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, uint64_t Pos, StringRef Name,
171                      const sys::TimePoint<std::chrono::seconds> &ModTime,
172                      unsigned UID, unsigned GID, unsigned Perms,
173                      unsigned Size) {
174   uint64_t PosAfterHeader = Pos + 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   while (Pad--)
183     Out.write(uint8_t(0));
184 }
185 
186 static bool useStringTable(bool Thin, StringRef Name) {
187   return Thin || Name.size() >= 16 || Name.contains('/');
188 }
189 
190 // Compute the relative path from From to To.
191 static std::string computeRelativePath(StringRef From, StringRef To) {
192   if (sys::path::is_absolute(From) || sys::path::is_absolute(To))
193     return To;
194 
195   StringRef DirFrom = sys::path::parent_path(From);
196   auto FromI = sys::path::begin(DirFrom);
197   auto ToI = sys::path::begin(To);
198   while (*FromI == *ToI) {
199     ++FromI;
200     ++ToI;
201   }
202 
203   SmallString<128> Relative;
204   for (auto FromE = sys::path::end(DirFrom); FromI != FromE; ++FromI)
205     sys::path::append(Relative, "..");
206 
207   for (auto ToE = sys::path::end(To); ToI != ToE; ++ToI)
208     sys::path::append(Relative, *ToI);
209 
210 #ifdef LLVM_ON_WIN32
211   // Replace backslashes with slashes so that the path is portable between *nix
212   // and Windows.
213   std::replace(Relative.begin(), Relative.end(), '\\', '/');
214 #endif
215 
216   return Relative.str();
217 }
218 
219 static void addToStringTable(raw_ostream &Out, StringRef ArcName,
220                              const NewArchiveMember &M, bool Thin) {
221   StringRef ID = M.Buf->getBufferIdentifier();
222   if (Thin) {
223     if (M.IsNew)
224       Out << computeRelativePath(ArcName, ID);
225     else
226       Out << ID;
227   } else
228     Out << M.MemberName;
229   Out << "/\n";
230 }
231 
232 static void printMemberHeader(raw_ostream &Out, uint64_t Pos,
233                               raw_ostream &StringTable,
234                               object::Archive::Kind Kind, bool Thin,
235                               StringRef ArcName, const NewArchiveMember &M,
236                               unsigned Size) {
237   if (isBSDLike(Kind))
238     return printBSDMemberHeader(Out, Pos, M.MemberName, M.ModTime, M.UID, M.GID,
239                                 M.Perms, Size);
240   if (!useStringTable(Thin, M.MemberName))
241     return printGNUSmallMemberHeader(Out, M.MemberName, M.ModTime, M.UID, M.GID,
242                                      M.Perms, Size);
243   Out << '/';
244   uint64_t NamePos = StringTable.tell();
245   addToStringTable(StringTable, ArcName, M, Thin);
246   printWithSpacePadding(Out, NamePos, 15);
247   printRestOfMemberHeader(Out, M.ModTime, M.UID, M.GID, M.Perms, Size);
248 }
249 
250 namespace {
251 struct MemberData {
252   std::vector<unsigned> Symbols;
253   std::string Header;
254   StringRef Data;
255   StringRef Padding;
256 };
257 } // namespace
258 
259 static MemberData computeStringTable(StringRef Names) {
260   unsigned Size = Names.size();
261   unsigned Pad = OffsetToAlignment(Size, 2);
262   std::string Header;
263   raw_string_ostream Out(Header);
264   printWithSpacePadding(Out, "//", 48);
265   printWithSpacePadding(Out, Size + Pad, 10);
266   Out << "`\n";
267   Out.flush();
268   return {{}, std::move(Header), Names, Pad ? "\n" : ""};
269 }
270 
271 static sys::TimePoint<std::chrono::seconds> now(bool Deterministic) {
272   using namespace std::chrono;
273 
274   if (!Deterministic)
275     return time_point_cast<seconds>(system_clock::now());
276   return sys::TimePoint<seconds>();
277 }
278 
279 static bool isArchiveSymbol(const object::BasicSymbolRef &S) {
280   uint32_t Symflags = S.getFlags();
281   if (Symflags & object::SymbolRef::SF_FormatSpecific)
282     return false;
283   if (!(Symflags & object::SymbolRef::SF_Global))
284     return false;
285   if (Symflags & object::SymbolRef::SF_Undefined &&
286       !(Symflags & object::SymbolRef::SF_Indirect))
287     return false;
288   return true;
289 }
290 
291 static void writeSymbolTable(raw_ostream &Out, object::Archive::Kind Kind,
292                              bool Deterministic, ArrayRef<MemberData> Members,
293                              StringRef StringTable) {
294   if (StringTable.empty())
295     return;
296 
297   unsigned NumSyms = 0;
298   for (const MemberData &M : Members)
299     NumSyms += M.Symbols.size();
300 
301   unsigned Size = 0;
302   Size += 4; // Number of entries
303   if (isBSDLike(Kind))
304     Size += NumSyms * 8; // Table
305   else
306     Size += NumSyms * 4; // Table
307   if (isBSDLike(Kind))
308     Size += 4; // byte count
309   Size += StringTable.size();
310   // ld64 expects the members to be 8-byte aligned for 64-bit content and at
311   // least 4-byte aligned for 32-bit content.  Opt for the larger encoding
312   // uniformly.
313   // We do this for all bsd formats because it simplifies aligning members.
314   unsigned Alignment = isBSDLike(Kind) ? 8 : 2;
315   unsigned Pad = OffsetToAlignment(Size, Alignment);
316   Size += Pad;
317 
318   if (isBSDLike(Kind))
319     printBSDMemberHeader(Out, Out.tell(), "__.SYMDEF", now(Deterministic), 0, 0,
320                          0, Size);
321   else
322     printGNUSmallMemberHeader(Out, "", now(Deterministic), 0, 0, 0, Size);
323 
324   uint64_t Pos = Out.tell() + Size;
325 
326   if (isBSDLike(Kind))
327     print32(Out, Kind, NumSyms * 8);
328   else
329     print32(Out, Kind, NumSyms);
330 
331   for (const MemberData &M : Members) {
332     for (unsigned StringOffset : M.Symbols) {
333       if (isBSDLike(Kind))
334         print32(Out, Kind, StringOffset);
335       print32(Out, Kind, Pos); // member offset
336     }
337     Pos += M.Header.size() + M.Data.size() + M.Padding.size();
338   }
339 
340   if (isBSDLike(Kind))
341     print32(Out, Kind, StringTable.size()); // byte count of the string table
342   Out << StringTable;
343 
344   while (Pad--)
345     Out.write(uint8_t(0));
346 }
347 
348 static Expected<std::vector<unsigned>>
349 getSymbols(MemoryBufferRef Buf, raw_ostream &SymNames, bool &HasObject) {
350   std::vector<unsigned> Ret;
351   LLVMContext Context;
352 
353   Expected<std::unique_ptr<object::SymbolicFile>> ObjOrErr =
354       object::SymbolicFile::createSymbolicFile(Buf, llvm::file_magic::unknown,
355                                                &Context);
356   if (!ObjOrErr) {
357     // FIXME: check only for "not an object file" errors.
358     consumeError(ObjOrErr.takeError());
359     return Ret;
360   }
361 
362   HasObject = true;
363   object::SymbolicFile &Obj = *ObjOrErr.get();
364   for (const object::BasicSymbolRef &S : Obj.symbols()) {
365     if (!isArchiveSymbol(S))
366       continue;
367     Ret.push_back(SymNames.tell());
368     if (auto EC = S.printName(SymNames))
369       return errorCodeToError(EC);
370     SymNames << '\0';
371   }
372   return Ret;
373 }
374 
375 static Expected<std::vector<MemberData>>
376 computeMemberData(raw_ostream &StringTable, raw_ostream &SymNames,
377                   object::Archive::Kind Kind, bool Thin, StringRef ArcName,
378                   ArrayRef<NewArchiveMember> NewMembers) {
379   static char PaddingData[8] = {'\n', '\n', '\n', '\n', '\n', '\n', '\n', '\n'};
380 
381   // This ignores the symbol table, but we only need the value mod 8 and the
382   // symbol table is aligned to be a multiple of 8 bytes
383   uint64_t Pos = 0;
384 
385   std::vector<MemberData> Ret;
386   bool HasObject = false;
387   for (const NewArchiveMember &M : NewMembers) {
388     std::string Header;
389     raw_string_ostream Out(Header);
390 
391     MemoryBufferRef Buf = M.Buf->getMemBufferRef();
392     StringRef Data = Thin ? "" : Buf.getBuffer();
393 
394     // ld64 expects the members to be 8-byte aligned for 64-bit content and at
395     // least 4-byte aligned for 32-bit content.  Opt for the larger encoding
396     // uniformly.  This matches the behaviour with cctools and ensures that ld64
397     // is happy with archives that we generate.
398     unsigned MemberPadding = Kind == object::Archive::K_DARWIN
399                                  ? OffsetToAlignment(Data.size(), 8)
400                                  : 0;
401     unsigned TailPadding = OffsetToAlignment(Data.size() + MemberPadding, 2);
402     StringRef Padding = StringRef(PaddingData, MemberPadding + TailPadding);
403 
404     printMemberHeader(Out, Pos, StringTable, Kind, Thin, ArcName, M,
405                       Buf.getBufferSize() + MemberPadding);
406     Out.flush();
407 
408     Expected<std::vector<unsigned>> Symbols =
409         getSymbols(Buf, SymNames, HasObject);
410     if (auto E = Symbols.takeError())
411       return std::move(E);
412 
413     Pos += Header.size() + Data.size() + Padding.size();
414     Ret.push_back({std::move(*Symbols), std::move(Header), Data, Padding});
415   }
416   // If there are no symbols, emit an empty symbol table, to satisfy Solaris
417   // tools, older versions of which expect a symbol table in a non-empty
418   // archive, regardless of whether there are any symbols in it.
419   if (HasObject && SymNames.tell() == 0)
420     SymNames << '\0' << '\0' << '\0';
421   return Ret;
422 }
423 
424 Error llvm::writeArchive(StringRef ArcName,
425                          ArrayRef<NewArchiveMember> NewMembers,
426                          bool WriteSymtab, object::Archive::Kind Kind,
427                          bool Deterministic, bool Thin,
428                          std::unique_ptr<MemoryBuffer> OldArchiveBuf) {
429   assert((!Thin || !isBSDLike(Kind)) && "Only the gnu format has a thin mode");
430 
431   SmallString<0> SymNamesBuf;
432   raw_svector_ostream SymNames(SymNamesBuf);
433   SmallString<0> StringTableBuf;
434   raw_svector_ostream StringTable(StringTableBuf);
435 
436   Expected<std::vector<MemberData>> DataOrErr =
437       computeMemberData(StringTable, SymNames, Kind, Thin, ArcName, NewMembers);
438   if (Error E = DataOrErr.takeError())
439     return E;
440   std::vector<MemberData> &Data = *DataOrErr;
441 
442   if (!StringTableBuf.empty())
443     Data.insert(Data.begin(), computeStringTable(StringTableBuf));
444 
445   SmallString<128> TmpArchive;
446   int TmpArchiveFD;
447   if (auto EC = sys::fs::createUniqueFile(ArcName + ".temp-archive-%%%%%%%.a",
448                                           TmpArchiveFD, TmpArchive))
449     return errorCodeToError(EC);
450 
451   ToolOutputFile Output(TmpArchive, TmpArchiveFD);
452   raw_fd_ostream &Out = Output.os();
453   if (Thin)
454     Out << "!<thin>\n";
455   else
456     Out << "!<arch>\n";
457 
458   if (WriteSymtab)
459     writeSymbolTable(Out, Kind, Deterministic, Data, SymNamesBuf);
460 
461   for (const MemberData &M : Data)
462     Out << M.Header << M.Data << M.Padding;
463 
464   Output.keep();
465   Out.close();
466 
467   // At this point, we no longer need whatever backing memory
468   // was used to generate the NewMembers. On Windows, this buffer
469   // could be a mapped view of the file we want to replace (if
470   // we're updating an existing archive, say). In that case, the
471   // rename would still succeed, but it would leave behind a
472   // temporary file (actually the original file renamed) because
473   // a file cannot be deleted while there's a handle open on it,
474   // only renamed. So by freeing this buffer, this ensures that
475   // the last open handle on the destination file, if any, is
476   // closed before we attempt to rename.
477   OldArchiveBuf.reset();
478 
479   sys::fs::rename(TmpArchive, ArcName);
480   return Error::success();
481 }
482