1 //===- Object.cpp ---------------------------------------------------------===//
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 #include "Object.h"
10 #include "llvm-objcopy.h"
11 #include "llvm/ADT/ArrayRef.h"
12 #include "llvm/ADT/STLExtras.h"
13 #include "llvm/ADT/StringRef.h"
14 #include "llvm/ADT/Twine.h"
15 #include "llvm/ADT/iterator_range.h"
16 #include "llvm/BinaryFormat/ELF.h"
17 #include "llvm/MC/MCTargetOptions.h"
18 #include "llvm/Object/ELFObjectFile.h"
19 #include "llvm/Support/Compression.h"
20 #include "llvm/Support/Endian.h"
21 #include "llvm/Support/ErrorHandling.h"
22 #include "llvm/Support/FileOutputBuffer.h"
23 #include "llvm/Support/Path.h"
24 #include <algorithm>
25 #include <cstddef>
26 #include <cstdint>
27 #include <iterator>
28 #include <unordered_set>
29 #include <utility>
30 #include <vector>
31 
32 namespace llvm {
33 namespace objcopy {
34 namespace elf {
35 
36 using namespace object;
37 using namespace ELF;
38 
39 template <class ELFT> void ELFWriter<ELFT>::writePhdr(const Segment &Seg) {
40   uint8_t *B = Buf.getBufferStart() + Obj.ProgramHdrSegment.Offset +
41                Seg.Index * sizeof(Elf_Phdr);
42   Elf_Phdr &Phdr = *reinterpret_cast<Elf_Phdr *>(B);
43   Phdr.p_type = Seg.Type;
44   Phdr.p_flags = Seg.Flags;
45   Phdr.p_offset = Seg.Offset;
46   Phdr.p_vaddr = Seg.VAddr;
47   Phdr.p_paddr = Seg.PAddr;
48   Phdr.p_filesz = Seg.FileSize;
49   Phdr.p_memsz = Seg.MemSize;
50   Phdr.p_align = Seg.Align;
51 }
52 
53 Error SectionBase::removeSectionReferences(
54     bool AllowBrokenLinks,
55     function_ref<bool(const SectionBase *)> ToRemove) {
56   return Error::success();
57 }
58 
59 Error SectionBase::removeSymbols(function_ref<bool(const Symbol &)> ToRemove) {
60   return Error::success();
61 }
62 
63 void SectionBase::initialize(SectionTableRef SecTable) {}
64 void SectionBase::finalize() {}
65 void SectionBase::markSymbols() {}
66 void SectionBase::replaceSectionReferences(
67     const DenseMap<SectionBase *, SectionBase *> &) {}
68 
69 template <class ELFT> void ELFWriter<ELFT>::writeShdr(const SectionBase &Sec) {
70   uint8_t *B = Buf.getBufferStart() + Sec.HeaderOffset;
71   Elf_Shdr &Shdr = *reinterpret_cast<Elf_Shdr *>(B);
72   Shdr.sh_name = Sec.NameIndex;
73   Shdr.sh_type = Sec.Type;
74   Shdr.sh_flags = Sec.Flags;
75   Shdr.sh_addr = Sec.Addr;
76   Shdr.sh_offset = Sec.Offset;
77   Shdr.sh_size = Sec.Size;
78   Shdr.sh_link = Sec.Link;
79   Shdr.sh_info = Sec.Info;
80   Shdr.sh_addralign = Sec.Align;
81   Shdr.sh_entsize = Sec.EntrySize;
82 }
83 
84 template <class ELFT> void ELFSectionSizer<ELFT>::visit(Section &Sec) {}
85 
86 template <class ELFT>
87 void ELFSectionSizer<ELFT>::visit(OwnedDataSection &Sec) {}
88 
89 template <class ELFT>
90 void ELFSectionSizer<ELFT>::visit(StringTableSection &Sec) {}
91 
92 template <class ELFT>
93 void ELFSectionSizer<ELFT>::visit(DynamicRelocationSection &Sec) {}
94 
95 template <class ELFT>
96 void ELFSectionSizer<ELFT>::visit(SymbolTableSection &Sec) {
97   Sec.EntrySize = sizeof(Elf_Sym);
98   Sec.Size = Sec.Symbols.size() * Sec.EntrySize;
99   // Align to the largest field in Elf_Sym.
100   Sec.Align = ELFT::Is64Bits ? sizeof(Elf_Xword) : sizeof(Elf_Word);
101 }
102 
103 template <class ELFT>
104 void ELFSectionSizer<ELFT>::visit(RelocationSection &Sec) {
105   Sec.EntrySize = Sec.Type == SHT_REL ? sizeof(Elf_Rel) : sizeof(Elf_Rela);
106   Sec.Size = Sec.Relocations.size() * Sec.EntrySize;
107   // Align to the largest field in Elf_Rel(a).
108   Sec.Align = ELFT::Is64Bits ? sizeof(Elf_Xword) : sizeof(Elf_Word);
109 }
110 
111 template <class ELFT>
112 void ELFSectionSizer<ELFT>::visit(GnuDebugLinkSection &Sec) {}
113 
114 template <class ELFT> void ELFSectionSizer<ELFT>::visit(GroupSection &Sec) {}
115 
116 template <class ELFT>
117 void ELFSectionSizer<ELFT>::visit(SectionIndexSection &Sec) {}
118 
119 template <class ELFT>
120 void ELFSectionSizer<ELFT>::visit(CompressedSection &Sec) {}
121 
122 template <class ELFT>
123 void ELFSectionSizer<ELFT>::visit(DecompressedSection &Sec) {}
124 
125 void BinarySectionWriter::visit(const SectionIndexSection &Sec) {
126   error("cannot write symbol section index table '" + Sec.Name + "' ");
127 }
128 
129 void BinarySectionWriter::visit(const SymbolTableSection &Sec) {
130   error("cannot write symbol table '" + Sec.Name + "' out to binary");
131 }
132 
133 void BinarySectionWriter::visit(const RelocationSection &Sec) {
134   error("cannot write relocation section '" + Sec.Name + "' out to binary");
135 }
136 
137 void BinarySectionWriter::visit(const GnuDebugLinkSection &Sec) {
138   error("cannot write '" + Sec.Name + "' out to binary");
139 }
140 
141 void BinarySectionWriter::visit(const GroupSection &Sec) {
142   error("cannot write '" + Sec.Name + "' out to binary");
143 }
144 
145 void SectionWriter::visit(const Section &Sec) {
146   if (Sec.Type != SHT_NOBITS)
147     llvm::copy(Sec.Contents, Out.getBufferStart() + Sec.Offset);
148 }
149 
150 static bool addressOverflows32bit(uint64_t Addr) {
151   // Sign extended 32 bit addresses (e.g 0xFFFFFFFF80000000) are ok
152   return Addr > UINT32_MAX && Addr + 0x80000000 > UINT32_MAX;
153 }
154 
155 template <class T> static T checkedGetHex(StringRef S) {
156   T Value;
157   bool Fail = S.getAsInteger(16, Value);
158   assert(!Fail);
159   (void)Fail;
160   return Value;
161 }
162 
163 // Fills exactly Len bytes of buffer with hexadecimal characters
164 // representing value 'X'
165 template <class T, class Iterator>
166 static Iterator utohexstr(T X, Iterator It, size_t Len) {
167   // Fill range with '0'
168   std::fill(It, It + Len, '0');
169 
170   for (long I = Len - 1; I >= 0; --I) {
171     unsigned char Mod = static_cast<unsigned char>(X) & 15;
172     *(It + I) = hexdigit(Mod, false);
173     X >>= 4;
174   }
175   assert(X == 0);
176   return It + Len;
177 }
178 
179 uint8_t IHexRecord::getChecksum(StringRef S) {
180   assert((S.size() & 1) == 0);
181   uint8_t Checksum = 0;
182   while (!S.empty()) {
183     Checksum += checkedGetHex<uint8_t>(S.take_front(2));
184     S = S.drop_front(2);
185   }
186   return -Checksum;
187 }
188 
189 IHexLineData IHexRecord::getLine(uint8_t Type, uint16_t Addr,
190                                  ArrayRef<uint8_t> Data) {
191   IHexLineData Line(getLineLength(Data.size()));
192   assert(Line.size());
193   auto Iter = Line.begin();
194   *Iter++ = ':';
195   Iter = utohexstr(Data.size(), Iter, 2);
196   Iter = utohexstr(Addr, Iter, 4);
197   Iter = utohexstr(Type, Iter, 2);
198   for (uint8_t X : Data)
199     Iter = utohexstr(X, Iter, 2);
200   StringRef S(Line.data() + 1, std::distance(Line.begin() + 1, Iter));
201   Iter = utohexstr(getChecksum(S), Iter, 2);
202   *Iter++ = '\r';
203   *Iter++ = '\n';
204   assert(Iter == Line.end());
205   return Line;
206 }
207 
208 static Error checkRecord(const IHexRecord &R) {
209   switch (R.Type) {
210   case IHexRecord::Data:
211     if (R.HexData.size() == 0)
212       return createStringError(
213           errc::invalid_argument,
214           "zero data length is not allowed for data records");
215     break;
216   case IHexRecord::EndOfFile:
217     break;
218   case IHexRecord::SegmentAddr:
219     // 20-bit segment address. Data length must be 2 bytes
220     // (4 bytes in hex)
221     if (R.HexData.size() != 4)
222       return createStringError(
223           errc::invalid_argument,
224           "segment address data should be 2 bytes in size");
225     break;
226   case IHexRecord::StartAddr80x86:
227   case IHexRecord::StartAddr:
228     if (R.HexData.size() != 8)
229       return createStringError(errc::invalid_argument,
230                                "start address data should be 4 bytes in size");
231     // According to Intel HEX specification '03' record
232     // only specifies the code address within the 20-bit
233     // segmented address space of the 8086/80186. This
234     // means 12 high order bits should be zeroes.
235     if (R.Type == IHexRecord::StartAddr80x86 &&
236         R.HexData.take_front(3) != "000")
237       return createStringError(errc::invalid_argument,
238                                "start address exceeds 20 bit for 80x86");
239     break;
240   case IHexRecord::ExtendedAddr:
241     // 16-31 bits of linear base address
242     if (R.HexData.size() != 4)
243       return createStringError(
244           errc::invalid_argument,
245           "extended address data should be 2 bytes in size");
246     break;
247   default:
248     // Unknown record type
249     return createStringError(errc::invalid_argument, "unknown record type: %u",
250                              static_cast<unsigned>(R.Type));
251   }
252   return Error::success();
253 }
254 
255 // Checks that IHEX line contains valid characters.
256 // This allows converting hexadecimal data to integers
257 // without extra verification.
258 static Error checkChars(StringRef Line) {
259   assert(!Line.empty());
260   if (Line[0] != ':')
261     return createStringError(errc::invalid_argument,
262                              "missing ':' in the beginning of line.");
263 
264   for (size_t Pos = 1; Pos < Line.size(); ++Pos)
265     if (hexDigitValue(Line[Pos]) == -1U)
266       return createStringError(errc::invalid_argument,
267                                "invalid character at position %zu.", Pos + 1);
268   return Error::success();
269 }
270 
271 Expected<IHexRecord> IHexRecord::parse(StringRef Line) {
272   assert(!Line.empty());
273 
274   // ':' + Length + Address + Type + Checksum with empty data ':LLAAAATTCC'
275   if (Line.size() < 11)
276     return createStringError(errc::invalid_argument,
277                              "line is too short: %zu chars.", Line.size());
278 
279   if (Error E = checkChars(Line))
280     return std::move(E);
281 
282   IHexRecord Rec;
283   size_t DataLen = checkedGetHex<uint8_t>(Line.substr(1, 2));
284   if (Line.size() != getLength(DataLen))
285     return createStringError(errc::invalid_argument,
286                              "invalid line length %zu (should be %zu)",
287                              Line.size(), getLength(DataLen));
288 
289   Rec.Addr = checkedGetHex<uint16_t>(Line.substr(3, 4));
290   Rec.Type = checkedGetHex<uint8_t>(Line.substr(7, 2));
291   Rec.HexData = Line.substr(9, DataLen * 2);
292 
293   if (getChecksum(Line.drop_front(1)) != 0)
294     return createStringError(errc::invalid_argument, "incorrect checksum.");
295   if (Error E = checkRecord(Rec))
296     return std::move(E);
297   return Rec;
298 }
299 
300 static uint64_t sectionPhysicalAddr(const SectionBase *Sec) {
301   Segment *Seg = Sec->ParentSegment;
302   if (Seg && Seg->Type != ELF::PT_LOAD)
303     Seg = nullptr;
304   return Seg ? Seg->PAddr + Sec->OriginalOffset - Seg->OriginalOffset
305              : Sec->Addr;
306 }
307 
308 void IHexSectionWriterBase::writeSection(const SectionBase *Sec,
309                                          ArrayRef<uint8_t> Data) {
310   assert(Data.size() == Sec->Size);
311   const uint32_t ChunkSize = 16;
312   uint32_t Addr = sectionPhysicalAddr(Sec) & 0xFFFFFFFFU;
313   while (!Data.empty()) {
314     uint64_t DataSize = std::min<uint64_t>(Data.size(), ChunkSize);
315     if (Addr > SegmentAddr + BaseAddr + 0xFFFFU) {
316       if (Addr > 0xFFFFFU) {
317         // Write extended address record, zeroing segment address
318         // if needed.
319         if (SegmentAddr != 0)
320           SegmentAddr = writeSegmentAddr(0U);
321         BaseAddr = writeBaseAddr(Addr);
322       } else {
323         // We can still remain 16-bit
324         SegmentAddr = writeSegmentAddr(Addr);
325       }
326     }
327     uint64_t SegOffset = Addr - BaseAddr - SegmentAddr;
328     assert(SegOffset <= 0xFFFFU);
329     DataSize = std::min(DataSize, 0x10000U - SegOffset);
330     writeData(0, SegOffset, Data.take_front(DataSize));
331     Addr += DataSize;
332     Data = Data.drop_front(DataSize);
333   }
334 }
335 
336 uint64_t IHexSectionWriterBase::writeSegmentAddr(uint64_t Addr) {
337   assert(Addr <= 0xFFFFFU);
338   uint8_t Data[] = {static_cast<uint8_t>((Addr & 0xF0000U) >> 12), 0};
339   writeData(2, 0, Data);
340   return Addr & 0xF0000U;
341 }
342 
343 uint64_t IHexSectionWriterBase::writeBaseAddr(uint64_t Addr) {
344   assert(Addr <= 0xFFFFFFFFU);
345   uint64_t Base = Addr & 0xFFFF0000U;
346   uint8_t Data[] = {static_cast<uint8_t>(Base >> 24),
347                     static_cast<uint8_t>((Base >> 16) & 0xFF)};
348   writeData(4, 0, Data);
349   return Base;
350 }
351 
352 void IHexSectionWriterBase::writeData(uint8_t Type, uint16_t Addr,
353                                       ArrayRef<uint8_t> Data) {
354   Offset += IHexRecord::getLineLength(Data.size());
355 }
356 
357 void IHexSectionWriterBase::visit(const Section &Sec) {
358   writeSection(&Sec, Sec.Contents);
359 }
360 
361 void IHexSectionWriterBase::visit(const OwnedDataSection &Sec) {
362   writeSection(&Sec, Sec.Data);
363 }
364 
365 void IHexSectionWriterBase::visit(const StringTableSection &Sec) {
366   // Check that sizer has already done its work
367   assert(Sec.Size == Sec.StrTabBuilder.getSize());
368   // We are free to pass an invalid pointer to writeSection as long
369   // as we don't actually write any data. The real writer class has
370   // to override this method .
371   writeSection(&Sec, {nullptr, static_cast<size_t>(Sec.Size)});
372 }
373 
374 void IHexSectionWriterBase::visit(const DynamicRelocationSection &Sec) {
375   writeSection(&Sec, Sec.Contents);
376 }
377 
378 void IHexSectionWriter::writeData(uint8_t Type, uint16_t Addr,
379                                   ArrayRef<uint8_t> Data) {
380   IHexLineData HexData = IHexRecord::getLine(Type, Addr, Data);
381   memcpy(Out.getBufferStart() + Offset, HexData.data(), HexData.size());
382   Offset += HexData.size();
383 }
384 
385 void IHexSectionWriter::visit(const StringTableSection &Sec) {
386   assert(Sec.Size == Sec.StrTabBuilder.getSize());
387   std::vector<uint8_t> Data(Sec.Size);
388   Sec.StrTabBuilder.write(Data.data());
389   writeSection(&Sec, Data);
390 }
391 
392 void Section::accept(SectionVisitor &Visitor) const { Visitor.visit(*this); }
393 
394 void Section::accept(MutableSectionVisitor &Visitor) { Visitor.visit(*this); }
395 
396 void SectionWriter::visit(const OwnedDataSection &Sec) {
397   llvm::copy(Sec.Data, Out.getBufferStart() + Sec.Offset);
398 }
399 
400 static constexpr std::array<uint8_t, 4> ZlibGnuMagic = {{'Z', 'L', 'I', 'B'}};
401 
402 static bool isDataGnuCompressed(ArrayRef<uint8_t> Data) {
403   return Data.size() > ZlibGnuMagic.size() &&
404          std::equal(ZlibGnuMagic.begin(), ZlibGnuMagic.end(), Data.data());
405 }
406 
407 template <class ELFT>
408 static std::tuple<uint64_t, uint64_t>
409 getDecompressedSizeAndAlignment(ArrayRef<uint8_t> Data) {
410   const bool IsGnuDebug = isDataGnuCompressed(Data);
411   const uint64_t DecompressedSize =
412       IsGnuDebug
413           ? support::endian::read64be(Data.data() + ZlibGnuMagic.size())
414           : reinterpret_cast<const Elf_Chdr_Impl<ELFT> *>(Data.data())->ch_size;
415   const uint64_t DecompressedAlign =
416       IsGnuDebug ? 1
417                  : reinterpret_cast<const Elf_Chdr_Impl<ELFT> *>(Data.data())
418                        ->ch_addralign;
419 
420   return std::make_tuple(DecompressedSize, DecompressedAlign);
421 }
422 
423 template <class ELFT>
424 void ELFSectionWriter<ELFT>::visit(const DecompressedSection &Sec) {
425   const size_t DataOffset = isDataGnuCompressed(Sec.OriginalData)
426                                 ? (ZlibGnuMagic.size() + sizeof(Sec.Size))
427                                 : sizeof(Elf_Chdr_Impl<ELFT>);
428 
429   StringRef CompressedContent(
430       reinterpret_cast<const char *>(Sec.OriginalData.data()) + DataOffset,
431       Sec.OriginalData.size() - DataOffset);
432 
433   SmallVector<char, 128> DecompressedContent;
434   if (Error E = zlib::uncompress(CompressedContent, DecompressedContent,
435                                  static_cast<size_t>(Sec.Size)))
436     reportError(Sec.Name, std::move(E));
437 
438   uint8_t *Buf = Out.getBufferStart() + Sec.Offset;
439   std::copy(DecompressedContent.begin(), DecompressedContent.end(), Buf);
440 }
441 
442 void BinarySectionWriter::visit(const DecompressedSection &Sec) {
443   error("cannot write compressed section '" + Sec.Name + "' ");
444 }
445 
446 void DecompressedSection::accept(SectionVisitor &Visitor) const {
447   Visitor.visit(*this);
448 }
449 
450 void DecompressedSection::accept(MutableSectionVisitor &Visitor) {
451   Visitor.visit(*this);
452 }
453 
454 void OwnedDataSection::accept(SectionVisitor &Visitor) const {
455   Visitor.visit(*this);
456 }
457 
458 void OwnedDataSection::accept(MutableSectionVisitor &Visitor) {
459   Visitor.visit(*this);
460 }
461 
462 void OwnedDataSection::appendHexData(StringRef HexData) {
463   assert((HexData.size() & 1) == 0);
464   while (!HexData.empty()) {
465     Data.push_back(checkedGetHex<uint8_t>(HexData.take_front(2)));
466     HexData = HexData.drop_front(2);
467   }
468   Size = Data.size();
469 }
470 
471 void BinarySectionWriter::visit(const CompressedSection &Sec) {
472   error("cannot write compressed section '" + Sec.Name + "' ");
473 }
474 
475 template <class ELFT>
476 void ELFSectionWriter<ELFT>::visit(const CompressedSection &Sec) {
477   uint8_t *Buf = Out.getBufferStart() + Sec.Offset;
478   if (Sec.CompressionType == DebugCompressionType::None) {
479     std::copy(Sec.OriginalData.begin(), Sec.OriginalData.end(), Buf);
480     return;
481   }
482 
483   if (Sec.CompressionType == DebugCompressionType::GNU) {
484     const char *Magic = "ZLIB";
485     memcpy(Buf, Magic, strlen(Magic));
486     Buf += strlen(Magic);
487     const uint64_t DecompressedSize =
488         support::endian::read64be(&Sec.DecompressedSize);
489     memcpy(Buf, &DecompressedSize, sizeof(DecompressedSize));
490     Buf += sizeof(DecompressedSize);
491   } else {
492     Elf_Chdr_Impl<ELFT> Chdr;
493     Chdr.ch_type = ELF::ELFCOMPRESS_ZLIB;
494     Chdr.ch_size = Sec.DecompressedSize;
495     Chdr.ch_addralign = Sec.DecompressedAlign;
496     memcpy(Buf, &Chdr, sizeof(Chdr));
497     Buf += sizeof(Chdr);
498   }
499 
500   std::copy(Sec.CompressedData.begin(), Sec.CompressedData.end(), Buf);
501 }
502 
503 CompressedSection::CompressedSection(const SectionBase &Sec,
504                                      DebugCompressionType CompressionType)
505     : SectionBase(Sec), CompressionType(CompressionType),
506       DecompressedSize(Sec.OriginalData.size()), DecompressedAlign(Sec.Align) {
507   if (Error E = zlib::compress(
508           StringRef(reinterpret_cast<const char *>(OriginalData.data()),
509                     OriginalData.size()),
510           CompressedData))
511     reportError(Name, std::move(E));
512 
513   size_t ChdrSize;
514   if (CompressionType == DebugCompressionType::GNU) {
515     Name = ".z" + Sec.Name.substr(1);
516     ChdrSize = sizeof("ZLIB") - 1 + sizeof(uint64_t);
517   } else {
518     Flags |= ELF::SHF_COMPRESSED;
519     ChdrSize =
520         std::max(std::max(sizeof(object::Elf_Chdr_Impl<object::ELF64LE>),
521                           sizeof(object::Elf_Chdr_Impl<object::ELF64BE>)),
522                  std::max(sizeof(object::Elf_Chdr_Impl<object::ELF32LE>),
523                           sizeof(object::Elf_Chdr_Impl<object::ELF32BE>)));
524   }
525   Size = ChdrSize + CompressedData.size();
526   Align = 8;
527 }
528 
529 CompressedSection::CompressedSection(ArrayRef<uint8_t> CompressedData,
530                                      uint64_t DecompressedSize,
531                                      uint64_t DecompressedAlign)
532     : CompressionType(DebugCompressionType::None),
533       DecompressedSize(DecompressedSize), DecompressedAlign(DecompressedAlign) {
534   OriginalData = CompressedData;
535 }
536 
537 void CompressedSection::accept(SectionVisitor &Visitor) const {
538   Visitor.visit(*this);
539 }
540 
541 void CompressedSection::accept(MutableSectionVisitor &Visitor) {
542   Visitor.visit(*this);
543 }
544 
545 void StringTableSection::addString(StringRef Name) { StrTabBuilder.add(Name); }
546 
547 uint32_t StringTableSection::findIndex(StringRef Name) const {
548   return StrTabBuilder.getOffset(Name);
549 }
550 
551 void StringTableSection::prepareForLayout() {
552   StrTabBuilder.finalize();
553   Size = StrTabBuilder.getSize();
554 }
555 
556 void SectionWriter::visit(const StringTableSection &Sec) {
557   Sec.StrTabBuilder.write(Out.getBufferStart() + Sec.Offset);
558 }
559 
560 void StringTableSection::accept(SectionVisitor &Visitor) const {
561   Visitor.visit(*this);
562 }
563 
564 void StringTableSection::accept(MutableSectionVisitor &Visitor) {
565   Visitor.visit(*this);
566 }
567 
568 template <class ELFT>
569 void ELFSectionWriter<ELFT>::visit(const SectionIndexSection &Sec) {
570   uint8_t *Buf = Out.getBufferStart() + Sec.Offset;
571   llvm::copy(Sec.Indexes, reinterpret_cast<Elf_Word *>(Buf));
572 }
573 
574 void SectionIndexSection::initialize(SectionTableRef SecTable) {
575   Size = 0;
576   setSymTab(SecTable.getSectionOfType<SymbolTableSection>(
577       Link,
578       "Link field value " + Twine(Link) + " in section " + Name + " is invalid",
579       "Link field value " + Twine(Link) + " in section " + Name +
580           " is not a symbol table"));
581   Symbols->setShndxTable(this);
582 }
583 
584 void SectionIndexSection::finalize() { Link = Symbols->Index; }
585 
586 void SectionIndexSection::accept(SectionVisitor &Visitor) const {
587   Visitor.visit(*this);
588 }
589 
590 void SectionIndexSection::accept(MutableSectionVisitor &Visitor) {
591   Visitor.visit(*this);
592 }
593 
594 static bool isValidReservedSectionIndex(uint16_t Index, uint16_t Machine) {
595   switch (Index) {
596   case SHN_ABS:
597   case SHN_COMMON:
598     return true;
599   }
600 
601   if (Machine == EM_AMDGPU) {
602     return Index == SHN_AMDGPU_LDS;
603   }
604 
605   if (Machine == EM_HEXAGON) {
606     switch (Index) {
607     case SHN_HEXAGON_SCOMMON:
608     case SHN_HEXAGON_SCOMMON_2:
609     case SHN_HEXAGON_SCOMMON_4:
610     case SHN_HEXAGON_SCOMMON_8:
611       return true;
612     }
613   }
614   return false;
615 }
616 
617 // Large indexes force us to clarify exactly what this function should do. This
618 // function should return the value that will appear in st_shndx when written
619 // out.
620 uint16_t Symbol::getShndx() const {
621   if (DefinedIn != nullptr) {
622     if (DefinedIn->Index >= SHN_LORESERVE)
623       return SHN_XINDEX;
624     return DefinedIn->Index;
625   }
626 
627   if (ShndxType == SYMBOL_SIMPLE_INDEX) {
628     // This means that we don't have a defined section but we do need to
629     // output a legitimate section index.
630     return SHN_UNDEF;
631   }
632 
633   assert(ShndxType == SYMBOL_ABS || ShndxType == SYMBOL_COMMON ||
634          (ShndxType >= SYMBOL_LOPROC && ShndxType <= SYMBOL_HIPROC) ||
635          (ShndxType >= SYMBOL_LOOS && ShndxType <= SYMBOL_HIOS));
636   return static_cast<uint16_t>(ShndxType);
637 }
638 
639 bool Symbol::isCommon() const { return getShndx() == SHN_COMMON; }
640 
641 void SymbolTableSection::assignIndices() {
642   uint32_t Index = 0;
643   for (auto &Sym : Symbols)
644     Sym->Index = Index++;
645 }
646 
647 void SymbolTableSection::addSymbol(Twine Name, uint8_t Bind, uint8_t Type,
648                                    SectionBase *DefinedIn, uint64_t Value,
649                                    uint8_t Visibility, uint16_t Shndx,
650                                    uint64_t SymbolSize) {
651   Symbol Sym;
652   Sym.Name = Name.str();
653   Sym.Binding = Bind;
654   Sym.Type = Type;
655   Sym.DefinedIn = DefinedIn;
656   if (DefinedIn != nullptr)
657     DefinedIn->HasSymbol = true;
658   if (DefinedIn == nullptr) {
659     if (Shndx >= SHN_LORESERVE)
660       Sym.ShndxType = static_cast<SymbolShndxType>(Shndx);
661     else
662       Sym.ShndxType = SYMBOL_SIMPLE_INDEX;
663   }
664   Sym.Value = Value;
665   Sym.Visibility = Visibility;
666   Sym.Size = SymbolSize;
667   Sym.Index = Symbols.size();
668   Symbols.emplace_back(std::make_unique<Symbol>(Sym));
669   Size += this->EntrySize;
670 }
671 
672 Error SymbolTableSection::removeSectionReferences(
673     bool AllowBrokenLinks,
674     function_ref<bool(const SectionBase *)> ToRemove) {
675   if (ToRemove(SectionIndexTable))
676     SectionIndexTable = nullptr;
677   if (ToRemove(SymbolNames)) {
678     if (!AllowBrokenLinks)
679       return createStringError(
680           llvm::errc::invalid_argument,
681           "string table '%s' cannot be removed because it is "
682           "referenced by the symbol table '%s'",
683           SymbolNames->Name.data(), this->Name.data());
684     SymbolNames = nullptr;
685   }
686   return removeSymbols(
687       [ToRemove](const Symbol &Sym) { return ToRemove(Sym.DefinedIn); });
688 }
689 
690 void SymbolTableSection::updateSymbols(function_ref<void(Symbol &)> Callable) {
691   std::for_each(std::begin(Symbols) + 1, std::end(Symbols),
692                 [Callable](SymPtr &Sym) { Callable(*Sym); });
693   std::stable_partition(
694       std::begin(Symbols), std::end(Symbols),
695       [](const SymPtr &Sym) { return Sym->Binding == STB_LOCAL; });
696   assignIndices();
697 }
698 
699 Error SymbolTableSection::removeSymbols(
700     function_ref<bool(const Symbol &)> ToRemove) {
701   Symbols.erase(
702       std::remove_if(std::begin(Symbols) + 1, std::end(Symbols),
703                      [ToRemove](const SymPtr &Sym) { return ToRemove(*Sym); }),
704       std::end(Symbols));
705   Size = Symbols.size() * EntrySize;
706   assignIndices();
707   return Error::success();
708 }
709 
710 void SymbolTableSection::replaceSectionReferences(
711     const DenseMap<SectionBase *, SectionBase *> &FromTo) {
712   for (std::unique_ptr<Symbol> &Sym : Symbols)
713     if (SectionBase *To = FromTo.lookup(Sym->DefinedIn))
714       Sym->DefinedIn = To;
715 }
716 
717 void SymbolTableSection::initialize(SectionTableRef SecTable) {
718   Size = 0;
719   setStrTab(SecTable.getSectionOfType<StringTableSection>(
720       Link,
721       "Symbol table has link index of " + Twine(Link) +
722           " which is not a valid index",
723       "Symbol table has link index of " + Twine(Link) +
724           " which is not a string table"));
725 }
726 
727 void SymbolTableSection::finalize() {
728   uint32_t MaxLocalIndex = 0;
729   for (std::unique_ptr<Symbol> &Sym : Symbols) {
730     Sym->NameIndex =
731         SymbolNames == nullptr ? 0 : SymbolNames->findIndex(Sym->Name);
732     if (Sym->Binding == STB_LOCAL)
733       MaxLocalIndex = std::max(MaxLocalIndex, Sym->Index);
734   }
735   // Now we need to set the Link and Info fields.
736   Link = SymbolNames == nullptr ? 0 : SymbolNames->Index;
737   Info = MaxLocalIndex + 1;
738 }
739 
740 void SymbolTableSection::prepareForLayout() {
741   // Reserve proper amount of space in section index table, so we can
742   // layout sections correctly. We will fill the table with correct
743   // indexes later in fillShdnxTable.
744   if (SectionIndexTable)
745     SectionIndexTable->reserve(Symbols.size());
746 
747   // Add all of our strings to SymbolNames so that SymbolNames has the right
748   // size before layout is decided.
749   // If the symbol names section has been removed, don't try to add strings to
750   // the table.
751   if (SymbolNames != nullptr)
752     for (std::unique_ptr<Symbol> &Sym : Symbols)
753       SymbolNames->addString(Sym->Name);
754 }
755 
756 void SymbolTableSection::fillShndxTable() {
757   if (SectionIndexTable == nullptr)
758     return;
759   // Fill section index table with real section indexes. This function must
760   // be called after assignOffsets.
761   for (const std::unique_ptr<Symbol> &Sym : Symbols) {
762     if (Sym->DefinedIn != nullptr && Sym->DefinedIn->Index >= SHN_LORESERVE)
763       SectionIndexTable->addIndex(Sym->DefinedIn->Index);
764     else
765       SectionIndexTable->addIndex(SHN_UNDEF);
766   }
767 }
768 
769 const Symbol *SymbolTableSection::getSymbolByIndex(uint32_t Index) const {
770   if (Symbols.size() <= Index)
771     error("invalid symbol index: " + Twine(Index));
772   return Symbols[Index].get();
773 }
774 
775 Symbol *SymbolTableSection::getSymbolByIndex(uint32_t Index) {
776   return const_cast<Symbol *>(
777       static_cast<const SymbolTableSection *>(this)->getSymbolByIndex(Index));
778 }
779 
780 template <class ELFT>
781 void ELFSectionWriter<ELFT>::visit(const SymbolTableSection &Sec) {
782   Elf_Sym *Sym = reinterpret_cast<Elf_Sym *>(Out.getBufferStart() + Sec.Offset);
783   // Loop though symbols setting each entry of the symbol table.
784   for (const std::unique_ptr<Symbol> &Symbol : Sec.Symbols) {
785     Sym->st_name = Symbol->NameIndex;
786     Sym->st_value = Symbol->Value;
787     Sym->st_size = Symbol->Size;
788     Sym->st_other = Symbol->Visibility;
789     Sym->setBinding(Symbol->Binding);
790     Sym->setType(Symbol->Type);
791     Sym->st_shndx = Symbol->getShndx();
792     ++Sym;
793   }
794 }
795 
796 void SymbolTableSection::accept(SectionVisitor &Visitor) const {
797   Visitor.visit(*this);
798 }
799 
800 void SymbolTableSection::accept(MutableSectionVisitor &Visitor) {
801   Visitor.visit(*this);
802 }
803 
804 Error RelocationSection::removeSectionReferences(
805     bool AllowBrokenLinks,
806     function_ref<bool(const SectionBase *)> ToRemove) {
807   if (ToRemove(Symbols)) {
808     if (!AllowBrokenLinks)
809       return createStringError(
810           llvm::errc::invalid_argument,
811           "symbol table '%s' cannot be removed because it is "
812           "referenced by the relocation section '%s'",
813           Symbols->Name.data(), this->Name.data());
814     Symbols = nullptr;
815   }
816 
817   for (const Relocation &R : Relocations) {
818     if (!R.RelocSymbol || !R.RelocSymbol->DefinedIn ||
819         !ToRemove(R.RelocSymbol->DefinedIn))
820       continue;
821     return createStringError(llvm::errc::invalid_argument,
822                              "section '%s' cannot be removed: (%s+0x%" PRIx64
823                              ") has relocation against symbol '%s'",
824                              R.RelocSymbol->DefinedIn->Name.data(),
825                              SecToApplyRel->Name.data(), R.Offset,
826                              R.RelocSymbol->Name.c_str());
827   }
828 
829   return Error::success();
830 }
831 
832 template <class SymTabType>
833 void RelocSectionWithSymtabBase<SymTabType>::initialize(
834     SectionTableRef SecTable) {
835   if (Link != SHN_UNDEF)
836     setSymTab(SecTable.getSectionOfType<SymTabType>(
837         Link,
838         "Link field value " + Twine(Link) + " in section " + Name +
839             " is invalid",
840         "Link field value " + Twine(Link) + " in section " + Name +
841             " is not a symbol table"));
842 
843   if (Info != SHN_UNDEF)
844     setSection(SecTable.getSection(Info, "Info field value " + Twine(Info) +
845                                              " in section " + Name +
846                                              " is invalid"));
847   else
848     setSection(nullptr);
849 }
850 
851 template <class SymTabType>
852 void RelocSectionWithSymtabBase<SymTabType>::finalize() {
853   this->Link = Symbols ? Symbols->Index : 0;
854 
855   if (SecToApplyRel != nullptr)
856     this->Info = SecToApplyRel->Index;
857 }
858 
859 template <class ELFT>
860 static void setAddend(Elf_Rel_Impl<ELFT, false> &Rel, uint64_t Addend) {}
861 
862 template <class ELFT>
863 static void setAddend(Elf_Rel_Impl<ELFT, true> &Rela, uint64_t Addend) {
864   Rela.r_addend = Addend;
865 }
866 
867 template <class RelRange, class T>
868 static void writeRel(const RelRange &Relocations, T *Buf) {
869   for (const auto &Reloc : Relocations) {
870     Buf->r_offset = Reloc.Offset;
871     setAddend(*Buf, Reloc.Addend);
872     Buf->setSymbolAndType(Reloc.RelocSymbol ? Reloc.RelocSymbol->Index : 0,
873                           Reloc.Type, false);
874     ++Buf;
875   }
876 }
877 
878 template <class ELFT>
879 void ELFSectionWriter<ELFT>::visit(const RelocationSection &Sec) {
880   uint8_t *Buf = Out.getBufferStart() + Sec.Offset;
881   if (Sec.Type == SHT_REL)
882     writeRel(Sec.Relocations, reinterpret_cast<Elf_Rel *>(Buf));
883   else
884     writeRel(Sec.Relocations, reinterpret_cast<Elf_Rela *>(Buf));
885 }
886 
887 void RelocationSection::accept(SectionVisitor &Visitor) const {
888   Visitor.visit(*this);
889 }
890 
891 void RelocationSection::accept(MutableSectionVisitor &Visitor) {
892   Visitor.visit(*this);
893 }
894 
895 Error RelocationSection::removeSymbols(
896     function_ref<bool(const Symbol &)> ToRemove) {
897   for (const Relocation &Reloc : Relocations)
898     if (Reloc.RelocSymbol && ToRemove(*Reloc.RelocSymbol))
899       return createStringError(
900           llvm::errc::invalid_argument,
901           "not stripping symbol '%s' because it is named in a relocation",
902           Reloc.RelocSymbol->Name.data());
903   return Error::success();
904 }
905 
906 void RelocationSection::markSymbols() {
907   for (const Relocation &Reloc : Relocations)
908     if (Reloc.RelocSymbol)
909       Reloc.RelocSymbol->Referenced = true;
910 }
911 
912 void RelocationSection::replaceSectionReferences(
913     const DenseMap<SectionBase *, SectionBase *> &FromTo) {
914   // Update the target section if it was replaced.
915   if (SectionBase *To = FromTo.lookup(SecToApplyRel))
916     SecToApplyRel = To;
917 }
918 
919 void SectionWriter::visit(const DynamicRelocationSection &Sec) {
920   llvm::copy(Sec.Contents, Out.getBufferStart() + Sec.Offset);
921 }
922 
923 void DynamicRelocationSection::accept(SectionVisitor &Visitor) const {
924   Visitor.visit(*this);
925 }
926 
927 void DynamicRelocationSection::accept(MutableSectionVisitor &Visitor) {
928   Visitor.visit(*this);
929 }
930 
931 Error DynamicRelocationSection::removeSectionReferences(
932     bool AllowBrokenLinks, function_ref<bool(const SectionBase *)> ToRemove) {
933   if (ToRemove(Symbols)) {
934     if (!AllowBrokenLinks)
935       return createStringError(
936           llvm::errc::invalid_argument,
937           "symbol table '%s' cannot be removed because it is "
938           "referenced by the relocation section '%s'",
939           Symbols->Name.data(), this->Name.data());
940     Symbols = nullptr;
941   }
942 
943   // SecToApplyRel contains a section referenced by sh_info field. It keeps
944   // a section to which the relocation section applies. When we remove any
945   // sections we also remove their relocation sections. Since we do that much
946   // earlier, this assert should never be triggered.
947   assert(!SecToApplyRel || !ToRemove(SecToApplyRel));
948   return Error::success();
949 }
950 
951 Error Section::removeSectionReferences(
952     bool AllowBrokenDependency,
953     function_ref<bool(const SectionBase *)> ToRemove) {
954   if (ToRemove(LinkSection)) {
955     if (!AllowBrokenDependency)
956       return createStringError(llvm::errc::invalid_argument,
957                                "section '%s' cannot be removed because it is "
958                                "referenced by the section '%s'",
959                                LinkSection->Name.data(), this->Name.data());
960     LinkSection = nullptr;
961   }
962   return Error::success();
963 }
964 
965 void GroupSection::finalize() {
966   this->Info = Sym->Index;
967   this->Link = SymTab->Index;
968 }
969 
970 Error GroupSection::removeSymbols(function_ref<bool(const Symbol &)> ToRemove) {
971   if (ToRemove(*Sym))
972     return createStringError(llvm::errc::invalid_argument,
973                              "symbol '%s' cannot be removed because it is "
974                              "referenced by the section '%s[%d]'",
975                              Sym->Name.data(), this->Name.data(), this->Index);
976   return Error::success();
977 }
978 
979 void GroupSection::markSymbols() {
980   if (Sym)
981     Sym->Referenced = true;
982 }
983 
984 void GroupSection::replaceSectionReferences(
985     const DenseMap<SectionBase *, SectionBase *> &FromTo) {
986   for (SectionBase *&Sec : GroupMembers)
987     if (SectionBase *To = FromTo.lookup(Sec))
988       Sec = To;
989 }
990 
991 void Section::initialize(SectionTableRef SecTable) {
992   if (Link == ELF::SHN_UNDEF)
993     return;
994   LinkSection =
995       SecTable.getSection(Link, "Link field value " + Twine(Link) +
996                                     " in section " + Name + " is invalid");
997   if (LinkSection->Type == ELF::SHT_SYMTAB)
998     LinkSection = nullptr;
999 }
1000 
1001 void Section::finalize() { this->Link = LinkSection ? LinkSection->Index : 0; }
1002 
1003 void GnuDebugLinkSection::init(StringRef File) {
1004   FileName = sys::path::filename(File);
1005   // The format for the .gnu_debuglink starts with the file name and is
1006   // followed by a null terminator and then the CRC32 of the file. The CRC32
1007   // should be 4 byte aligned. So we add the FileName size, a 1 for the null
1008   // byte, and then finally push the size to alignment and add 4.
1009   Size = alignTo(FileName.size() + 1, 4) + 4;
1010   // The CRC32 will only be aligned if we align the whole section.
1011   Align = 4;
1012   Type = OriginalType = ELF::SHT_PROGBITS;
1013   Name = ".gnu_debuglink";
1014   // For sections not found in segments, OriginalOffset is only used to
1015   // establish the order that sections should go in. By using the maximum
1016   // possible offset we cause this section to wind up at the end.
1017   OriginalOffset = std::numeric_limits<uint64_t>::max();
1018 }
1019 
1020 GnuDebugLinkSection::GnuDebugLinkSection(StringRef File,
1021                                          uint32_t PrecomputedCRC)
1022     : FileName(File), CRC32(PrecomputedCRC) {
1023   init(File);
1024 }
1025 
1026 template <class ELFT>
1027 void ELFSectionWriter<ELFT>::visit(const GnuDebugLinkSection &Sec) {
1028   unsigned char *Buf = Out.getBufferStart() + Sec.Offset;
1029   Elf_Word *CRC =
1030       reinterpret_cast<Elf_Word *>(Buf + Sec.Size - sizeof(Elf_Word));
1031   *CRC = Sec.CRC32;
1032   llvm::copy(Sec.FileName, Buf);
1033 }
1034 
1035 void GnuDebugLinkSection::accept(SectionVisitor &Visitor) const {
1036   Visitor.visit(*this);
1037 }
1038 
1039 void GnuDebugLinkSection::accept(MutableSectionVisitor &Visitor) {
1040   Visitor.visit(*this);
1041 }
1042 
1043 template <class ELFT>
1044 void ELFSectionWriter<ELFT>::visit(const GroupSection &Sec) {
1045   ELF::Elf32_Word *Buf =
1046       reinterpret_cast<ELF::Elf32_Word *>(Out.getBufferStart() + Sec.Offset);
1047   *Buf++ = Sec.FlagWord;
1048   for (SectionBase *S : Sec.GroupMembers)
1049     support::endian::write32<ELFT::TargetEndianness>(Buf++, S->Index);
1050 }
1051 
1052 void GroupSection::accept(SectionVisitor &Visitor) const {
1053   Visitor.visit(*this);
1054 }
1055 
1056 void GroupSection::accept(MutableSectionVisitor &Visitor) {
1057   Visitor.visit(*this);
1058 }
1059 
1060 // Returns true IFF a section is wholly inside the range of a segment
1061 static bool sectionWithinSegment(const SectionBase &Sec, const Segment &Seg) {
1062   // If a section is empty it should be treated like it has a size of 1. This is
1063   // to clarify the case when an empty section lies on a boundary between two
1064   // segments and ensures that the section "belongs" to the second segment and
1065   // not the first.
1066   uint64_t SecSize = Sec.Size ? Sec.Size : 1;
1067 
1068   if (Sec.Type == SHT_NOBITS) {
1069     if (!(Sec.Flags & SHF_ALLOC))
1070       return false;
1071 
1072     bool SectionIsTLS = Sec.Flags & SHF_TLS;
1073     bool SegmentIsTLS = Seg.Type == PT_TLS;
1074     if (SectionIsTLS != SegmentIsTLS)
1075       return false;
1076 
1077     return Seg.VAddr <= Sec.Addr &&
1078            Seg.VAddr + Seg.MemSize >= Sec.Addr + SecSize;
1079   }
1080 
1081   return Seg.Offset <= Sec.OriginalOffset &&
1082          Seg.Offset + Seg.FileSize >= Sec.OriginalOffset + SecSize;
1083 }
1084 
1085 // Returns true IFF a segment's original offset is inside of another segment's
1086 // range.
1087 static bool segmentOverlapsSegment(const Segment &Child,
1088                                    const Segment &Parent) {
1089 
1090   return Parent.OriginalOffset <= Child.OriginalOffset &&
1091          Parent.OriginalOffset + Parent.FileSize > Child.OriginalOffset;
1092 }
1093 
1094 static bool compareSegmentsByOffset(const Segment *A, const Segment *B) {
1095   // Any segment without a parent segment should come before a segment
1096   // that has a parent segment.
1097   if (A->OriginalOffset < B->OriginalOffset)
1098     return true;
1099   if (A->OriginalOffset > B->OriginalOffset)
1100     return false;
1101   return A->Index < B->Index;
1102 }
1103 
1104 static bool compareSegmentsByPAddr(const Segment *A, const Segment *B) {
1105   if (A->PAddr < B->PAddr)
1106     return true;
1107   if (A->PAddr > B->PAddr)
1108     return false;
1109   return A->Index < B->Index;
1110 }
1111 
1112 void BasicELFBuilder::initFileHeader() {
1113   Obj->Flags = 0x0;
1114   Obj->Type = ET_REL;
1115   Obj->OSABI = ELFOSABI_NONE;
1116   Obj->ABIVersion = 0;
1117   Obj->Entry = 0x0;
1118   Obj->Machine = EM_NONE;
1119   Obj->Version = 1;
1120 }
1121 
1122 void BasicELFBuilder::initHeaderSegment() { Obj->ElfHdrSegment.Index = 0; }
1123 
1124 StringTableSection *BasicELFBuilder::addStrTab() {
1125   auto &StrTab = Obj->addSection<StringTableSection>();
1126   StrTab.Name = ".strtab";
1127 
1128   Obj->SectionNames = &StrTab;
1129   return &StrTab;
1130 }
1131 
1132 SymbolTableSection *BasicELFBuilder::addSymTab(StringTableSection *StrTab) {
1133   auto &SymTab = Obj->addSection<SymbolTableSection>();
1134 
1135   SymTab.Name = ".symtab";
1136   SymTab.Link = StrTab->Index;
1137 
1138   // The symbol table always needs a null symbol
1139   SymTab.addSymbol("", 0, 0, nullptr, 0, 0, 0, 0);
1140 
1141   Obj->SymbolTable = &SymTab;
1142   return &SymTab;
1143 }
1144 
1145 void BasicELFBuilder::initSections() {
1146   for (SectionBase &Sec : Obj->sections())
1147     Sec.initialize(Obj->sections());
1148 }
1149 
1150 void BinaryELFBuilder::addData(SymbolTableSection *SymTab) {
1151   auto Data = ArrayRef<uint8_t>(
1152       reinterpret_cast<const uint8_t *>(MemBuf->getBufferStart()),
1153       MemBuf->getBufferSize());
1154   auto &DataSection = Obj->addSection<Section>(Data);
1155   DataSection.Name = ".data";
1156   DataSection.Type = ELF::SHT_PROGBITS;
1157   DataSection.Size = Data.size();
1158   DataSection.Flags = ELF::SHF_ALLOC | ELF::SHF_WRITE;
1159 
1160   std::string SanitizedFilename = MemBuf->getBufferIdentifier().str();
1161   std::replace_if(std::begin(SanitizedFilename), std::end(SanitizedFilename),
1162                   [](char C) { return !isalnum(C); }, '_');
1163   Twine Prefix = Twine("_binary_") + SanitizedFilename;
1164 
1165   SymTab->addSymbol(Prefix + "_start", STB_GLOBAL, STT_NOTYPE, &DataSection,
1166                     /*Value=*/0, NewSymbolVisibility, 0, 0);
1167   SymTab->addSymbol(Prefix + "_end", STB_GLOBAL, STT_NOTYPE, &DataSection,
1168                     /*Value=*/DataSection.Size, NewSymbolVisibility, 0, 0);
1169   SymTab->addSymbol(Prefix + "_size", STB_GLOBAL, STT_NOTYPE, nullptr,
1170                     /*Value=*/DataSection.Size, NewSymbolVisibility, SHN_ABS,
1171                     0);
1172 }
1173 
1174 std::unique_ptr<Object> BinaryELFBuilder::build() {
1175   initFileHeader();
1176   initHeaderSegment();
1177 
1178   SymbolTableSection *SymTab = addSymTab(addStrTab());
1179   initSections();
1180   addData(SymTab);
1181 
1182   return std::move(Obj);
1183 }
1184 
1185 // Adds sections from IHEX data file. Data should have been
1186 // fully validated by this time.
1187 void IHexELFBuilder::addDataSections() {
1188   OwnedDataSection *Section = nullptr;
1189   uint64_t SegmentAddr = 0, BaseAddr = 0;
1190   uint32_t SecNo = 1;
1191 
1192   for (const IHexRecord &R : Records) {
1193     uint64_t RecAddr;
1194     switch (R.Type) {
1195     case IHexRecord::Data:
1196       // Ignore empty data records
1197       if (R.HexData.empty())
1198         continue;
1199       RecAddr = R.Addr + SegmentAddr + BaseAddr;
1200       if (!Section || Section->Addr + Section->Size != RecAddr)
1201         // OriginalOffset field is only used to sort section properly, so
1202         // instead of keeping track of real offset in IHEX file, we use
1203         // section number.
1204         Section = &Obj->addSection<OwnedDataSection>(
1205             ".sec" + std::to_string(SecNo++), RecAddr,
1206             ELF::SHF_ALLOC | ELF::SHF_WRITE, SecNo);
1207       Section->appendHexData(R.HexData);
1208       break;
1209     case IHexRecord::EndOfFile:
1210       break;
1211     case IHexRecord::SegmentAddr:
1212       // 20-bit segment address.
1213       SegmentAddr = checkedGetHex<uint16_t>(R.HexData) << 4;
1214       break;
1215     case IHexRecord::StartAddr80x86:
1216     case IHexRecord::StartAddr:
1217       Obj->Entry = checkedGetHex<uint32_t>(R.HexData);
1218       assert(Obj->Entry <= 0xFFFFFU);
1219       break;
1220     case IHexRecord::ExtendedAddr:
1221       // 16-31 bits of linear base address
1222       BaseAddr = checkedGetHex<uint16_t>(R.HexData) << 16;
1223       break;
1224     default:
1225       llvm_unreachable("unknown record type");
1226     }
1227   }
1228 }
1229 
1230 std::unique_ptr<Object> IHexELFBuilder::build() {
1231   initFileHeader();
1232   initHeaderSegment();
1233   StringTableSection *StrTab = addStrTab();
1234   addSymTab(StrTab);
1235   initSections();
1236   addDataSections();
1237 
1238   return std::move(Obj);
1239 }
1240 
1241 template <class ELFT> void ELFBuilder<ELFT>::setParentSegment(Segment &Child) {
1242   for (Segment &Parent : Obj.segments()) {
1243     // Every segment will overlap with itself but we don't want a segment to
1244     // be it's own parent so we avoid that situation.
1245     if (&Child != &Parent && segmentOverlapsSegment(Child, Parent)) {
1246       // We want a canonical "most parental" segment but this requires
1247       // inspecting the ParentSegment.
1248       if (compareSegmentsByOffset(&Parent, &Child))
1249         if (Child.ParentSegment == nullptr ||
1250             compareSegmentsByOffset(&Parent, Child.ParentSegment)) {
1251           Child.ParentSegment = &Parent;
1252         }
1253     }
1254   }
1255 }
1256 
1257 template <class ELFT> void ELFBuilder<ELFT>::findEhdrOffset() {
1258   if (!ExtractPartition)
1259     return;
1260 
1261   for (const SectionBase &Sec : Obj.sections()) {
1262     if (Sec.Type == SHT_LLVM_PART_EHDR && Sec.Name == *ExtractPartition) {
1263       EhdrOffset = Sec.Offset;
1264       return;
1265     }
1266   }
1267   error("could not find partition named '" + *ExtractPartition + "'");
1268 }
1269 
1270 template <class ELFT>
1271 void ELFBuilder<ELFT>::readProgramHeaders(const ELFFile<ELFT> &HeadersFile) {
1272   uint32_t Index = 0;
1273   for (const auto &Phdr : unwrapOrError(HeadersFile.program_headers())) {
1274     if (Phdr.p_offset + Phdr.p_filesz > HeadersFile.getBufSize())
1275       error("program header with offset 0x" + Twine::utohexstr(Phdr.p_offset) +
1276             " and file size 0x" + Twine::utohexstr(Phdr.p_filesz) +
1277             " goes past the end of the file");
1278 
1279     ArrayRef<uint8_t> Data{HeadersFile.base() + Phdr.p_offset,
1280                            (size_t)Phdr.p_filesz};
1281     Segment &Seg = Obj.addSegment(Data);
1282     Seg.Type = Phdr.p_type;
1283     Seg.Flags = Phdr.p_flags;
1284     Seg.OriginalOffset = Phdr.p_offset + EhdrOffset;
1285     Seg.Offset = Phdr.p_offset + EhdrOffset;
1286     Seg.VAddr = Phdr.p_vaddr;
1287     Seg.PAddr = Phdr.p_paddr;
1288     Seg.FileSize = Phdr.p_filesz;
1289     Seg.MemSize = Phdr.p_memsz;
1290     Seg.Align = Phdr.p_align;
1291     Seg.Index = Index++;
1292     for (SectionBase &Sec : Obj.sections())
1293       if (sectionWithinSegment(Sec, Seg)) {
1294         Seg.addSection(&Sec);
1295         if (!Sec.ParentSegment || Sec.ParentSegment->Offset > Seg.Offset)
1296           Sec.ParentSegment = &Seg;
1297       }
1298   }
1299 
1300   auto &ElfHdr = Obj.ElfHdrSegment;
1301   ElfHdr.Index = Index++;
1302   ElfHdr.OriginalOffset = ElfHdr.Offset = EhdrOffset;
1303 
1304   const auto &Ehdr = *HeadersFile.getHeader();
1305   auto &PrHdr = Obj.ProgramHdrSegment;
1306   PrHdr.Type = PT_PHDR;
1307   PrHdr.Flags = 0;
1308   // The spec requires us to have p_vaddr % p_align == p_offset % p_align.
1309   // Whereas this works automatically for ElfHdr, here OriginalOffset is
1310   // always non-zero and to ensure the equation we assign the same value to
1311   // VAddr as well.
1312   PrHdr.OriginalOffset = PrHdr.Offset = PrHdr.VAddr = EhdrOffset + Ehdr.e_phoff;
1313   PrHdr.PAddr = 0;
1314   PrHdr.FileSize = PrHdr.MemSize = Ehdr.e_phentsize * Ehdr.e_phnum;
1315   // The spec requires us to naturally align all the fields.
1316   PrHdr.Align = sizeof(Elf_Addr);
1317   PrHdr.Index = Index++;
1318 
1319   // Now we do an O(n^2) loop through the segments in order to match up
1320   // segments.
1321   for (Segment &Child : Obj.segments())
1322     setParentSegment(Child);
1323   setParentSegment(ElfHdr);
1324   setParentSegment(PrHdr);
1325 }
1326 
1327 template <class ELFT>
1328 void ELFBuilder<ELFT>::initGroupSection(GroupSection *GroupSec) {
1329   if (GroupSec->Align % sizeof(ELF::Elf32_Word) != 0)
1330     error("invalid alignment " + Twine(GroupSec->Align) + " of group section '" +
1331           GroupSec->Name + "'");
1332   SectionTableRef SecTable = Obj.sections();
1333   auto SymTab = SecTable.template getSectionOfType<SymbolTableSection>(
1334       GroupSec->Link,
1335       "link field value '" + Twine(GroupSec->Link) + "' in section '" +
1336           GroupSec->Name + "' is invalid",
1337       "link field value '" + Twine(GroupSec->Link) + "' in section '" +
1338           GroupSec->Name + "' is not a symbol table");
1339   Symbol *Sym = SymTab->getSymbolByIndex(GroupSec->Info);
1340   if (!Sym)
1341     error("info field value '" + Twine(GroupSec->Info) + "' in section '" +
1342           GroupSec->Name + "' is not a valid symbol index");
1343   GroupSec->setSymTab(SymTab);
1344   GroupSec->setSymbol(Sym);
1345   if (GroupSec->Contents.size() % sizeof(ELF::Elf32_Word) ||
1346       GroupSec->Contents.empty())
1347     error("the content of the section " + GroupSec->Name + " is malformed");
1348   const ELF::Elf32_Word *Word =
1349       reinterpret_cast<const ELF::Elf32_Word *>(GroupSec->Contents.data());
1350   const ELF::Elf32_Word *End =
1351       Word + GroupSec->Contents.size() / sizeof(ELF::Elf32_Word);
1352   GroupSec->setFlagWord(*Word++);
1353   for (; Word != End; ++Word) {
1354     uint32_t Index = support::endian::read32<ELFT::TargetEndianness>(Word);
1355     GroupSec->addMember(SecTable.getSection(
1356         Index, "group member index " + Twine(Index) + " in section '" +
1357                    GroupSec->Name + "' is invalid"));
1358   }
1359 }
1360 
1361 template <class ELFT>
1362 void ELFBuilder<ELFT>::initSymbolTable(SymbolTableSection *SymTab) {
1363   const Elf_Shdr &Shdr = *unwrapOrError(ElfFile.getSection(SymTab->Index));
1364   StringRef StrTabData = unwrapOrError(ElfFile.getStringTableForSymtab(Shdr));
1365   ArrayRef<Elf_Word> ShndxData;
1366 
1367   auto Symbols = unwrapOrError(ElfFile.symbols(&Shdr));
1368   for (const auto &Sym : Symbols) {
1369     SectionBase *DefSection = nullptr;
1370     StringRef Name = unwrapOrError(Sym.getName(StrTabData));
1371 
1372     if (Sym.st_shndx == SHN_XINDEX) {
1373       if (SymTab->getShndxTable() == nullptr)
1374         error("symbol '" + Name +
1375               "' has index SHN_XINDEX but no SHT_SYMTAB_SHNDX section exists");
1376       if (ShndxData.data() == nullptr) {
1377         const Elf_Shdr &ShndxSec =
1378             *unwrapOrError(ElfFile.getSection(SymTab->getShndxTable()->Index));
1379         ShndxData = unwrapOrError(
1380             ElfFile.template getSectionContentsAsArray<Elf_Word>(&ShndxSec));
1381         if (ShndxData.size() != Symbols.size())
1382           error("symbol section index table does not have the same number of "
1383                 "entries as the symbol table");
1384       }
1385       Elf_Word Index = ShndxData[&Sym - Symbols.begin()];
1386       DefSection = Obj.sections().getSection(
1387           Index,
1388           "symbol '" + Name + "' has invalid section index " + Twine(Index));
1389     } else if (Sym.st_shndx >= SHN_LORESERVE) {
1390       if (!isValidReservedSectionIndex(Sym.st_shndx, Obj.Machine)) {
1391         error(
1392             "symbol '" + Name +
1393             "' has unsupported value greater than or equal to SHN_LORESERVE: " +
1394             Twine(Sym.st_shndx));
1395       }
1396     } else if (Sym.st_shndx != SHN_UNDEF) {
1397       DefSection = Obj.sections().getSection(
1398           Sym.st_shndx, "symbol '" + Name +
1399                             "' is defined has invalid section index " +
1400                             Twine(Sym.st_shndx));
1401     }
1402 
1403     SymTab->addSymbol(Name, Sym.getBinding(), Sym.getType(), DefSection,
1404                       Sym.getValue(), Sym.st_other, Sym.st_shndx, Sym.st_size);
1405   }
1406 }
1407 
1408 template <class ELFT>
1409 static void getAddend(uint64_t &ToSet, const Elf_Rel_Impl<ELFT, false> &Rel) {}
1410 
1411 template <class ELFT>
1412 static void getAddend(uint64_t &ToSet, const Elf_Rel_Impl<ELFT, true> &Rela) {
1413   ToSet = Rela.r_addend;
1414 }
1415 
1416 template <class T>
1417 static void initRelocations(RelocationSection *Relocs,
1418                             SymbolTableSection *SymbolTable, T RelRange) {
1419   for (const auto &Rel : RelRange) {
1420     Relocation ToAdd;
1421     ToAdd.Offset = Rel.r_offset;
1422     getAddend(ToAdd.Addend, Rel);
1423     ToAdd.Type = Rel.getType(false);
1424 
1425     if (uint32_t Sym = Rel.getSymbol(false)) {
1426       if (!SymbolTable)
1427         error("'" + Relocs->Name +
1428               "': relocation references symbol with index " + Twine(Sym) +
1429               ", but there is no symbol table");
1430       ToAdd.RelocSymbol = SymbolTable->getSymbolByIndex(Sym);
1431     }
1432 
1433     Relocs->addRelocation(ToAdd);
1434   }
1435 }
1436 
1437 SectionBase *SectionTableRef::getSection(uint32_t Index, Twine ErrMsg) {
1438   if (Index == SHN_UNDEF || Index > Sections.size())
1439     error(ErrMsg);
1440   return Sections[Index - 1].get();
1441 }
1442 
1443 template <class T>
1444 T *SectionTableRef::getSectionOfType(uint32_t Index, Twine IndexErrMsg,
1445                                      Twine TypeErrMsg) {
1446   if (T *Sec = dyn_cast<T>(getSection(Index, IndexErrMsg)))
1447     return Sec;
1448   error(TypeErrMsg);
1449 }
1450 
1451 template <class ELFT>
1452 SectionBase &ELFBuilder<ELFT>::makeSection(const Elf_Shdr &Shdr) {
1453   ArrayRef<uint8_t> Data;
1454   switch (Shdr.sh_type) {
1455   case SHT_REL:
1456   case SHT_RELA:
1457     if (Shdr.sh_flags & SHF_ALLOC) {
1458       Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1459       return Obj.addSection<DynamicRelocationSection>(Data);
1460     }
1461     return Obj.addSection<RelocationSection>();
1462   case SHT_STRTAB:
1463     // If a string table is allocated we don't want to mess with it. That would
1464     // mean altering the memory image. There are no special link types or
1465     // anything so we can just use a Section.
1466     if (Shdr.sh_flags & SHF_ALLOC) {
1467       Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1468       return Obj.addSection<Section>(Data);
1469     }
1470     return Obj.addSection<StringTableSection>();
1471   case SHT_HASH:
1472   case SHT_GNU_HASH:
1473     // Hash tables should refer to SHT_DYNSYM which we're not going to change.
1474     // Because of this we don't need to mess with the hash tables either.
1475     Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1476     return Obj.addSection<Section>(Data);
1477   case SHT_GROUP:
1478     Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1479     return Obj.addSection<GroupSection>(Data);
1480   case SHT_DYNSYM:
1481     Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1482     return Obj.addSection<DynamicSymbolTableSection>(Data);
1483   case SHT_DYNAMIC:
1484     Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1485     return Obj.addSection<DynamicSection>(Data);
1486   case SHT_SYMTAB: {
1487     auto &SymTab = Obj.addSection<SymbolTableSection>();
1488     Obj.SymbolTable = &SymTab;
1489     return SymTab;
1490   }
1491   case SHT_SYMTAB_SHNDX: {
1492     auto &ShndxSection = Obj.addSection<SectionIndexSection>();
1493     Obj.SectionIndexTable = &ShndxSection;
1494     return ShndxSection;
1495   }
1496   case SHT_NOBITS:
1497     return Obj.addSection<Section>(Data);
1498   default: {
1499     Data = unwrapOrError(ElfFile.getSectionContents(&Shdr));
1500 
1501     StringRef Name = unwrapOrError(ElfFile.getSectionName(&Shdr));
1502     if (Name.startswith(".zdebug") || (Shdr.sh_flags & ELF::SHF_COMPRESSED)) {
1503       uint64_t DecompressedSize, DecompressedAlign;
1504       std::tie(DecompressedSize, DecompressedAlign) =
1505           getDecompressedSizeAndAlignment<ELFT>(Data);
1506       return Obj.addSection<CompressedSection>(Data, DecompressedSize,
1507                                                DecompressedAlign);
1508     }
1509 
1510     return Obj.addSection<Section>(Data);
1511   }
1512   }
1513 }
1514 
1515 template <class ELFT> void ELFBuilder<ELFT>::readSectionHeaders() {
1516   uint32_t Index = 0;
1517   for (const auto &Shdr : unwrapOrError(ElfFile.sections())) {
1518     if (Index == 0) {
1519       ++Index;
1520       continue;
1521     }
1522     auto &Sec = makeSection(Shdr);
1523     Sec.Name = unwrapOrError(ElfFile.getSectionName(&Shdr));
1524     Sec.Type = Sec.OriginalType = Shdr.sh_type;
1525     Sec.Flags = Sec.OriginalFlags = Shdr.sh_flags;
1526     Sec.Addr = Shdr.sh_addr;
1527     Sec.Offset = Shdr.sh_offset;
1528     Sec.OriginalOffset = Shdr.sh_offset;
1529     Sec.Size = Shdr.sh_size;
1530     Sec.Link = Shdr.sh_link;
1531     Sec.Info = Shdr.sh_info;
1532     Sec.Align = Shdr.sh_addralign;
1533     Sec.EntrySize = Shdr.sh_entsize;
1534     Sec.Index = Index++;
1535     Sec.OriginalData =
1536         ArrayRef<uint8_t>(ElfFile.base() + Shdr.sh_offset,
1537                           (Shdr.sh_type == SHT_NOBITS) ? 0 : Shdr.sh_size);
1538   }
1539 }
1540 
1541 template <class ELFT> void ELFBuilder<ELFT>::readSections(bool EnsureSymtab) {
1542   uint32_t ShstrIndex = ElfFile.getHeader()->e_shstrndx;
1543   if (ShstrIndex == SHN_XINDEX)
1544     ShstrIndex = unwrapOrError(ElfFile.getSection(0))->sh_link;
1545 
1546   if (ShstrIndex == SHN_UNDEF)
1547     Obj.HadShdrs = false;
1548   else
1549     Obj.SectionNames =
1550         Obj.sections().template getSectionOfType<StringTableSection>(
1551             ShstrIndex,
1552             "e_shstrndx field value " + Twine(ShstrIndex) + " in elf header " +
1553                 " is invalid",
1554             "e_shstrndx field value " + Twine(ShstrIndex) + " in elf header " +
1555                 " does not reference a string table");
1556 
1557   // If a section index table exists we'll need to initialize it before we
1558   // initialize the symbol table because the symbol table might need to
1559   // reference it.
1560   if (Obj.SectionIndexTable)
1561     Obj.SectionIndexTable->initialize(Obj.sections());
1562 
1563   // Now that all of the sections have been added we can fill out some extra
1564   // details about symbol tables. We need the symbol table filled out before
1565   // any relocations.
1566   if (Obj.SymbolTable) {
1567     Obj.SymbolTable->initialize(Obj.sections());
1568     initSymbolTable(Obj.SymbolTable);
1569   } else if (EnsureSymtab) {
1570     // Reuse an existing SHT_STRTAB section if it exists.
1571     StringTableSection *StrTab = nullptr;
1572     for (auto &Sec : Obj.sections()) {
1573       if (Sec.Type == ELF::SHT_STRTAB && !(Sec.Flags & SHF_ALLOC)) {
1574         StrTab = static_cast<StringTableSection *>(&Sec);
1575 
1576         // Prefer a string table that is not the section header string table, if
1577         // such a table exists.
1578         if (Obj.SectionNames != &Sec)
1579           break;
1580       }
1581     }
1582     if (!StrTab)
1583       StrTab = &Obj.addSection<StringTableSection>();
1584 
1585     SymbolTableSection &SymTab = Obj.addSection<SymbolTableSection>();
1586     SymTab.Name = ".symtab";
1587     SymTab.Link = StrTab->Index;
1588     SymTab.initialize(Obj.sections());
1589     SymTab.addSymbol("", 0, 0, nullptr, 0, 0, 0, 0);
1590     Obj.SymbolTable = &SymTab;
1591   }
1592 
1593   // Now that all sections and symbols have been added we can add
1594   // relocations that reference symbols and set the link and info fields for
1595   // relocation sections.
1596   for (auto &Sec : Obj.sections()) {
1597     if (&Sec == Obj.SymbolTable)
1598       continue;
1599     Sec.initialize(Obj.sections());
1600     if (auto RelSec = dyn_cast<RelocationSection>(&Sec)) {
1601       auto Shdr = unwrapOrError(ElfFile.sections()).begin() + RelSec->Index;
1602       if (RelSec->Type == SHT_REL)
1603         initRelocations(RelSec, Obj.SymbolTable,
1604                         unwrapOrError(ElfFile.rels(Shdr)));
1605       else
1606         initRelocations(RelSec, Obj.SymbolTable,
1607                         unwrapOrError(ElfFile.relas(Shdr)));
1608     } else if (auto GroupSec = dyn_cast<GroupSection>(&Sec)) {
1609       initGroupSection(GroupSec);
1610     }
1611   }
1612 }
1613 
1614 template <class ELFT> void ELFBuilder<ELFT>::build(bool EnsureSymtab) {
1615   readSectionHeaders();
1616   findEhdrOffset();
1617 
1618   // The ELFFile whose ELF headers and program headers are copied into the
1619   // output file. Normally the same as ElfFile, but if we're extracting a
1620   // loadable partition it will point to the partition's headers.
1621   ELFFile<ELFT> HeadersFile = unwrapOrError(ELFFile<ELFT>::create(toStringRef(
1622       {ElfFile.base() + EhdrOffset, ElfFile.getBufSize() - EhdrOffset})));
1623 
1624   auto &Ehdr = *HeadersFile.getHeader();
1625   Obj.OSABI = Ehdr.e_ident[EI_OSABI];
1626   Obj.ABIVersion = Ehdr.e_ident[EI_ABIVERSION];
1627   Obj.Type = Ehdr.e_type;
1628   Obj.Machine = Ehdr.e_machine;
1629   Obj.Version = Ehdr.e_version;
1630   Obj.Entry = Ehdr.e_entry;
1631   Obj.Flags = Ehdr.e_flags;
1632 
1633   readSections(EnsureSymtab);
1634   readProgramHeaders(HeadersFile);
1635 }
1636 
1637 Writer::~Writer() {}
1638 
1639 Reader::~Reader() {}
1640 
1641 std::unique_ptr<Object> BinaryReader::create(bool /*EnsureSymtab*/) const {
1642   return BinaryELFBuilder(MemBuf, NewSymbolVisibility).build();
1643 }
1644 
1645 Expected<std::vector<IHexRecord>> IHexReader::parse() const {
1646   SmallVector<StringRef, 16> Lines;
1647   std::vector<IHexRecord> Records;
1648   bool HasSections = false;
1649 
1650   MemBuf->getBuffer().split(Lines, '\n');
1651   Records.reserve(Lines.size());
1652   for (size_t LineNo = 1; LineNo <= Lines.size(); ++LineNo) {
1653     StringRef Line = Lines[LineNo - 1].trim();
1654     if (Line.empty())
1655       continue;
1656 
1657     Expected<IHexRecord> R = IHexRecord::parse(Line);
1658     if (!R)
1659       return parseError(LineNo, R.takeError());
1660     if (R->Type == IHexRecord::EndOfFile)
1661       break;
1662     HasSections |= (R->Type == IHexRecord::Data);
1663     Records.push_back(*R);
1664   }
1665   if (!HasSections)
1666     return parseError(-1U, "no sections");
1667 
1668   return std::move(Records);
1669 }
1670 
1671 std::unique_ptr<Object> IHexReader::create(bool /*EnsureSymtab*/) const {
1672   std::vector<IHexRecord> Records = unwrapOrError(parse());
1673   return IHexELFBuilder(Records).build();
1674 }
1675 
1676 std::unique_ptr<Object> ELFReader::create(bool EnsureSymtab) const {
1677   auto Obj = std::make_unique<Object>();
1678   if (auto *O = dyn_cast<ELFObjectFile<ELF32LE>>(Bin)) {
1679     ELFBuilder<ELF32LE> Builder(*O, *Obj, ExtractPartition);
1680     Builder.build(EnsureSymtab);
1681     return Obj;
1682   } else if (auto *O = dyn_cast<ELFObjectFile<ELF64LE>>(Bin)) {
1683     ELFBuilder<ELF64LE> Builder(*O, *Obj, ExtractPartition);
1684     Builder.build(EnsureSymtab);
1685     return Obj;
1686   } else if (auto *O = dyn_cast<ELFObjectFile<ELF32BE>>(Bin)) {
1687     ELFBuilder<ELF32BE> Builder(*O, *Obj, ExtractPartition);
1688     Builder.build(EnsureSymtab);
1689     return Obj;
1690   } else if (auto *O = dyn_cast<ELFObjectFile<ELF64BE>>(Bin)) {
1691     ELFBuilder<ELF64BE> Builder(*O, *Obj, ExtractPartition);
1692     Builder.build(EnsureSymtab);
1693     return Obj;
1694   }
1695   error("invalid file type");
1696 }
1697 
1698 template <class ELFT> void ELFWriter<ELFT>::writeEhdr() {
1699   Elf_Ehdr &Ehdr = *reinterpret_cast<Elf_Ehdr *>(Buf.getBufferStart());
1700   std::fill(Ehdr.e_ident, Ehdr.e_ident + 16, 0);
1701   Ehdr.e_ident[EI_MAG0] = 0x7f;
1702   Ehdr.e_ident[EI_MAG1] = 'E';
1703   Ehdr.e_ident[EI_MAG2] = 'L';
1704   Ehdr.e_ident[EI_MAG3] = 'F';
1705   Ehdr.e_ident[EI_CLASS] = ELFT::Is64Bits ? ELFCLASS64 : ELFCLASS32;
1706   Ehdr.e_ident[EI_DATA] =
1707       ELFT::TargetEndianness == support::big ? ELFDATA2MSB : ELFDATA2LSB;
1708   Ehdr.e_ident[EI_VERSION] = EV_CURRENT;
1709   Ehdr.e_ident[EI_OSABI] = Obj.OSABI;
1710   Ehdr.e_ident[EI_ABIVERSION] = Obj.ABIVersion;
1711 
1712   Ehdr.e_type = Obj.Type;
1713   Ehdr.e_machine = Obj.Machine;
1714   Ehdr.e_version = Obj.Version;
1715   Ehdr.e_entry = Obj.Entry;
1716   // We have to use the fully-qualified name llvm::size
1717   // since some compilers complain on ambiguous resolution.
1718   Ehdr.e_phnum = llvm::size(Obj.segments());
1719   Ehdr.e_phoff = (Ehdr.e_phnum != 0) ? Obj.ProgramHdrSegment.Offset : 0;
1720   Ehdr.e_phentsize = (Ehdr.e_phnum != 0) ? sizeof(Elf_Phdr) : 0;
1721   Ehdr.e_flags = Obj.Flags;
1722   Ehdr.e_ehsize = sizeof(Elf_Ehdr);
1723   if (WriteSectionHeaders && Obj.sections().size() != 0) {
1724     Ehdr.e_shentsize = sizeof(Elf_Shdr);
1725     Ehdr.e_shoff = Obj.SHOff;
1726     // """
1727     // If the number of sections is greater than or equal to
1728     // SHN_LORESERVE (0xff00), this member has the value zero and the actual
1729     // number of section header table entries is contained in the sh_size field
1730     // of the section header at index 0.
1731     // """
1732     auto Shnum = Obj.sections().size() + 1;
1733     if (Shnum >= SHN_LORESERVE)
1734       Ehdr.e_shnum = 0;
1735     else
1736       Ehdr.e_shnum = Shnum;
1737     // """
1738     // If the section name string table section index is greater than or equal
1739     // to SHN_LORESERVE (0xff00), this member has the value SHN_XINDEX (0xffff)
1740     // and the actual index of the section name string table section is
1741     // contained in the sh_link field of the section header at index 0.
1742     // """
1743     if (Obj.SectionNames->Index >= SHN_LORESERVE)
1744       Ehdr.e_shstrndx = SHN_XINDEX;
1745     else
1746       Ehdr.e_shstrndx = Obj.SectionNames->Index;
1747   } else {
1748     Ehdr.e_shentsize = 0;
1749     Ehdr.e_shoff = 0;
1750     Ehdr.e_shnum = 0;
1751     Ehdr.e_shstrndx = 0;
1752   }
1753 }
1754 
1755 template <class ELFT> void ELFWriter<ELFT>::writePhdrs() {
1756   for (auto &Seg : Obj.segments())
1757     writePhdr(Seg);
1758 }
1759 
1760 template <class ELFT> void ELFWriter<ELFT>::writeShdrs() {
1761   // This reference serves to write the dummy section header at the begining
1762   // of the file. It is not used for anything else
1763   Elf_Shdr &Shdr =
1764       *reinterpret_cast<Elf_Shdr *>(Buf.getBufferStart() + Obj.SHOff);
1765   Shdr.sh_name = 0;
1766   Shdr.sh_type = SHT_NULL;
1767   Shdr.sh_flags = 0;
1768   Shdr.sh_addr = 0;
1769   Shdr.sh_offset = 0;
1770   // See writeEhdr for why we do this.
1771   uint64_t Shnum = Obj.sections().size() + 1;
1772   if (Shnum >= SHN_LORESERVE)
1773     Shdr.sh_size = Shnum;
1774   else
1775     Shdr.sh_size = 0;
1776   // See writeEhdr for why we do this.
1777   if (Obj.SectionNames != nullptr && Obj.SectionNames->Index >= SHN_LORESERVE)
1778     Shdr.sh_link = Obj.SectionNames->Index;
1779   else
1780     Shdr.sh_link = 0;
1781   Shdr.sh_info = 0;
1782   Shdr.sh_addralign = 0;
1783   Shdr.sh_entsize = 0;
1784 
1785   for (SectionBase &Sec : Obj.sections())
1786     writeShdr(Sec);
1787 }
1788 
1789 template <class ELFT> void ELFWriter<ELFT>::writeSectionData() {
1790   for (SectionBase &Sec : Obj.sections())
1791     // Segments are responsible for writing their contents, so only write the
1792     // section data if the section is not in a segment. Note that this renders
1793     // sections in segments effectively immutable.
1794     if (Sec.ParentSegment == nullptr)
1795       Sec.accept(*SecWriter);
1796 }
1797 
1798 template <class ELFT> void ELFWriter<ELFT>::writeSegmentData() {
1799   for (Segment &Seg : Obj.segments()) {
1800     size_t Size = std::min<size_t>(Seg.FileSize, Seg.getContents().size());
1801     std::memcpy(Buf.getBufferStart() + Seg.Offset, Seg.getContents().data(),
1802                 Size);
1803   }
1804 
1805   // Iterate over removed sections and overwrite their old data with zeroes.
1806   for (auto &Sec : Obj.removedSections()) {
1807     Segment *Parent = Sec.ParentSegment;
1808     if (Parent == nullptr || Sec.Type == SHT_NOBITS || Sec.Size == 0)
1809       continue;
1810     uint64_t Offset =
1811         Sec.OriginalOffset - Parent->OriginalOffset + Parent->Offset;
1812     std::memset(Buf.getBufferStart() + Offset, 0, Sec.Size);
1813   }
1814 }
1815 
1816 template <class ELFT>
1817 ELFWriter<ELFT>::ELFWriter(Object &Obj, Buffer &Buf, bool WSH,
1818                            bool OnlyKeepDebug)
1819     : Writer(Obj, Buf), WriteSectionHeaders(WSH && Obj.HadShdrs),
1820       OnlyKeepDebug(OnlyKeepDebug) {}
1821 
1822 Error Object::removeSections(bool AllowBrokenLinks,
1823     std::function<bool(const SectionBase &)> ToRemove) {
1824 
1825   auto Iter = std::stable_partition(
1826       std::begin(Sections), std::end(Sections), [=](const SecPtr &Sec) {
1827         if (ToRemove(*Sec))
1828           return false;
1829         if (auto RelSec = dyn_cast<RelocationSectionBase>(Sec.get())) {
1830           if (auto ToRelSec = RelSec->getSection())
1831             return !ToRemove(*ToRelSec);
1832         }
1833         return true;
1834       });
1835   if (SymbolTable != nullptr && ToRemove(*SymbolTable))
1836     SymbolTable = nullptr;
1837   if (SectionNames != nullptr && ToRemove(*SectionNames))
1838     SectionNames = nullptr;
1839   if (SectionIndexTable != nullptr && ToRemove(*SectionIndexTable))
1840     SectionIndexTable = nullptr;
1841   // Now make sure there are no remaining references to the sections that will
1842   // be removed. Sometimes it is impossible to remove a reference so we emit
1843   // an error here instead.
1844   std::unordered_set<const SectionBase *> RemoveSections;
1845   RemoveSections.reserve(std::distance(Iter, std::end(Sections)));
1846   for (auto &RemoveSec : make_range(Iter, std::end(Sections))) {
1847     for (auto &Segment : Segments)
1848       Segment->removeSection(RemoveSec.get());
1849     RemoveSections.insert(RemoveSec.get());
1850   }
1851 
1852   // For each section that remains alive, we want to remove the dead references.
1853   // This either might update the content of the section (e.g. remove symbols
1854   // from symbol table that belongs to removed section) or trigger an error if
1855   // a live section critically depends on a section being removed somehow
1856   // (e.g. the removed section is referenced by a relocation).
1857   for (auto &KeepSec : make_range(std::begin(Sections), Iter)) {
1858     if (Error E = KeepSec->removeSectionReferences(AllowBrokenLinks,
1859             [&RemoveSections](const SectionBase *Sec) {
1860               return RemoveSections.find(Sec) != RemoveSections.end();
1861             }))
1862       return E;
1863   }
1864 
1865   // Transfer removed sections into the Object RemovedSections container for use
1866   // later.
1867   std::move(Iter, Sections.end(), std::back_inserter(RemovedSections));
1868   // Now finally get rid of them all together.
1869   Sections.erase(Iter, std::end(Sections));
1870   return Error::success();
1871 }
1872 
1873 Error Object::removeSymbols(function_ref<bool(const Symbol &)> ToRemove) {
1874   if (SymbolTable)
1875     for (const SecPtr &Sec : Sections)
1876       if (Error E = Sec->removeSymbols(ToRemove))
1877         return E;
1878   return Error::success();
1879 }
1880 
1881 void Object::sortSections() {
1882   // Use stable_sort to maintain the original ordering as closely as possible.
1883   llvm::stable_sort(Sections, [](const SecPtr &A, const SecPtr &B) {
1884     // Put SHT_GROUP sections first, since group section headers must come
1885     // before the sections they contain. This also matches what GNU objcopy
1886     // does.
1887     if (A->Type != B->Type &&
1888         (A->Type == ELF::SHT_GROUP || B->Type == ELF::SHT_GROUP))
1889       return A->Type == ELF::SHT_GROUP;
1890     // For all other sections, sort by offset order.
1891     return A->OriginalOffset < B->OriginalOffset;
1892   });
1893 }
1894 
1895 // Orders segments such that if x = y->ParentSegment then y comes before x.
1896 static void orderSegments(std::vector<Segment *> &Segments) {
1897   llvm::stable_sort(Segments, compareSegmentsByOffset);
1898 }
1899 
1900 // This function finds a consistent layout for a list of segments starting from
1901 // an Offset. It assumes that Segments have been sorted by orderSegments and
1902 // returns an Offset one past the end of the last segment.
1903 static uint64_t layoutSegments(std::vector<Segment *> &Segments,
1904                                uint64_t Offset) {
1905   assert(std::is_sorted(std::begin(Segments), std::end(Segments),
1906                         compareSegmentsByOffset));
1907   // The only way a segment should move is if a section was between two
1908   // segments and that section was removed. If that section isn't in a segment
1909   // then it's acceptable, but not ideal, to simply move it to after the
1910   // segments. So we can simply layout segments one after the other accounting
1911   // for alignment.
1912   for (Segment *Seg : Segments) {
1913     // We assume that segments have been ordered by OriginalOffset and Index
1914     // such that a parent segment will always come before a child segment in
1915     // OrderedSegments. This means that the Offset of the ParentSegment should
1916     // already be set and we can set our offset relative to it.
1917     if (Seg->ParentSegment != nullptr) {
1918       Segment *Parent = Seg->ParentSegment;
1919       Seg->Offset =
1920           Parent->Offset + Seg->OriginalOffset - Parent->OriginalOffset;
1921     } else {
1922       Seg->Offset =
1923           alignTo(Offset, std::max<uint64_t>(Seg->Align, 1), Seg->VAddr);
1924     }
1925     Offset = std::max(Offset, Seg->Offset + Seg->FileSize);
1926   }
1927   return Offset;
1928 }
1929 
1930 // This function finds a consistent layout for a list of sections. It assumes
1931 // that the ->ParentSegment of each section has already been laid out. The
1932 // supplied starting Offset is used for the starting offset of any section that
1933 // does not have a ParentSegment. It returns either the offset given if all
1934 // sections had a ParentSegment or an offset one past the last section if there
1935 // was a section that didn't have a ParentSegment.
1936 template <class Range>
1937 static uint64_t layoutSections(Range Sections, uint64_t Offset) {
1938   // Now the offset of every segment has been set we can assign the offsets
1939   // of each section. For sections that are covered by a segment we should use
1940   // the segment's original offset and the section's original offset to compute
1941   // the offset from the start of the segment. Using the offset from the start
1942   // of the segment we can assign a new offset to the section. For sections not
1943   // covered by segments we can just bump Offset to the next valid location.
1944   uint32_t Index = 1;
1945   for (auto &Sec : Sections) {
1946     Sec.Index = Index++;
1947     if (Sec.ParentSegment != nullptr) {
1948       auto Segment = *Sec.ParentSegment;
1949       Sec.Offset =
1950           Segment.Offset + (Sec.OriginalOffset - Segment.OriginalOffset);
1951     } else {
1952       Offset = alignTo(Offset, Sec.Align == 0 ? 1 : Sec.Align);
1953       Sec.Offset = Offset;
1954       if (Sec.Type != SHT_NOBITS)
1955         Offset += Sec.Size;
1956     }
1957   }
1958   return Offset;
1959 }
1960 
1961 // Rewrite sh_offset after some sections are changed to SHT_NOBITS and thus
1962 // occupy no space in the file.
1963 static uint64_t layoutSectionsForOnlyKeepDebug(Object &Obj, uint64_t Off) {
1964   uint32_t Index = 1;
1965   for (auto &Sec : Obj.sections()) {
1966     Sec.Index = Index++;
1967 
1968     auto *FirstSec = Sec.ParentSegment && Sec.ParentSegment->Type == PT_LOAD
1969                          ? Sec.ParentSegment->firstSection()
1970                          : nullptr;
1971 
1972     // The first section in a PT_LOAD has to have congruent offset and address
1973     // modulo the alignment, which usually equals the maximum page size.
1974     if (FirstSec && FirstSec == &Sec)
1975       Off = alignTo(Off, Sec.ParentSegment->Align, Sec.Addr);
1976 
1977     // sh_offset is not significant for SHT_NOBITS sections, but the congruence
1978     // rule must be followed if it is the first section in a PT_LOAD. Do not
1979     // advance Off.
1980     if (Sec.Type == SHT_NOBITS) {
1981       Sec.Offset = Off;
1982       continue;
1983     }
1984 
1985     if (!FirstSec) {
1986       // FirstSec being nullptr generally means that Sec does not have the
1987       // SHF_ALLOC flag.
1988       Off = Sec.Align ? alignTo(Off, Sec.Align) : Off;
1989     } else if (FirstSec != &Sec) {
1990       // The offset is relative to the first section in the PT_LOAD segment. Use
1991       // sh_offset for non-SHF_ALLOC sections.
1992       Off = Sec.OriginalOffset - FirstSec->OriginalOffset + FirstSec->Offset;
1993     }
1994     Sec.Offset = Off;
1995     Off += Sec.Size;
1996   }
1997   return Off;
1998 }
1999 
2000 // Rewrite p_offset and p_filesz of non-empty non-PT_PHDR segments after
2001 // sh_offset values have been updated.
2002 static uint64_t layoutSegmentsForOnlyKeepDebug(std::vector<Segment *> &Segments,
2003                                                uint64_t HdrEnd) {
2004   uint64_t MaxOffset = 0;
2005   for (Segment *Seg : Segments) {
2006     const SectionBase *FirstSec = Seg->firstSection();
2007     if (Seg->Type == PT_PHDR || !FirstSec)
2008       continue;
2009 
2010     uint64_t Offset = FirstSec->Offset;
2011     uint64_t FileSize = 0;
2012     for (const SectionBase *Sec : Seg->Sections) {
2013       uint64_t Size = Sec->Type == SHT_NOBITS ? 0 : Sec->Size;
2014       if (Sec->Offset + Size > Offset)
2015         FileSize = std::max(FileSize, Sec->Offset + Size - Offset);
2016     }
2017 
2018     // If the segment includes EHDR and program headers, don't make it smaller
2019     // than the headers.
2020     if (Seg->Offset < HdrEnd && HdrEnd <= Seg->Offset + Seg->FileSize) {
2021       FileSize += Offset - Seg->Offset;
2022       Offset = Seg->Offset;
2023       FileSize = std::max(FileSize, HdrEnd - Offset);
2024     }
2025 
2026     Seg->Offset = Offset;
2027     Seg->FileSize = FileSize;
2028     MaxOffset = std::max(MaxOffset, Offset + FileSize);
2029   }
2030   return MaxOffset;
2031 }
2032 
2033 template <class ELFT> void ELFWriter<ELFT>::initEhdrSegment() {
2034   Segment &ElfHdr = Obj.ElfHdrSegment;
2035   ElfHdr.Type = PT_PHDR;
2036   ElfHdr.Flags = 0;
2037   ElfHdr.VAddr = 0;
2038   ElfHdr.PAddr = 0;
2039   ElfHdr.FileSize = ElfHdr.MemSize = sizeof(Elf_Ehdr);
2040   ElfHdr.Align = 0;
2041 }
2042 
2043 template <class ELFT> void ELFWriter<ELFT>::assignOffsets() {
2044   // We need a temporary list of segments that has a special order to it
2045   // so that we know that anytime ->ParentSegment is set that segment has
2046   // already had its offset properly set.
2047   std::vector<Segment *> OrderedSegments;
2048   for (Segment &Segment : Obj.segments())
2049     OrderedSegments.push_back(&Segment);
2050   OrderedSegments.push_back(&Obj.ElfHdrSegment);
2051   OrderedSegments.push_back(&Obj.ProgramHdrSegment);
2052   orderSegments(OrderedSegments);
2053 
2054   uint64_t Offset;
2055   if (OnlyKeepDebug) {
2056     // For --only-keep-debug, the sections that did not preserve contents were
2057     // changed to SHT_NOBITS. We now rewrite sh_offset fields of sections, and
2058     // then rewrite p_offset/p_filesz of program headers.
2059     uint64_t HdrEnd =
2060         sizeof(Elf_Ehdr) + llvm::size(Obj.segments()) * sizeof(Elf_Phdr);
2061     Offset = layoutSectionsForOnlyKeepDebug(Obj, HdrEnd);
2062     Offset = std::max(Offset,
2063                       layoutSegmentsForOnlyKeepDebug(OrderedSegments, HdrEnd));
2064   } else {
2065     // Offset is used as the start offset of the first segment to be laid out.
2066     // Since the ELF Header (ElfHdrSegment) must be at the start of the file,
2067     // we start at offset 0.
2068     Offset = layoutSegments(OrderedSegments, 0);
2069     Offset = layoutSections(Obj.sections(), Offset);
2070   }
2071   // If we need to write the section header table out then we need to align the
2072   // Offset so that SHOffset is valid.
2073   if (WriteSectionHeaders)
2074     Offset = alignTo(Offset, sizeof(Elf_Addr));
2075   Obj.SHOff = Offset;
2076 }
2077 
2078 template <class ELFT> size_t ELFWriter<ELFT>::totalSize() const {
2079   // We already have the section header offset so we can calculate the total
2080   // size by just adding up the size of each section header.
2081   if (!WriteSectionHeaders)
2082     return Obj.SHOff;
2083   size_t ShdrCount = Obj.sections().size() + 1; // Includes null shdr.
2084   return Obj.SHOff + ShdrCount * sizeof(Elf_Shdr);
2085 }
2086 
2087 template <class ELFT> Error ELFWriter<ELFT>::write() {
2088   // Segment data must be written first, so that the ELF header and program
2089   // header tables can overwrite it, if covered by a segment.
2090   writeSegmentData();
2091   writeEhdr();
2092   writePhdrs();
2093   writeSectionData();
2094   if (WriteSectionHeaders)
2095     writeShdrs();
2096   return Buf.commit();
2097 }
2098 
2099 static Error removeUnneededSections(Object &Obj) {
2100   // We can remove an empty symbol table from non-relocatable objects.
2101   // Relocatable objects typically have relocation sections whose
2102   // sh_link field points to .symtab, so we can't remove .symtab
2103   // even if it is empty.
2104   if (Obj.isRelocatable() || Obj.SymbolTable == nullptr ||
2105       !Obj.SymbolTable->empty())
2106     return Error::success();
2107 
2108   // .strtab can be used for section names. In such a case we shouldn't
2109   // remove it.
2110   auto *StrTab = Obj.SymbolTable->getStrTab() == Obj.SectionNames
2111                      ? nullptr
2112                      : Obj.SymbolTable->getStrTab();
2113   return Obj.removeSections(false, [&](const SectionBase &Sec) {
2114     return &Sec == Obj.SymbolTable || &Sec == StrTab;
2115   });
2116 }
2117 
2118 template <class ELFT> Error ELFWriter<ELFT>::finalize() {
2119   // It could happen that SectionNames has been removed and yet the user wants
2120   // a section header table output. We need to throw an error if a user tries
2121   // to do that.
2122   if (Obj.SectionNames == nullptr && WriteSectionHeaders)
2123     return createStringError(llvm::errc::invalid_argument,
2124                              "cannot write section header table because "
2125                              "section header string table was removed");
2126 
2127   if (Error E = removeUnneededSections(Obj))
2128     return E;
2129   Obj.sortSections();
2130 
2131   // We need to assign indexes before we perform layout because we need to know
2132   // if we need large indexes or not. We can assign indexes first and check as
2133   // we go to see if we will actully need large indexes.
2134   bool NeedsLargeIndexes = false;
2135   if (Obj.sections().size() >= SHN_LORESERVE) {
2136     SectionTableRef Sections = Obj.sections();
2137     NeedsLargeIndexes =
2138         std::any_of(Sections.begin() + SHN_LORESERVE, Sections.end(),
2139                     [](const SectionBase &Sec) { return Sec.HasSymbol; });
2140     // TODO: handle case where only one section needs the large index table but
2141     // only needs it because the large index table hasn't been removed yet.
2142   }
2143 
2144   if (NeedsLargeIndexes) {
2145     // This means we definitely need to have a section index table but if we
2146     // already have one then we should use it instead of making a new one.
2147     if (Obj.SymbolTable != nullptr && Obj.SectionIndexTable == nullptr) {
2148       // Addition of a section to the end does not invalidate the indexes of
2149       // other sections and assigns the correct index to the new section.
2150       auto &Shndx = Obj.addSection<SectionIndexSection>();
2151       Obj.SymbolTable->setShndxTable(&Shndx);
2152       Shndx.setSymTab(Obj.SymbolTable);
2153     }
2154   } else {
2155     // Since we don't need SectionIndexTable we should remove it and all
2156     // references to it.
2157     if (Obj.SectionIndexTable != nullptr) {
2158       // We do not support sections referring to the section index table.
2159       if (Error E = Obj.removeSections(false /*AllowBrokenLinks*/,
2160                                        [this](const SectionBase &Sec) {
2161                                          return &Sec == Obj.SectionIndexTable;
2162                                        }))
2163         return E;
2164     }
2165   }
2166 
2167   // Make sure we add the names of all the sections. Importantly this must be
2168   // done after we decide to add or remove SectionIndexes.
2169   if (Obj.SectionNames != nullptr)
2170     for (const SectionBase &Sec : Obj.sections())
2171       Obj.SectionNames->addString(Sec.Name);
2172 
2173   initEhdrSegment();
2174 
2175   // Before we can prepare for layout the indexes need to be finalized.
2176   // Also, the output arch may not be the same as the input arch, so fix up
2177   // size-related fields before doing layout calculations.
2178   uint64_t Index = 0;
2179   auto SecSizer = std::make_unique<ELFSectionSizer<ELFT>>();
2180   for (SectionBase &Sec : Obj.sections()) {
2181     Sec.Index = Index++;
2182     Sec.accept(*SecSizer);
2183   }
2184 
2185   // The symbol table does not update all other sections on update. For
2186   // instance, symbol names are not added as new symbols are added. This means
2187   // that some sections, like .strtab, don't yet have their final size.
2188   if (Obj.SymbolTable != nullptr)
2189     Obj.SymbolTable->prepareForLayout();
2190 
2191   // Now that all strings are added we want to finalize string table builders,
2192   // because that affects section sizes which in turn affects section offsets.
2193   for (SectionBase &Sec : Obj.sections())
2194     if (auto StrTab = dyn_cast<StringTableSection>(&Sec))
2195       StrTab->prepareForLayout();
2196 
2197   assignOffsets();
2198 
2199   // layoutSections could have modified section indexes, so we need
2200   // to fill the index table after assignOffsets.
2201   if (Obj.SymbolTable != nullptr)
2202     Obj.SymbolTable->fillShndxTable();
2203 
2204   // Finally now that all offsets and indexes have been set we can finalize any
2205   // remaining issues.
2206   uint64_t Offset = Obj.SHOff + sizeof(Elf_Shdr);
2207   for (SectionBase &Sec : Obj.sections()) {
2208     Sec.HeaderOffset = Offset;
2209     Offset += sizeof(Elf_Shdr);
2210     if (WriteSectionHeaders)
2211       Sec.NameIndex = Obj.SectionNames->findIndex(Sec.Name);
2212     Sec.finalize();
2213   }
2214 
2215   if (Error E = Buf.allocate(totalSize()))
2216     return E;
2217   SecWriter = std::make_unique<ELFSectionWriter<ELFT>>(Buf);
2218   return Error::success();
2219 }
2220 
2221 Error BinaryWriter::write() {
2222   for (const SectionBase &Sec : Obj.allocSections())
2223     Sec.accept(*SecWriter);
2224   return Buf.commit();
2225 }
2226 
2227 Error BinaryWriter::finalize() {
2228   // We need a temporary list of segments that has a special order to it
2229   // so that we know that anytime ->ParentSegment is set that segment has
2230   // already had it's offset properly set. We only want to consider the segments
2231   // that will affect layout of allocated sections so we only add those.
2232   std::vector<Segment *> OrderedSegments;
2233   for (const SectionBase &Sec : Obj.allocSections())
2234     if (Sec.ParentSegment != nullptr)
2235       OrderedSegments.push_back(Sec.ParentSegment);
2236 
2237   // For binary output, we're going to use physical addresses instead of
2238   // virtual addresses, since a binary output is used for cases like ROM
2239   // loading and physical addresses are intended for ROM loading.
2240   // However, if no segment has a physical address, we'll fallback to using
2241   // virtual addresses for all.
2242   if (all_of(OrderedSegments,
2243              [](const Segment *Seg) { return Seg->PAddr == 0; }))
2244     for (Segment *Seg : OrderedSegments)
2245       Seg->PAddr = Seg->VAddr;
2246 
2247   llvm::stable_sort(OrderedSegments, compareSegmentsByPAddr);
2248 
2249   // Because we add a ParentSegment for each section we might have duplicate
2250   // segments in OrderedSegments. If there were duplicates then layoutSegments
2251   // would do very strange things.
2252   auto End =
2253       std::unique(std::begin(OrderedSegments), std::end(OrderedSegments));
2254   OrderedSegments.erase(End, std::end(OrderedSegments));
2255 
2256   // Compute the section LMA based on its sh_offset and the containing segment's
2257   // p_offset and p_paddr. Also compute the minimum LMA of all non-empty
2258   // sections as MinAddr. In the output, the contents between address 0 and
2259   // MinAddr will be skipped.
2260   uint64_t MinAddr = UINT64_MAX;
2261   for (SectionBase &Sec : Obj.allocSections()) {
2262     if (Sec.ParentSegment != nullptr)
2263       Sec.Addr =
2264           Sec.Offset - Sec.ParentSegment->Offset + Sec.ParentSegment->PAddr;
2265     if (Sec.Size > 0)
2266       MinAddr = std::min(MinAddr, Sec.Addr);
2267   }
2268 
2269   // Now that every section has been laid out we just need to compute the total
2270   // file size. This might not be the same as the offset returned by
2271   // layoutSections, because we want to truncate the last segment to the end of
2272   // its last non-empty section, to match GNU objcopy's behaviour.
2273   TotalSize = 0;
2274   for (SectionBase &Sec : Obj.allocSections())
2275     if (Sec.Type != SHT_NOBITS && Sec.Size > 0) {
2276       Sec.Offset = Sec.Addr - MinAddr;
2277       TotalSize = std::max(TotalSize, Sec.Offset + Sec.Size);
2278     }
2279 
2280   if (Error E = Buf.allocate(TotalSize))
2281     return E;
2282   SecWriter = std::make_unique<BinarySectionWriter>(Buf);
2283   return Error::success();
2284 }
2285 
2286 bool IHexWriter::SectionCompare::operator()(const SectionBase *Lhs,
2287                                             const SectionBase *Rhs) const {
2288   return (sectionPhysicalAddr(Lhs) & 0xFFFFFFFFU) <
2289          (sectionPhysicalAddr(Rhs) & 0xFFFFFFFFU);
2290 }
2291 
2292 uint64_t IHexWriter::writeEntryPointRecord(uint8_t *Buf) {
2293   IHexLineData HexData;
2294   uint8_t Data[4] = {};
2295   // We don't write entry point record if entry is zero.
2296   if (Obj.Entry == 0)
2297     return 0;
2298 
2299   if (Obj.Entry <= 0xFFFFFU) {
2300     Data[0] = ((Obj.Entry & 0xF0000U) >> 12) & 0xFF;
2301     support::endian::write(&Data[2], static_cast<uint16_t>(Obj.Entry),
2302                            support::big);
2303     HexData = IHexRecord::getLine(IHexRecord::StartAddr80x86, 0, Data);
2304   } else {
2305     support::endian::write(Data, static_cast<uint32_t>(Obj.Entry),
2306                            support::big);
2307     HexData = IHexRecord::getLine(IHexRecord::StartAddr, 0, Data);
2308   }
2309   memcpy(Buf, HexData.data(), HexData.size());
2310   return HexData.size();
2311 }
2312 
2313 uint64_t IHexWriter::writeEndOfFileRecord(uint8_t *Buf) {
2314   IHexLineData HexData = IHexRecord::getLine(IHexRecord::EndOfFile, 0, {});
2315   memcpy(Buf, HexData.data(), HexData.size());
2316   return HexData.size();
2317 }
2318 
2319 Error IHexWriter::write() {
2320   IHexSectionWriter Writer(Buf);
2321   // Write sections.
2322   for (const SectionBase *Sec : Sections)
2323     Sec->accept(Writer);
2324 
2325   uint64_t Offset = Writer.getBufferOffset();
2326   // Write entry point address.
2327   Offset += writeEntryPointRecord(Buf.getBufferStart() + Offset);
2328   // Write EOF.
2329   Offset += writeEndOfFileRecord(Buf.getBufferStart() + Offset);
2330   assert(Offset == TotalSize);
2331   return Buf.commit();
2332 }
2333 
2334 Error IHexWriter::checkSection(const SectionBase &Sec) {
2335   uint64_t Addr = sectionPhysicalAddr(&Sec);
2336   if (addressOverflows32bit(Addr) || addressOverflows32bit(Addr + Sec.Size - 1))
2337     return createStringError(
2338         errc::invalid_argument,
2339         "Section '%s' address range [0x%llx, 0x%llx] is not 32 bit", Sec.Name.c_str(),
2340         Addr, Addr + Sec.Size - 1);
2341   return Error::success();
2342 }
2343 
2344 Error IHexWriter::finalize() {
2345   bool UseSegments = false;
2346   auto ShouldWrite = [](const SectionBase &Sec) {
2347     return (Sec.Flags & ELF::SHF_ALLOC) && (Sec.Type != ELF::SHT_NOBITS);
2348   };
2349   auto IsInPtLoad = [](const SectionBase &Sec) {
2350     return Sec.ParentSegment && Sec.ParentSegment->Type == ELF::PT_LOAD;
2351   };
2352 
2353   // We can't write 64-bit addresses.
2354   if (addressOverflows32bit(Obj.Entry))
2355     return createStringError(errc::invalid_argument,
2356                              "Entry point address 0x%llx overflows 32 bits.",
2357                              Obj.Entry);
2358 
2359   // If any section we're to write has segment then we
2360   // switch to using physical addresses. Otherwise we
2361   // use section virtual address.
2362   for (const SectionBase &Sec : Obj.sections())
2363     if (ShouldWrite(Sec) && IsInPtLoad(Sec)) {
2364       UseSegments = true;
2365       break;
2366     }
2367 
2368   for (const SectionBase &Sec : Obj.sections())
2369     if (ShouldWrite(Sec) && (!UseSegments || IsInPtLoad(Sec))) {
2370       if (Error E = checkSection(Sec))
2371         return E;
2372       Sections.insert(&Sec);
2373     }
2374 
2375   IHexSectionWriterBase LengthCalc(Buf);
2376   for (const SectionBase *Sec : Sections)
2377     Sec->accept(LengthCalc);
2378 
2379   // We need space to write section records + StartAddress record
2380   // (if start adress is not zero) + EndOfFile record.
2381   TotalSize = LengthCalc.getBufferOffset() +
2382               (Obj.Entry ? IHexRecord::getLineLength(4) : 0) +
2383               IHexRecord::getLineLength(0);
2384   if (Error E = Buf.allocate(TotalSize))
2385     return E;
2386   return Error::success();
2387 }
2388 
2389 template class ELFBuilder<ELF64LE>;
2390 template class ELFBuilder<ELF64BE>;
2391 template class ELFBuilder<ELF32LE>;
2392 template class ELFBuilder<ELF32BE>;
2393 
2394 template class ELFWriter<ELF64LE>;
2395 template class ELFWriter<ELF64BE>;
2396 template class ELFWriter<ELF32LE>;
2397 template class ELFWriter<ELF32BE>;
2398 
2399 } // end namespace elf
2400 } // end namespace objcopy
2401 } // end namespace llvm
2402