LLVM 20.0.0git
MachOWriter.cpp
Go to the documentation of this file.
1//===- MachOWriter.cpp ------------------------------------------*- C++ -*-===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8
9#include "MachOWriter.h"
10#include "MachOLayoutBuilder.h"
11#include "MachOObject.h"
12#include "llvm/ADT/STLExtras.h"
14#include "llvm/Support/Errc.h"
16#include "llvm/Support/SHA256.h"
17#include <memory>
18
19#if defined(__APPLE__)
20#include <sys/mman.h>
21#endif
22
23using namespace llvm;
24using namespace llvm::objcopy::macho;
25using namespace llvm::support::endian;
26
27size_t MachOWriter::headerSize() const {
28 return Is64Bit ? sizeof(MachO::mach_header_64) : sizeof(MachO::mach_header);
29}
30
31size_t MachOWriter::loadCommandsSize() const { return O.Header.SizeOfCmds; }
32
33size_t MachOWriter::symTableSize() const {
34 return O.SymTable.Symbols.size() *
35 (Is64Bit ? sizeof(MachO::nlist_64) : sizeof(MachO::nlist));
36}
37
38size_t MachOWriter::totalSize() const {
39 // Going from tail to head and looking for an appropriate "anchor" to
40 // calculate the total size assuming that all the offsets are either valid
41 // ("true") or 0 (0 indicates that the corresponding part is missing).
42
44 if (O.SymTabCommandIndex) {
45 const MachO::symtab_command &SymTabCommand =
47 .MachOLoadCommand.symtab_command_data;
48 if (SymTabCommand.symoff)
49 Ends.push_back(SymTabCommand.symoff + symTableSize());
50 if (SymTabCommand.stroff)
51 Ends.push_back(SymTabCommand.stroff + SymTabCommand.strsize);
52 }
54 const MachO::dyld_info_command &DyLdInfoCommand =
56 .MachOLoadCommand.dyld_info_command_data;
57 if (DyLdInfoCommand.rebase_off) {
58 assert((DyLdInfoCommand.rebase_size == O.Rebases.Opcodes.size()) &&
59 "Incorrect rebase opcodes size");
60 Ends.push_back(DyLdInfoCommand.rebase_off + DyLdInfoCommand.rebase_size);
61 }
62 if (DyLdInfoCommand.bind_off) {
63 assert((DyLdInfoCommand.bind_size == O.Binds.Opcodes.size()) &&
64 "Incorrect bind opcodes size");
65 Ends.push_back(DyLdInfoCommand.bind_off + DyLdInfoCommand.bind_size);
66 }
67 if (DyLdInfoCommand.weak_bind_off) {
68 assert((DyLdInfoCommand.weak_bind_size == O.WeakBinds.Opcodes.size()) &&
69 "Incorrect weak bind opcodes size");
70 Ends.push_back(DyLdInfoCommand.weak_bind_off +
71 DyLdInfoCommand.weak_bind_size);
72 }
73 if (DyLdInfoCommand.lazy_bind_off) {
74 assert((DyLdInfoCommand.lazy_bind_size == O.LazyBinds.Opcodes.size()) &&
75 "Incorrect lazy bind opcodes size");
76 Ends.push_back(DyLdInfoCommand.lazy_bind_off +
77 DyLdInfoCommand.lazy_bind_size);
78 }
79 if (DyLdInfoCommand.export_off) {
80 assert((DyLdInfoCommand.export_size == O.Exports.Trie.size()) &&
81 "Incorrect trie size");
82 Ends.push_back(DyLdInfoCommand.export_off + DyLdInfoCommand.export_size);
83 }
84 }
85
87 const MachO::dysymtab_command &DySymTabCommand =
89 .MachOLoadCommand.dysymtab_command_data;
90
91 if (DySymTabCommand.indirectsymoff)
92 Ends.push_back(DySymTabCommand.indirectsymoff +
93 sizeof(uint32_t) * O.IndirectSymTable.Symbols.size());
94 }
95
96 for (std::optional<size_t> LinkEditDataCommandIndex :
101 if (LinkEditDataCommandIndex) {
102 const MachO::linkedit_data_command &LinkEditDataCommand =
103 O.LoadCommands[*LinkEditDataCommandIndex]
104 .MachOLoadCommand.linkedit_data_command_data;
105 if (LinkEditDataCommand.dataoff)
106 Ends.push_back(LinkEditDataCommand.dataoff +
107 LinkEditDataCommand.datasize);
108 }
109
110 // Otherwise, use the last section / reloction.
111 for (const LoadCommand &LC : O.LoadCommands)
112 for (const std::unique_ptr<Section> &S : LC.Sections) {
113 if (!S->hasValidOffset()) {
114 assert((S->Offset == 0) && "Skipped section's offset must be zero");
115 assert((S->isVirtualSection() || S->Size == 0) &&
116 "Non-zero-fill sections with zero offset must have zero size");
117 continue;
118 }
119 assert((S->Offset != 0) &&
120 "Non-zero-fill section's offset cannot be zero");
121 Ends.push_back(S->Offset + S->Size);
122 if (S->RelOff)
123 Ends.push_back(S->RelOff +
124 S->NReloc * sizeof(MachO::any_relocation_info));
125 }
126
127 if (!Ends.empty())
128 return *llvm::max_element(Ends);
129
130 // Otherwise, we have only Mach header and load commands.
131 return headerSize() + loadCommandsSize();
132}
133
134void MachOWriter::writeHeader() {
136
137 Header.magic = O.Header.Magic;
138 Header.cputype = O.Header.CPUType;
139 Header.cpusubtype = O.Header.CPUSubType;
140 Header.filetype = O.Header.FileType;
141 Header.ncmds = O.Header.NCmds;
142 Header.sizeofcmds = O.Header.SizeOfCmds;
143 Header.flags = O.Header.Flags;
144 Header.reserved = O.Header.Reserved;
145
146 if (IsLittleEndian != sys::IsLittleEndianHost)
147 MachO::swapStruct(Header);
148
149 auto HeaderSize =
150 Is64Bit ? sizeof(MachO::mach_header_64) : sizeof(MachO::mach_header);
151 memcpy(Buf->getBufferStart(), &Header, HeaderSize);
152}
153
154void MachOWriter::writeLoadCommands() {
155 uint8_t *Begin =
156 reinterpret_cast<uint8_t *>(Buf->getBufferStart()) + headerSize();
157 for (const LoadCommand &LC : O.LoadCommands) {
158 // Construct a load command.
160 switch (MLC.load_command_data.cmd) {
161 case MachO::LC_SEGMENT:
162 if (IsLittleEndian != sys::IsLittleEndianHost)
163 MachO::swapStruct(MLC.segment_command_data);
164 memcpy(Begin, &MLC.segment_command_data, sizeof(MachO::segment_command));
165 Begin += sizeof(MachO::segment_command);
166
167 for (const std::unique_ptr<Section> &Sec : LC.Sections)
168 writeSectionInLoadCommand<MachO::section>(*Sec, Begin);
169 continue;
170 case MachO::LC_SEGMENT_64:
171 if (IsLittleEndian != sys::IsLittleEndianHost)
172 MachO::swapStruct(MLC.segment_command_64_data);
173 memcpy(Begin, &MLC.segment_command_64_data,
175 Begin += sizeof(MachO::segment_command_64);
176
177 for (const std::unique_ptr<Section> &Sec : LC.Sections)
178 writeSectionInLoadCommand<MachO::section_64>(*Sec, Begin);
179 continue;
180 }
181
182#define HANDLE_LOAD_COMMAND(LCName, LCValue, LCStruct) \
183 case MachO::LCName: \
184 assert(sizeof(MachO::LCStruct) + LC.Payload.size() == \
185 MLC.load_command_data.cmdsize); \
186 if (IsLittleEndian != sys::IsLittleEndianHost) \
187 MachO::swapStruct(MLC.LCStruct##_data); \
188 memcpy(Begin, &MLC.LCStruct##_data, sizeof(MachO::LCStruct)); \
189 Begin += sizeof(MachO::LCStruct); \
190 if (!LC.Payload.empty()) \
191 memcpy(Begin, LC.Payload.data(), LC.Payload.size()); \
192 Begin += LC.Payload.size(); \
193 break;
194
195 // Copy the load command as it is.
196 switch (MLC.load_command_data.cmd) {
197 default:
198 assert(sizeof(MachO::load_command) + LC.Payload.size() ==
199 MLC.load_command_data.cmdsize);
200 if (IsLittleEndian != sys::IsLittleEndianHost)
201 MachO::swapStruct(MLC.load_command_data);
202 memcpy(Begin, &MLC.load_command_data, sizeof(MachO::load_command));
203 Begin += sizeof(MachO::load_command);
204 if (!LC.Payload.empty())
205 memcpy(Begin, LC.Payload.data(), LC.Payload.size());
206 Begin += LC.Payload.size();
207 break;
208#include "llvm/BinaryFormat/MachO.def"
209 }
210 }
211}
212
213template <typename StructType>
214void MachOWriter::writeSectionInLoadCommand(const Section &Sec, uint8_t *&Out) {
215 StructType Temp;
216 assert(Sec.Segname.size() <= sizeof(Temp.segname) && "too long segment name");
217 assert(Sec.Sectname.size() <= sizeof(Temp.sectname) &&
218 "too long section name");
219 memset(&Temp, 0, sizeof(StructType));
220 memcpy(Temp.segname, Sec.Segname.data(), Sec.Segname.size());
221 memcpy(Temp.sectname, Sec.Sectname.data(), Sec.Sectname.size());
222 Temp.addr = Sec.Addr;
223 Temp.size = Sec.Size;
224 Temp.offset = Sec.Offset;
225 Temp.align = Sec.Align;
226 Temp.reloff = Sec.RelOff;
227 Temp.nreloc = Sec.NReloc;
228 Temp.flags = Sec.Flags;
229 Temp.reserved1 = Sec.Reserved1;
230 Temp.reserved2 = Sec.Reserved2;
231
232 if (IsLittleEndian != sys::IsLittleEndianHost)
233 MachO::swapStruct(Temp);
234 memcpy(Out, &Temp, sizeof(StructType));
235 Out += sizeof(StructType);
236}
237
238void MachOWriter::writeSections() {
239 for (const LoadCommand &LC : O.LoadCommands)
240 for (const std::unique_ptr<Section> &Sec : LC.Sections) {
241 if (!Sec->hasValidOffset()) {
242 assert((Sec->Offset == 0) && "Skipped section's offset must be zero");
243 assert((Sec->isVirtualSection() || Sec->Size == 0) &&
244 "Non-zero-fill sections with zero offset must have zero size");
245 continue;
246 }
247
248 assert(Sec->Offset && "Section offset can not be zero");
249 assert((Sec->Size == Sec->Content.size()) && "Incorrect section size");
250 memcpy(Buf->getBufferStart() + Sec->Offset, Sec->Content.data(),
251 Sec->Content.size());
252 for (size_t Index = 0; Index < Sec->Relocations.size(); ++Index) {
253 RelocationInfo RelocInfo = Sec->Relocations[Index];
254 if (!RelocInfo.Scattered && !RelocInfo.IsAddend) {
255 const uint32_t SymbolNum = RelocInfo.Extern
256 ? (*RelocInfo.Symbol)->Index
257 : (*RelocInfo.Sec)->Index;
258 RelocInfo.setPlainRelocationSymbolNum(SymbolNum, IsLittleEndian);
259 }
260 if (IsLittleEndian != sys::IsLittleEndianHost)
262 reinterpret_cast<MachO::any_relocation_info &>(RelocInfo.Info));
263 memcpy(Buf->getBufferStart() + Sec->RelOff +
264 Index * sizeof(MachO::any_relocation_info),
265 &RelocInfo.Info, sizeof(RelocInfo.Info));
266 }
267 }
268}
269
270template <typename NListType>
271void writeNListEntry(const SymbolEntry &SE, bool IsLittleEndian, char *&Out,
272 uint32_t Nstrx) {
273 NListType ListEntry;
274 ListEntry.n_strx = Nstrx;
275 ListEntry.n_type = SE.n_type;
276 ListEntry.n_sect = SE.n_sect;
277 ListEntry.n_desc = SE.n_desc;
278 ListEntry.n_value = SE.n_value;
279
280 if (IsLittleEndian != sys::IsLittleEndianHost)
281 MachO::swapStruct(ListEntry);
282 memcpy(Out, reinterpret_cast<const char *>(&ListEntry), sizeof(NListType));
283 Out += sizeof(NListType);
284}
285
286void MachOWriter::writeStringTable() {
287 if (!O.SymTabCommandIndex)
288 return;
289 const MachO::symtab_command &SymTabCommand =
291 .MachOLoadCommand.symtab_command_data;
292
293 uint8_t *StrTable = (uint8_t *)Buf->getBufferStart() + SymTabCommand.stroff;
294 LayoutBuilder.getStringTableBuilder().write(StrTable);
295}
296
297void MachOWriter::writeSymbolTable() {
298 if (!O.SymTabCommandIndex)
299 return;
300 const MachO::symtab_command &SymTabCommand =
302 .MachOLoadCommand.symtab_command_data;
303
304 char *SymTable = (char *)Buf->getBufferStart() + SymTabCommand.symoff;
305 for (auto &Symbol : O.SymTable.Symbols) {
306 SymbolEntry *Sym = Symbol.get();
307 uint32_t Nstrx = LayoutBuilder.getStringTableBuilder().getOffset(Sym->Name);
308
309 if (Is64Bit)
310 writeNListEntry<MachO::nlist_64>(*Sym, IsLittleEndian, SymTable, Nstrx);
311 else
312 writeNListEntry<MachO::nlist>(*Sym, IsLittleEndian, SymTable, Nstrx);
313 }
314}
315
316void MachOWriter::writeRebaseInfo() {
318 return;
319 const MachO::dyld_info_command &DyLdInfoCommand =
321 .MachOLoadCommand.dyld_info_command_data;
322 char *Out = (char *)Buf->getBufferStart() + DyLdInfoCommand.rebase_off;
323 assert((DyLdInfoCommand.rebase_size == O.Rebases.Opcodes.size()) &&
324 "Incorrect rebase opcodes size");
325 memcpy(Out, O.Rebases.Opcodes.data(), O.Rebases.Opcodes.size());
326}
327
328void MachOWriter::writeBindInfo() {
330 return;
331 const MachO::dyld_info_command &DyLdInfoCommand =
333 .MachOLoadCommand.dyld_info_command_data;
334 char *Out = (char *)Buf->getBufferStart() + DyLdInfoCommand.bind_off;
335 assert((DyLdInfoCommand.bind_size == O.Binds.Opcodes.size()) &&
336 "Incorrect bind opcodes size");
337 memcpy(Out, O.Binds.Opcodes.data(), O.Binds.Opcodes.size());
338}
339
340void MachOWriter::writeWeakBindInfo() {
342 return;
343 const MachO::dyld_info_command &DyLdInfoCommand =
345 .MachOLoadCommand.dyld_info_command_data;
346 char *Out = (char *)Buf->getBufferStart() + DyLdInfoCommand.weak_bind_off;
347 assert((DyLdInfoCommand.weak_bind_size == O.WeakBinds.Opcodes.size()) &&
348 "Incorrect weak bind opcodes size");
349 memcpy(Out, O.WeakBinds.Opcodes.data(), O.WeakBinds.Opcodes.size());
350}
351
352void MachOWriter::writeLazyBindInfo() {
354 return;
355 const MachO::dyld_info_command &DyLdInfoCommand =
357 .MachOLoadCommand.dyld_info_command_data;
358 char *Out = (char *)Buf->getBufferStart() + DyLdInfoCommand.lazy_bind_off;
359 assert((DyLdInfoCommand.lazy_bind_size == O.LazyBinds.Opcodes.size()) &&
360 "Incorrect lazy bind opcodes size");
361 memcpy(Out, O.LazyBinds.Opcodes.data(), O.LazyBinds.Opcodes.size());
362}
363
364void MachOWriter::writeExportInfo() {
366 return;
367 const MachO::dyld_info_command &DyLdInfoCommand =
369 .MachOLoadCommand.dyld_info_command_data;
370 char *Out = (char *)Buf->getBufferStart() + DyLdInfoCommand.export_off;
371 assert((DyLdInfoCommand.export_size == O.Exports.Trie.size()) &&
372 "Incorrect export trie size");
373 memcpy(Out, O.Exports.Trie.data(), O.Exports.Trie.size());
374}
375
376void MachOWriter::writeIndirectSymbolTable() {
378 return;
379
380 const MachO::dysymtab_command &DySymTabCommand =
382 .MachOLoadCommand.dysymtab_command_data;
383
384 uint32_t *Out =
385 (uint32_t *)(Buf->getBufferStart() + DySymTabCommand.indirectsymoff);
387 uint32_t Entry = (Sym.Symbol) ? (*Sym.Symbol)->Index : Sym.OriginalIndex;
388 if (IsLittleEndian != sys::IsLittleEndianHost)
389 sys::swapByteOrder(Entry);
390 *Out++ = Entry;
391 }
392}
393
394void MachOWriter::writeLinkData(std::optional<size_t> LCIndex,
395 const LinkData &LD) {
396 if (!LCIndex)
397 return;
398 const MachO::linkedit_data_command &LinkEditDataCommand =
399 O.LoadCommands[*LCIndex].MachOLoadCommand.linkedit_data_command_data;
400 char *Out = (char *)Buf->getBufferStart() + LinkEditDataCommand.dataoff;
401 assert((LinkEditDataCommand.datasize == LD.Data.size()) &&
402 "Incorrect data size");
403 memcpy(Out, LD.Data.data(), LD.Data.size());
404}
405
406static uint64_t
407getSegmentFileOffset(const LoadCommand &TextSegmentLoadCommand) {
408 const MachO::macho_load_command &MLC =
409 TextSegmentLoadCommand.MachOLoadCommand;
410 switch (MLC.load_command_data.cmd) {
411 case MachO::LC_SEGMENT:
412 return MLC.segment_command_data.fileoff;
413 case MachO::LC_SEGMENT_64:
414 return MLC.segment_command_64_data.fileoff;
415 default:
416 return 0;
417 }
418}
419
420static uint64_t getSegmentFileSize(const LoadCommand &TextSegmentLoadCommand) {
421 const MachO::macho_load_command &MLC =
422 TextSegmentLoadCommand.MachOLoadCommand;
423 switch (MLC.load_command_data.cmd) {
424 case MachO::LC_SEGMENT:
425 return MLC.segment_command_data.filesize;
426 case MachO::LC_SEGMENT_64:
427 return MLC.segment_command_64_data.filesize;
428 default:
429 return 0;
430 }
431}
432
433void MachOWriter::writeCodeSignatureData() {
434 // NOTE: This CodeSignature section behaviour must be kept in sync with that
435 // performed in LLD's CodeSignatureSection::write /
436 // CodeSignatureSection::writeHashes. Furthermore, this call must occur only
437 // after the rest of the binary has already been written to the buffer. This
438 // is because the buffer is read from to perform the necessary hashing.
439
440 // The CodeSignature section is the last section in the MachO binary and
441 // contains a hash of all content in the binary before it. Since llvm-objcopy
442 // has likely modified the target binary, the hash must be regenerated
443 // entirely. To generate this hash, we must read from the start of the binary
444 // (HashReadStart) to just before the start of the CodeSignature section
445 // (HashReadEnd).
446
447 const CodeSignatureInfo &CodeSignature = LayoutBuilder.getCodeSignature();
448
449 uint8_t *BufferStart = reinterpret_cast<uint8_t *>(Buf->getBufferStart());
450 uint8_t *HashReadStart = BufferStart;
451 uint8_t *HashReadEnd = BufferStart + CodeSignature.StartOffset;
452
453 // The CodeSignature section begins with a header, after which the hashes
454 // of each page of the binary are written.
455 uint8_t *HashWriteStart = HashReadEnd + CodeSignature.AllHeadersSize;
456
457 uint32_t TextSegmentFileOff = 0;
458 uint32_t TextSegmentFileSize = 0;
460 const LoadCommand &TextSegmentLoadCommand =
462 assert(TextSegmentLoadCommand.MachOLoadCommand.load_command_data.cmd ==
463 MachO::LC_SEGMENT ||
464 TextSegmentLoadCommand.MachOLoadCommand.load_command_data.cmd ==
465 MachO::LC_SEGMENT_64);
466 assert(StringRef(TextSegmentLoadCommand.MachOLoadCommand
467 .segment_command_data.segname) == "__TEXT");
468 TextSegmentFileOff = getSegmentFileOffset(TextSegmentLoadCommand);
469 TextSegmentFileSize = getSegmentFileSize(TextSegmentLoadCommand);
470 }
471
472 const uint32_t FileNamePad = CodeSignature.AllHeadersSize -
473 CodeSignature.FixedHeadersSize -
474 CodeSignature.OutputFileName.size();
475
476 // Write code section header.
477 auto *SuperBlob = reinterpret_cast<MachO::CS_SuperBlob *>(HashReadEnd);
479 write32be(&SuperBlob->length, CodeSignature.Size);
480 write32be(&SuperBlob->count, 1);
481 auto *BlobIndex = reinterpret_cast<MachO::CS_BlobIndex *>(&SuperBlob[1]);
482 write32be(&BlobIndex->type, MachO::CSSLOT_CODEDIRECTORY);
483 write32be(&BlobIndex->offset, CodeSignature.BlobHeadersSize);
484 auto *CodeDirectory = reinterpret_cast<MachO::CS_CodeDirectory *>(
485 HashReadEnd + CodeSignature.BlobHeadersSize);
486 write32be(&CodeDirectory->magic, MachO::CSMAGIC_CODEDIRECTORY);
487 write32be(&CodeDirectory->length,
488 CodeSignature.Size - CodeSignature.BlobHeadersSize);
489 write32be(&CodeDirectory->version, MachO::CS_SUPPORTSEXECSEG);
490 write32be(&CodeDirectory->flags, MachO::CS_ADHOC | MachO::CS_LINKER_SIGNED);
491 write32be(&CodeDirectory->hashOffset,
493 CodeSignature.OutputFileName.size() + FileNamePad);
494 write32be(&CodeDirectory->identOffset, sizeof(MachO::CS_CodeDirectory));
495 CodeDirectory->nSpecialSlots = 0;
496 write32be(&CodeDirectory->nCodeSlots, CodeSignature.BlockCount);
497 write32be(&CodeDirectory->codeLimit, CodeSignature.StartOffset);
498 CodeDirectory->hashSize = static_cast<uint8_t>(CodeSignature.HashSize);
499 CodeDirectory->hashType = MachO::kSecCodeSignatureHashSHA256;
500 CodeDirectory->platform = 0;
501 CodeDirectory->pageSize = CodeSignature.BlockSizeShift;
502 CodeDirectory->spare2 = 0;
503 CodeDirectory->scatterOffset = 0;
504 CodeDirectory->teamOffset = 0;
505 CodeDirectory->spare3 = 0;
506 CodeDirectory->codeLimit64 = 0;
507 write64be(&CodeDirectory->execSegBase, TextSegmentFileOff);
508 write64be(&CodeDirectory->execSegLimit, TextSegmentFileSize);
509 write64be(&CodeDirectory->execSegFlags, O.Header.FileType == MachO::MH_EXECUTE
511 : 0);
512
513 auto *Id = reinterpret_cast<char *>(&CodeDirectory[1]);
514 memcpy(Id, CodeSignature.OutputFileName.begin(),
515 CodeSignature.OutputFileName.size());
516 memset(Id + CodeSignature.OutputFileName.size(), 0, FileNamePad);
517
518 // Write the hashes.
519 uint8_t *CurrHashReadPosition = HashReadStart;
520 uint8_t *CurrHashWritePosition = HashWriteStart;
521 while (CurrHashReadPosition < HashReadEnd) {
522 StringRef Block(reinterpret_cast<char *>(CurrHashReadPosition),
523 std::min(static_cast<size_t>(HashReadEnd
524 - CurrHashReadPosition),
525 static_cast<size_t>(CodeSignature.BlockSize)));
526 SHA256 Hasher;
527 Hasher.update(Block);
528 std::array<uint8_t, 32> Hash = Hasher.final();
529 assert(Hash.size() == CodeSignature.HashSize);
530 memcpy(CurrHashWritePosition, Hash.data(), CodeSignature.HashSize);
531 CurrHashReadPosition += CodeSignature.BlockSize;
532 CurrHashWritePosition += CodeSignature.HashSize;
533 }
534#if defined(__APPLE__)
535 // This is macOS-specific work-around and makes no sense for any
536 // other host OS. See https://openradar.appspot.com/FB8914231
537 //
538 // The macOS kernel maintains a signature-verification cache to
539 // quickly validate applications at time of execve(2). The trouble
540 // is that for the kernel creates the cache entry at the time of the
541 // mmap(2) call, before we have a chance to write either the code to
542 // sign or the signature header+hashes. The fix is to invalidate
543 // all cached data associated with the output file, thus discarding
544 // the bogus prematurely-cached signature.
545 msync(BufferStart, CodeSignature.StartOffset + CodeSignature.Size,
546 MS_INVALIDATE);
547#endif
548}
549
550void MachOWriter::writeDataInCodeData() {
551 return writeLinkData(O.DataInCodeCommandIndex, O.DataInCode);
552}
553
554void MachOWriter::writeLinkerOptimizationHint() {
555 return writeLinkData(O.LinkerOptimizationHintCommandIndex,
557}
558
559void MachOWriter::writeFunctionStartsData() {
560 return writeLinkData(O.FunctionStartsCommandIndex, O.FunctionStarts);
561}
562
563void MachOWriter::writeDylibCodeSignDRsData() {
564 return writeLinkData(O.DylibCodeSignDRsIndex, O.DylibCodeSignDRs);
565}
566
567void MachOWriter::writeChainedFixupsData() {
568 return writeLinkData(O.ChainedFixupsCommandIndex, O.ChainedFixups);
569}
570
571void MachOWriter::writeExportsTrieData() {
573 return;
574 const MachO::linkedit_data_command &ExportsTrieCmd =
576 .MachOLoadCommand.linkedit_data_command_data;
577 char *Out = (char *)Buf->getBufferStart() + ExportsTrieCmd.dataoff;
578 assert((ExportsTrieCmd.datasize == O.Exports.Trie.size()) &&
579 "Incorrect export trie size");
580 memcpy(Out, O.Exports.Trie.data(), O.Exports.Trie.size());
581}
582
583void MachOWriter::writeTail() {
584 typedef void (MachOWriter::*WriteHandlerType)();
585 typedef std::pair<uint64_t, WriteHandlerType> WriteOperation;
587
588 if (O.SymTabCommandIndex) {
589 const MachO::symtab_command &SymTabCommand =
591 .MachOLoadCommand.symtab_command_data;
592 if (SymTabCommand.symoff)
593 Queue.push_back({SymTabCommand.symoff, &MachOWriter::writeSymbolTable});
594 if (SymTabCommand.stroff)
595 Queue.push_back({SymTabCommand.stroff, &MachOWriter::writeStringTable});
596 }
597
598 if (O.DyLdInfoCommandIndex) {
599 const MachO::dyld_info_command &DyLdInfoCommand =
601 .MachOLoadCommand.dyld_info_command_data;
602 if (DyLdInfoCommand.rebase_off)
603 Queue.push_back(
604 {DyLdInfoCommand.rebase_off, &MachOWriter::writeRebaseInfo});
605 if (DyLdInfoCommand.bind_off)
606 Queue.push_back({DyLdInfoCommand.bind_off, &MachOWriter::writeBindInfo});
607 if (DyLdInfoCommand.weak_bind_off)
608 Queue.push_back(
609 {DyLdInfoCommand.weak_bind_off, &MachOWriter::writeWeakBindInfo});
610 if (DyLdInfoCommand.lazy_bind_off)
611 Queue.push_back(
612 {DyLdInfoCommand.lazy_bind_off, &MachOWriter::writeLazyBindInfo});
613 if (DyLdInfoCommand.export_off)
614 Queue.push_back(
615 {DyLdInfoCommand.export_off, &MachOWriter::writeExportInfo});
616 }
617
618 if (O.DySymTabCommandIndex) {
619 const MachO::dysymtab_command &DySymTabCommand =
621 .MachOLoadCommand.dysymtab_command_data;
622
623 if (DySymTabCommand.indirectsymoff)
624 Queue.emplace_back(DySymTabCommand.indirectsymoff,
625 &MachOWriter::writeIndirectSymbolTable);
626 }
627
628 std::initializer_list<std::pair<std::optional<size_t>, WriteHandlerType>>
629 LinkEditDataCommandWriters = {
630 {O.CodeSignatureCommandIndex, &MachOWriter::writeCodeSignatureData},
631 {O.DylibCodeSignDRsIndex, &MachOWriter::writeDylibCodeSignDRsData},
632 {O.DataInCodeCommandIndex, &MachOWriter::writeDataInCodeData},
634 &MachOWriter::writeLinkerOptimizationHint},
635 {O.FunctionStartsCommandIndex, &MachOWriter::writeFunctionStartsData},
636 {O.ChainedFixupsCommandIndex, &MachOWriter::writeChainedFixupsData},
637 {O.ExportsTrieCommandIndex, &MachOWriter::writeExportsTrieData}};
638 for (const auto &W : LinkEditDataCommandWriters) {
639 std::optional<size_t> LinkEditDataCommandIndex;
640 WriteHandlerType WriteHandler;
641 std::tie(LinkEditDataCommandIndex, WriteHandler) = W;
642 if (LinkEditDataCommandIndex) {
643 const MachO::linkedit_data_command &LinkEditDataCommand =
644 O.LoadCommands[*LinkEditDataCommandIndex]
645 .MachOLoadCommand.linkedit_data_command_data;
646 if (LinkEditDataCommand.dataoff)
647 Queue.emplace_back(LinkEditDataCommand.dataoff, WriteHandler);
648 }
649 }
650
652
653 for (auto WriteOp : Queue)
654 (this->*WriteOp.second)();
655}
656
657Error MachOWriter::finalize() { return LayoutBuilder.layout(); }
658
660 size_t TotalSize = totalSize();
662 if (!Buf)
664 "failed to allocate memory buffer of " +
665 Twine::utohexstr(TotalSize) + " bytes");
666 writeHeader();
667 writeLoadCommands();
668 writeSections();
669 writeTail();
670
671 // TODO: Implement direct writing to the output stream (without intermediate
672 // memory buffer Buf).
673 Out.write(Buf->getBufferStart(), Buf->getBufferSize());
674 return Error::success();
675}
Symbol * Sym
Definition: ELF_riscv.cpp:479
static uint64_t getSegmentFileOffset(const LoadCommand &TextSegmentLoadCommand)
static uint64_t getSegmentFileSize(const LoadCommand &TextSegmentLoadCommand)
void writeNListEntry(const SymbolEntry &SE, bool IsLittleEndian, char *&Out, uint32_t Nstrx)
assert(ImpDefSCC.getReg()==AMDGPU::SCC &&ImpDefSCC.isDef())
This file contains some templates that are useful if you are working with the STL at all.
size_t size() const
size - Get the array size.
Definition: ArrayRef.h:168
const T * data() const
Definition: ArrayRef.h:165
Lightweight error class with error context and mandatory checking.
Definition: Error.h:160
static ErrorSuccess success()
Create a success value.
Definition: Error.h:337
std::array< uint8_t, 32 > final()
Return the current raw 256-bits SHA256 for the digested data since the last call to init().
Definition: SHA256.cpp:258
void update(ArrayRef< uint8_t > Data)
Digest more data.
Definition: SHA256.cpp:187
bool empty() const
Definition: SmallVector.h:81
void push_back(const T &Elt)
Definition: SmallVector.h:413
This is a 'vector' (really, a variable-sized array), optimized for the case when the array is small.
Definition: SmallVector.h:1196
StringRef - Represent a constant reference to a string, i.e.
Definition: StringRef.h:51
iterator begin() const
Definition: StringRef.h:116
constexpr size_t size() const
size - Get the string size.
Definition: StringRef.h:150
size_t getOffset(CachedHashStringRef S) const
Get the offest of a string in the string table.
void write(raw_ostream &OS) const
Class to represent struct types.
Definition: DerivedTypes.h:218
static Twine utohexstr(const uint64_t &Val)
Definition: Twine.h:416
static std::unique_ptr< WritableMemoryBuffer > getNewMemBuffer(size_t Size, const Twine &BufferName="")
Allocate a new zero-initialized MemoryBuffer of the specified size.
const CodeSignatureInfo & getCodeSignature() const
@ Entry
Definition: COFF.h:844
@ CS_EXECSEG_MAIN_BINARY
Definition: MachO.h:2278
@ MH_EXECUTE
Definition: MachO.h:44
@ CSMAGIC_EMBEDDED_SIGNATURE
Definition: MachO.h:2295
@ CSMAGIC_CODEDIRECTORY
Definition: MachO.h:2294
@ CSSLOT_CODEDIRECTORY
Definition: MachO.h:2309
@ CS_SUPPORTSEXECSEG
Definition: MachO.h:2305
@ kSecCodeSignatureHashSHA256
Definition: MachO.h:2404
void swapStruct(fat_header &mh)
Definition: MachO.h:1140
@ CS_LINKER_SIGNED
Definition: MachO.h:2241
@ CS_ADHOC
Definition: MachO.h:2218
void write32be(void *P, uint32_t V)
Definition: Endian.h:477
void write64be(void *P, uint64_t V)
Definition: Endian.h:480
static const bool IsLittleEndianHost
Definition: SwapByteOrder.h:29
void swapByteOrder(T &Value)
Definition: SwapByteOrder.h:61
This is an optimization pass for GlobalISel generic memory operations.
Definition: AddressRanges.h:18
Error createStringError(std::error_code EC, char const *Fmt, const Ts &... Vals)
Create formatted StringError object.
Definition: Error.h:1291
void sort(IteratorTy Start, IteratorTy End)
Definition: STLExtras.h:1664
auto max_element(R &&Range)
Provide wrappers to std::max_element which take ranges instead of having to pass begin/end explicitly...
Definition: STLExtras.h:2014
Function object to check whether the first component of a container supported by std::get (like std::...
Definition: STLExtras.h:1467
ArrayRef< uint8_t > Opcodes
Definition: MachOObject.h:230
When MachO binaries include a LC_CODE_SIGNATURE load command, the __LINKEDIT data segment will includ...
static constexpr uint32_t FixedHeadersSize
std::vector< IndirectSymbolEntry > Symbols
Definition: MachOObject.h:163
MachO::macho_load_command MachOLoadCommand
Definition: MachOObject.h:82
std::vector< std::unique_ptr< Section > > Sections
Definition: MachOObject.h:93
std::vector< uint8_t > Payload
Definition: MachOObject.h:87
std::optional< size_t > FunctionStartsCommandIndex
The index LC_FUNCTION_STARTS load command if present.
Definition: MachOObject.h:335
std::optional< size_t > ChainedFixupsCommandIndex
The index LC_DYLD_CHAINED_FIXUPS load command if present.
Definition: MachOObject.h:337
std::optional< size_t > ExportsTrieCommandIndex
The index LC_DYLD_EXPORTS_TRIE load command if present.
Definition: MachOObject.h:339
std::optional< size_t > DylibCodeSignDRsIndex
The index of LC_DYLIB_CODE_SIGN_DRS load command if present.
Definition: MachOObject.h:323
std::optional< size_t > SymTabCommandIndex
The index of LC_SYMTAB load command if present.
Definition: MachOObject.h:325
std::optional< size_t > DyLdInfoCommandIndex
The index of LC_DYLD_INFO or LC_DYLD_INFO_ONLY load command if present.
Definition: MachOObject.h:327
std::vector< LoadCommand > LoadCommands
Definition: MachOObject.h:300
std::optional< size_t > DataInCodeCommandIndex
The index LC_DATA_IN_CODE load command if present.
Definition: MachOObject.h:331
std::optional< size_t > DySymTabCommandIndex
The index LC_DYSYMTAB load command if present.
Definition: MachOObject.h:329
std::optional< size_t > TextSegmentCommandIndex
The index of the LC_SEGMENT or LC_SEGMENT_64 load command corresponding to the __TEXT segment.
Definition: MachOObject.h:342
IndirectSymbolTable IndirectSymTable
Definition: MachOObject.h:310
std::optional< size_t > CodeSignatureCommandIndex
The index of LC_CODE_SIGNATURE load command if present.
Definition: MachOObject.h:321
std::optional< size_t > LinkerOptimizationHintCommandIndex
The index of LC_LINKER_OPTIMIZATIN_HINT load command if present.
Definition: MachOObject.h:333
ArrayRef< uint8_t > Opcodes
Definition: MachOObject.h:214
MachO::any_relocation_info Info
Definition: MachOObject.h:184
std::optional< const SymbolEntry * > Symbol
Definition: MachOObject.h:174
std::optional< const Section * > Sec
Definition: MachOObject.h:176
void setPlainRelocationSymbolNum(unsigned SymbolNum, bool IsLittleEndian)
Definition: MachOObject.h:192
std::vector< std::unique_ptr< SymbolEntry > > Symbols
Definition: MachOObject.h:135