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 
41 Expected<NewArchiveMember>
42 NewArchiveMember::getOldMember(const object::Archive::Child &OldMember,
43                                bool Deterministic) {
44   Expected<llvm::MemoryBufferRef> BufOrErr = OldMember.getMemoryBufferRef();
45   if (!BufOrErr)
46     return BufOrErr.takeError();
47 
48   NewArchiveMember M;
49   assert(M.IsNew == false);
50   M.Buf = MemoryBuffer::getMemBuffer(*BufOrErr, false);
51   if (!Deterministic) {
52     auto ModTimeOrErr = OldMember.getLastModified();
53     if (!ModTimeOrErr)
54       return ModTimeOrErr.takeError();
55     M.ModTime = ModTimeOrErr.get();
56     Expected<unsigned> UIDOrErr = OldMember.getUID();
57     if (!UIDOrErr)
58       return UIDOrErr.takeError();
59     M.UID = UIDOrErr.get();
60     Expected<unsigned> GIDOrErr = OldMember.getGID();
61     if (!GIDOrErr)
62       return GIDOrErr.takeError();
63     M.GID = GIDOrErr.get();
64     Expected<sys::fs::perms> AccessModeOrErr = OldMember.getAccessMode();
65     if (!AccessModeOrErr)
66       return AccessModeOrErr.takeError();
67     M.Perms = AccessModeOrErr.get();
68   }
69   return std::move(M);
70 }
71 
72 Expected<NewArchiveMember> NewArchiveMember::getFile(StringRef FileName,
73                                                      bool Deterministic) {
74   sys::fs::file_status Status;
75   int FD;
76   if (auto EC = sys::fs::openFileForRead(FileName, FD))
77     return errorCodeToError(EC);
78   assert(FD != -1);
79 
80   if (auto EC = sys::fs::status(FD, Status))
81     return errorCodeToError(EC);
82 
83   // Opening a directory doesn't make sense. Let it fail.
84   // Linux cannot open directories with open(2), although
85   // cygwin and *bsd can.
86   if (Status.type() == sys::fs::file_type::directory_file)
87     return errorCodeToError(make_error_code(errc::is_a_directory));
88 
89   ErrorOr<std::unique_ptr<MemoryBuffer>> MemberBufferOrErr =
90       MemoryBuffer::getOpenFile(FD, FileName, Status.getSize(), false);
91   if (!MemberBufferOrErr)
92     return errorCodeToError(MemberBufferOrErr.getError());
93 
94   if (close(FD) != 0)
95     return errorCodeToError(std::error_code(errno, std::generic_category()));
96 
97   NewArchiveMember M;
98   M.IsNew = true;
99   M.Buf = std::move(*MemberBufferOrErr);
100   if (!Deterministic) {
101     M.ModTime = std::chrono::time_point_cast<std::chrono::seconds>(
102         Status.getLastModificationTime());
103     M.UID = Status.getUser();
104     M.GID = Status.getGroup();
105     M.Perms = Status.permissions();
106   }
107   return std::move(M);
108 }
109 
110 template <typename T>
111 static void printWithSpacePadding(raw_fd_ostream &OS, T Data, unsigned Size,
112                                   bool MayTruncate = false) {
113   uint64_t OldPos = OS.tell();
114   OS << Data;
115   unsigned SizeSoFar = OS.tell() - OldPos;
116   if (Size > SizeSoFar) {
117     OS.indent(Size - SizeSoFar);
118   } else if (Size < SizeSoFar) {
119     assert(MayTruncate && "Data doesn't fit in Size");
120     // Some of the data this is used for (like UID) can be larger than the
121     // space available in the archive format. Truncate in that case.
122     OS.seek(OldPos + Size);
123   }
124 }
125 
126 static bool isBSDLike(object::Archive::Kind Kind) {
127   switch (Kind) {
128   case object::Archive::K_GNU:
129     return false;
130   case object::Archive::K_BSD:
131   case object::Archive::K_DARWIN:
132     return true;
133   case object::Archive::K_MIPS64:
134   case object::Archive::K_DARWIN64:
135   case object::Archive::K_COFF:
136     break;
137   }
138   llvm_unreachable("not supported for writting");
139 }
140 
141 static void print32(raw_ostream &Out, object::Archive::Kind Kind,
142                     uint32_t Val) {
143   if (isBSDLike(Kind))
144     support::endian::Writer<support::little>(Out).write(Val);
145   else
146     support::endian::Writer<support::big>(Out).write(Val);
147 }
148 
149 static void printRestOfMemberHeader(
150     raw_fd_ostream &Out, const sys::TimePoint<std::chrono::seconds> &ModTime,
151     unsigned UID, unsigned GID, unsigned Perms, unsigned Size) {
152   printWithSpacePadding(Out, sys::toTimeT(ModTime), 12);
153   printWithSpacePadding(Out, UID, 6, true);
154   printWithSpacePadding(Out, GID, 6, true);
155   printWithSpacePadding(Out, format("%o", Perms), 8);
156   printWithSpacePadding(Out, Size, 10);
157   Out << "`\n";
158 }
159 
160 static void
161 printGNUSmallMemberHeader(raw_fd_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_fd_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;
189 }
190 
191 static void
192 printMemberHeader(raw_fd_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 = sys::path::filename(Path);
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 // Returns the offset of the first reference to a member offset.
281 static ErrorOr<unsigned>
282 writeSymbolTable(raw_fd_ostream &Out, object::Archive::Kind Kind,
283                  ArrayRef<NewArchiveMember> Members,
284                  std::vector<unsigned> &MemberOffsetRefs, bool Deterministic) {
285   unsigned HeaderStartOffset = 0;
286   unsigned BodyStartOffset = 0;
287   SmallString<128> NameBuf;
288   raw_svector_ostream NameOS(NameBuf);
289   LLVMContext Context;
290   for (unsigned MemberNum = 0, N = Members.size(); MemberNum < N; ++MemberNum) {
291     MemoryBufferRef MemberBuffer = Members[MemberNum].Buf->getMemBufferRef();
292     Expected<std::unique_ptr<object::SymbolicFile>> ObjOrErr =
293         object::SymbolicFile::createSymbolicFile(
294             MemberBuffer, llvm::file_magic::unknown, &Context);
295     if (!ObjOrErr) {
296       // FIXME: check only for "not an object file" errors.
297       consumeError(ObjOrErr.takeError());
298       continue;
299     }
300     object::SymbolicFile &Obj = *ObjOrErr.get();
301 
302     if (!HeaderStartOffset) {
303       HeaderStartOffset = Out.tell();
304       if (isBSDLike(Kind))
305         printBSDMemberHeader(Out, "__.SYMDEF", now(Deterministic), 0, 0, 0, 0);
306       else
307         printGNUSmallMemberHeader(Out, "", now(Deterministic), 0, 0, 0, 0);
308       BodyStartOffset = Out.tell();
309       print32(Out, Kind, 0); // number of entries or bytes
310     }
311 
312     for (const object::BasicSymbolRef &S : Obj.symbols()) {
313       uint32_t Symflags = S.getFlags();
314       if (Symflags & object::SymbolRef::SF_FormatSpecific)
315         continue;
316       if (!(Symflags & object::SymbolRef::SF_Global))
317         continue;
318       if (Symflags & object::SymbolRef::SF_Undefined)
319         continue;
320 
321       unsigned NameOffset = NameOS.tell();
322       if (auto EC = S.printName(NameOS))
323         return EC;
324       NameOS << '\0';
325       MemberOffsetRefs.push_back(MemberNum);
326       if (isBSDLike(Kind))
327         print32(Out, Kind, NameOffset);
328       print32(Out, Kind, 0); // member offset
329     }
330   }
331 
332   if (HeaderStartOffset == 0)
333     return 0;
334 
335   // ld64 prefers the cctools type archive which pads its string table to a
336   // boundary of sizeof(int32_t).
337   if (isBSDLike(Kind))
338     for (unsigned P = OffsetToAlignment(NameOS.tell(), sizeof(int32_t)); P--;)
339       NameOS << '\0';
340 
341   StringRef StringTable = NameOS.str();
342   if (isBSDLike(Kind))
343     print32(Out, Kind, StringTable.size()); // byte count of the string table
344   Out << StringTable;
345   // If there are no symbols, emit an empty symbol table, to satisfy Solaris
346   // tools, older versions of which expect a symbol table in a non-empty
347   // archive, regardless of whether there are any symbols in it.
348   if (StringTable.size() == 0)
349     print32(Out, Kind, 0);
350 
351   // ld64 requires the next member header to start at an offset that is
352   // 4 bytes aligned.
353   unsigned Pad = OffsetToAlignment(Out.tell(), 4);
354   while (Pad--)
355     Out.write(uint8_t(0));
356 
357   // Patch up the size of the symbol table now that we know how big it is.
358   unsigned Pos = Out.tell();
359   const unsigned MemberHeaderSize = 60;
360   Out.seek(HeaderStartOffset + 48); // offset of the size field.
361   printWithSpacePadding(Out, Pos - MemberHeaderSize - HeaderStartOffset, 10);
362 
363   // Patch up the number of symbols.
364   Out.seek(BodyStartOffset);
365   unsigned NumSyms = MemberOffsetRefs.size();
366   if (isBSDLike(Kind))
367     print32(Out, Kind, NumSyms * 8);
368   else
369     print32(Out, Kind, NumSyms);
370 
371   Out.seek(Pos);
372   return BodyStartOffset + 4;
373 }
374 
375 std::pair<StringRef, std::error_code>
376 llvm::writeArchive(StringRef ArcName,
377                    std::vector<NewArchiveMember> &NewMembers,
378                    bool WriteSymtab, object::Archive::Kind Kind,
379                    bool Deterministic, bool Thin,
380                    std::unique_ptr<MemoryBuffer> OldArchiveBuf) {
381   assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode");
382   SmallString<128> TmpArchive;
383   int TmpArchiveFD;
384   if (auto EC = sys::fs::createUniqueFile(ArcName + ".temp-archive-%%%%%%%.a",
385                                           TmpArchiveFD, TmpArchive))
386     return std::make_pair(ArcName, EC);
387 
388   tool_output_file Output(TmpArchive, TmpArchiveFD);
389   raw_fd_ostream &Out = Output.os();
390   if (Thin)
391     Out << "!<thin>\n";
392   else
393     Out << "!<arch>\n";
394 
395   std::vector<unsigned> MemberOffsetRefs;
396 
397   unsigned MemberReferenceOffset = 0;
398   if (WriteSymtab) {
399     ErrorOr<unsigned> MemberReferenceOffsetOrErr = writeSymbolTable(
400         Out, Kind, NewMembers, MemberOffsetRefs, Deterministic);
401     if (auto EC = MemberReferenceOffsetOrErr.getError())
402       return std::make_pair(ArcName, EC);
403     MemberReferenceOffset = MemberReferenceOffsetOrErr.get();
404   }
405 
406   std::vector<unsigned> StringMapIndexes;
407   if (!isBSDLike(Kind))
408     writeStringTable(Out, ArcName, NewMembers, StringMapIndexes, Thin);
409 
410   std::vector<unsigned>::iterator StringMapIndexIter = StringMapIndexes.begin();
411   std::vector<unsigned> MemberOffset;
412   for (const NewArchiveMember &M : NewMembers) {
413     MemoryBufferRef File = M.Buf->getMemBufferRef();
414     unsigned Padding = 0;
415 
416     unsigned Pos = Out.tell();
417     MemberOffset.push_back(Pos);
418 
419     // ld64 expects the members to be 8-byte aligned for 64-bit content and at
420     // least 4-byte aligned for 32-bit content.  Opt for the larger encoding
421     // uniformly.  This matches the behaviour with cctools and ensures that ld64
422     // is happy with archives that we generate.
423     if (Kind == object::Archive::K_DARWIN)
424       Padding = OffsetToAlignment(M.Buf->getBufferSize(), 8);
425 
426     printMemberHeader(Out, Kind, Thin,
427                       sys::path::filename(M.Buf->getBufferIdentifier()),
428                       StringMapIndexIter, M.ModTime, M.UID, M.GID, M.Perms,
429                       M.Buf->getBufferSize() + Padding);
430 
431     if (!Thin)
432       Out << File.getBuffer();
433 
434     while (Padding--)
435       Out << '\n';
436     if (Out.tell() % 2)
437       Out << '\n';
438   }
439 
440   if (MemberReferenceOffset) {
441     Out.seek(MemberReferenceOffset);
442     for (unsigned MemberNum : MemberOffsetRefs) {
443       if (isBSDLike(Kind))
444         Out.seek(Out.tell() + 4); // skip over the string offset
445       print32(Out, Kind, MemberOffset[MemberNum]);
446     }
447   }
448 
449   Output.keep();
450   Out.close();
451 
452   // At this point, we no longer need whatever backing memory
453   // was used to generate the NewMembers. On Windows, this buffer
454   // could be a mapped view of the file we want to replace (if
455   // we're updating an existing archive, say). In that case, the
456   // rename would still succeed, but it would leave behind a
457   // temporary file (actually the original file renamed) because
458   // a file cannot be deleted while there's a handle open on it,
459   // only renamed. So by freeing this buffer, this ensures that
460   // the last open handle on the destination file, if any, is
461   // closed before we attempt to rename.
462   OldArchiveBuf.reset();
463 
464   sys::fs::rename(TmpArchive, ArcName);
465   return std::make_pair("", std::error_code());
466 }
467