::success();
474 }
475