1//===- MachOObject.cpp - Mach-O object file model ---------------*- 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 "MachOObject.h"
10#include "llvm/ADT/SmallPtrSet.h"
11#include "llvm/Support/SystemZ/zOSSupport.h"
12#include <unordered_set>
13
14using namespace llvm;
15using namespace llvm::objcopy::macho;
16
17Section::Section(StringRef SegName, StringRef SectName)
18 : Segname(SegName), Sectname(SectName),
19 CanonicalName((Twine(SegName) + Twine(',') + SectName).str()) {}
20
21Section::Section(StringRef SegName, StringRef SectName, StringRef Content)
22 : Segname(SegName), Sectname(SectName),
23 CanonicalName((Twine(SegName) + Twine(',') + SectName).str()),
24 Content(Content) {}
25
26const SymbolEntry *SymbolTable::getSymbolByIndex(uint32_t Index) const {
27 assert(Index < Symbols.size() && "invalid symbol index");
28 return Symbols[Index].get();
29}
30
31SymbolEntry *SymbolTable::getSymbolByIndex(uint32_t Index) {
32 return const_cast<SymbolEntry *>(
33 static_cast<const SymbolTable *>(this)->getSymbolByIndex(Index));
34}
35
36void SymbolTable::updateSymbols(function_ref<void(SymbolEntry &)> Callable) {
37 for (auto &Sym : Symbols)
38 Callable(*Sym);
39
40 // Partition symbols: local < defined external < undefined external.
41 auto ExternalBegin = std::stable_partition(
42 first: std::begin(cont&: Symbols), last: std::end(cont&: Symbols),
43 pred: [](const auto &Sym) { return Sym->isLocalSymbol(); });
44 std::stable_partition(first: ExternalBegin, last: std::end(cont&: Symbols), pred: [](const auto &Sym) {
45 return !Sym->isUndefinedSymbol();
46 });
47}
48
49void SymbolTable::removeSymbols(
50 function_ref<bool(const std::unique_ptr<SymbolEntry> &)> ToRemove) {
51 llvm::erase_if(C&: Symbols, P: ToRemove);
52}
53
54void Object::updateLoadCommandIndexes() {
55 static constexpr char TextSegmentName[] = "__TEXT";
56 // Update indices of special load commands
57 for (size_t Index = 0, Size = LoadCommands.size(); Index < Size; ++Index) {
58 LoadCommand &LC = LoadCommands[Index];
59 switch (LC.MachOLoadCommand.load_command_data.cmd) {
60 case MachO::LC_CODE_SIGNATURE:
61 CodeSignatureCommandIndex = Index;
62 break;
63 case MachO::LC_SEGMENT:
64 if (StringRef(LC.MachOLoadCommand.segment_command_data.segname) ==
65 TextSegmentName)
66 TextSegmentCommandIndex = Index;
67 break;
68 case MachO::LC_SEGMENT_64:
69 if (StringRef(LC.MachOLoadCommand.segment_command_64_data.segname) ==
70 TextSegmentName)
71 TextSegmentCommandIndex = Index;
72 break;
73 case MachO::LC_SYMTAB:
74 SymTabCommandIndex = Index;
75 break;
76 case MachO::LC_DYSYMTAB:
77 DySymTabCommandIndex = Index;
78 break;
79 case MachO::LC_DYLD_INFO:
80 case MachO::LC_DYLD_INFO_ONLY:
81 DyLdInfoCommandIndex = Index;
82 break;
83 case MachO::LC_DATA_IN_CODE:
84 DataInCodeCommandIndex = Index;
85 break;
86 case MachO::LC_LINKER_OPTIMIZATION_HINT:
87 LinkerOptimizationHintCommandIndex = Index;
88 break;
89 case MachO::LC_FUNCTION_STARTS:
90 FunctionStartsCommandIndex = Index;
91 break;
92 case MachO::LC_DYLIB_CODE_SIGN_DRS:
93 DylibCodeSignDRsIndex = Index;
94 break;
95 case MachO::LC_DYLD_CHAINED_FIXUPS:
96 ChainedFixupsCommandIndex = Index;
97 break;
98 case MachO::LC_DYLD_EXPORTS_TRIE:
99 ExportsTrieCommandIndex = Index;
100 break;
101 }
102 }
103}
104
105Error Object::removeLoadCommands(
106 function_ref<bool(const LoadCommand &)> ToRemove) {
107 auto It = std::stable_partition(
108 first: LoadCommands.begin(), last: LoadCommands.end(),
109 pred: [&](const LoadCommand &LC) { return !ToRemove(LC); });
110 LoadCommands.erase(first: It, last: LoadCommands.end());
111
112 updateLoadCommandIndexes();
113 return Error::success();
114}
115
116Error Object::removeSections(
117 function_ref<bool(const std::unique_ptr<Section> &)> ToRemove) {
118 DenseMap<uint32_t, const Section *> OldIndexToSection;
119 uint32_t NextSectionIndex = 1;
120 for (LoadCommand &LC : LoadCommands) {
121 auto It = std::stable_partition(
122 first: std::begin(cont&: LC.Sections), last: std::end(cont&: LC.Sections),
123 pred: [&](const std::unique_ptr<Section> &Sec) { return !ToRemove(Sec); });
124 for (auto I = LC.Sections.begin(), End = It; I != End; ++I) {
125 OldIndexToSection[(*I)->Index] = I->get();
126 (*I)->Index = NextSectionIndex++;
127 }
128 LC.Sections.erase(first: It, last: LC.Sections.end());
129 }
130
131 auto IsDead = [&](const std::unique_ptr<SymbolEntry> &S) -> bool {
132 std::optional<uint32_t> Section = S->section();
133 return (Section && !OldIndexToSection.count(Val: *Section));
134 };
135
136 SmallPtrSet<const SymbolEntry *, 2> DeadSymbols;
137 for (const std::unique_ptr<SymbolEntry> &Sym : SymTable.Symbols)
138 if (IsDead(Sym))
139 DeadSymbols.insert(Ptr: Sym.get());
140
141 for (const LoadCommand &LC : LoadCommands)
142 for (const std::unique_ptr<Section> &Sec : LC.Sections)
143 for (const RelocationInfo &R : Sec->Relocations)
144 if (R.Symbol && *R.Symbol && DeadSymbols.count(Ptr: *R.Symbol))
145 return createStringError(EC: std::errc::invalid_argument,
146 Fmt: "symbol '%s' defined in section with index "
147 "'%u' cannot be removed because it is "
148 "referenced by a relocation in section '%s'",
149 Vals: (*R.Symbol)->Name.c_str(),
150 Vals: *((*R.Symbol)->section()),
151 Vals: Sec->CanonicalName.c_str());
152 SymTable.removeSymbols(ToRemove: IsDead);
153 for (std::unique_ptr<SymbolEntry> &S : SymTable.Symbols)
154 if (S->section())
155 S->n_sect = OldIndexToSection[S->n_sect]->Index;
156 return Error::success();
157}
158
159uint64_t Object::nextAvailableSegmentAddress() const {
160 uint64_t HeaderSize =
161 is64Bit() ? sizeof(MachO::mach_header_64) : sizeof(MachO::mach_header);
162 uint64_t Addr = HeaderSize + Header.SizeOfCmds;
163 for (const LoadCommand &LC : LoadCommands) {
164 const MachO::macho_load_command &MLC = LC.MachOLoadCommand;
165 switch (MLC.load_command_data.cmd) {
166 case MachO::LC_SEGMENT:
167 Addr = std::max(a: Addr,
168 b: static_cast<uint64_t>(MLC.segment_command_data.vmaddr) +
169 MLC.segment_command_data.vmsize);
170 break;
171 case MachO::LC_SEGMENT_64:
172 Addr = std::max(a: Addr, b: MLC.segment_command_64_data.vmaddr +
173 MLC.segment_command_64_data.vmsize);
174 break;
175 default:
176 continue;
177 }
178 }
179 return Addr;
180}
181
182template <typename SegmentType>
183static void
184constructSegment(SegmentType &Seg, llvm::MachO::LoadCommandType CmdType,
185 StringRef SegName, uint64_t SegVMAddr, uint64_t SegVMSize) {
186 assert(SegName.size() <= sizeof(Seg.segname) && "too long segment name");
187 memset(&Seg, 0, sizeof(SegmentType));
188 Seg.cmd = CmdType;
189 strncpy(Seg.segname, SegName.data(), SegName.size());
190 Seg.maxprot |=
191 (MachO::VM_PROT_READ | MachO::VM_PROT_WRITE | MachO::VM_PROT_EXECUTE);
192 Seg.initprot |=
193 (MachO::VM_PROT_READ | MachO::VM_PROT_WRITE | MachO::VM_PROT_EXECUTE);
194 Seg.vmaddr = SegVMAddr;
195 Seg.vmsize = SegVMSize;
196}
197
198LoadCommand &Object::addSegment(StringRef SegName, uint64_t SegVMSize) {
199 LoadCommand LC;
200 const uint64_t SegVMAddr = nextAvailableSegmentAddress();
201 if (is64Bit())
202 constructSegment(Seg&: LC.MachOLoadCommand.segment_command_64_data,
203 CmdType: MachO::LC_SEGMENT_64, SegName, SegVMAddr, SegVMSize);
204 else
205 constructSegment(Seg&: LC.MachOLoadCommand.segment_command_data,
206 CmdType: MachO::LC_SEGMENT, SegName, SegVMAddr, SegVMSize);
207
208 LoadCommands.push_back(x: std::move(LC));
209 return LoadCommands.back();
210}
211
212/// Extracts a segment name from a string which is possibly non-null-terminated.
213static StringRef extractSegmentName(const char *SegName) {
214 return StringRef(SegName,
215 strnlen(string: SegName, maxlen: sizeof(MachO::segment_command::segname)));
216}
217
218std::optional<StringRef> LoadCommand::getSegmentName() const {
219 const MachO::macho_load_command &MLC = MachOLoadCommand;
220 switch (MLC.load_command_data.cmd) {
221 case MachO::LC_SEGMENT:
222 return extractSegmentName(SegName: MLC.segment_command_data.segname);
223 case MachO::LC_SEGMENT_64:
224 return extractSegmentName(SegName: MLC.segment_command_64_data.segname);
225 default:
226 return std::nullopt;
227 }
228}
229
230std::optional<uint64_t> LoadCommand::getSegmentVMAddr() const {
231 const MachO::macho_load_command &MLC = MachOLoadCommand;
232 switch (MLC.load_command_data.cmd) {
233 case MachO::LC_SEGMENT:
234 return MLC.segment_command_data.vmaddr;
235 case MachO::LC_SEGMENT_64:
236 return MLC.segment_command_64_data.vmaddr;
237 default:
238 return std::nullopt;
239 }
240}
241