1 //===- bolt/Core/BinarySection.cpp - Section in a binary file -------------===//
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 // This file implements the BinarySection class.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #include "bolt/Core/BinarySection.h"
14 #include "bolt/Core/BinaryContext.h"
15 #include "bolt/Utils/Utils.h"
16 #include "llvm/MC/MCStreamer.h"
17 #include "llvm/Support/CommandLine.h"
18 
19 #define DEBUG_TYPE "bolt"
20 
21 using namespace llvm;
22 using namespace bolt;
23 
24 namespace opts {
25 extern cl::opt<bool> PrintRelocations;
26 extern cl::opt<bool> HotData;
27 } // namespace opts
28 
isELF() const29 bool BinarySection::isELF() const { return BC.isELF(); }
30 
isMachO() const31 bool BinarySection::isMachO() const { return BC.isMachO(); }
32 
33 uint64_t
hash(const BinaryData & BD,std::map<const BinaryData *,uint64_t> & Cache) const34 BinarySection::hash(const BinaryData &BD,
35                     std::map<const BinaryData *, uint64_t> &Cache) const {
36   auto Itr = Cache.find(&BD);
37   if (Itr != Cache.end())
38     return Itr->second;
39 
40   Cache[&BD] = 0;
41 
42   uint64_t Offset = BD.getAddress() - getAddress();
43   const uint64_t EndOffset = BD.getEndAddress() - getAddress();
44   auto Begin = Relocations.lower_bound(Relocation{Offset, 0, 0, 0, 0});
45   auto End = Relocations.upper_bound(Relocation{EndOffset, 0, 0, 0, 0});
46   const StringRef Contents = getContents();
47 
48   hash_code Hash =
49       hash_combine(hash_value(BD.getSize()), hash_value(BD.getSectionName()));
50 
51   while (Begin != End) {
52     const Relocation &Rel = *Begin++;
53     Hash = hash_combine(
54         Hash, hash_value(Contents.substr(Offset, Begin->Offset - Offset)));
55     if (BinaryData *RelBD = BC.getBinaryDataByName(Rel.Symbol->getName()))
56       Hash = hash_combine(Hash, hash(*RelBD, Cache));
57     Offset = Rel.Offset + Rel.getSize();
58   }
59 
60   Hash = hash_combine(Hash,
61                       hash_value(Contents.substr(Offset, EndOffset - Offset)));
62 
63   Cache[&BD] = Hash;
64 
65   return Hash;
66 }
67 
emitAsData(MCStreamer & Streamer,StringRef NewName) const68 void BinarySection::emitAsData(MCStreamer &Streamer, StringRef NewName) const {
69   StringRef SectionName = !NewName.empty() ? NewName : getName();
70   StringRef SectionContents = getContents();
71   MCSectionELF *ELFSection =
72       BC.Ctx->getELFSection(SectionName, getELFType(), getELFFlags());
73 
74   Streamer.switchSection(ELFSection);
75   Streamer.emitValueToAlignment(getAlignment());
76 
77   if (BC.HasRelocations && opts::HotData && isReordered())
78     Streamer.emitLabel(BC.Ctx->getOrCreateSymbol("__hot_data_start"));
79 
80   LLVM_DEBUG(dbgs() << "BOLT-DEBUG: emitting "
81                     << (isAllocatable() ? "" : "non-")
82                     << "allocatable data section " << SectionName << '\n');
83 
84   if (!hasRelocations()) {
85     Streamer.emitBytes(SectionContents);
86   } else {
87     uint64_t SectionOffset = 0;
88     for (const Relocation &Relocation : relocations()) {
89       assert(Relocation.Offset < SectionContents.size() && "overflow detected");
90       // Skip undefined symbols.
91       if (BC.UndefinedSymbols.count(Relocation.Symbol))
92         continue;
93       if (SectionOffset < Relocation.Offset) {
94         Streamer.emitBytes(SectionContents.substr(
95             SectionOffset, Relocation.Offset - SectionOffset));
96         SectionOffset = Relocation.Offset;
97       }
98       LLVM_DEBUG(dbgs() << "BOLT-DEBUG: emitting relocation for symbol "
99                         << (Relocation.Symbol ? Relocation.Symbol->getName()
100                                               : StringRef("<none>"))
101                         << " at offset 0x"
102                         << Twine::utohexstr(Relocation.Offset) << " with size "
103                         << Relocation::getSizeForType(Relocation.Type) << '\n');
104       size_t RelocationSize = Relocation.emit(&Streamer);
105       SectionOffset += RelocationSize;
106     }
107     assert(SectionOffset <= SectionContents.size() && "overflow error");
108     if (SectionOffset < SectionContents.size())
109       Streamer.emitBytes(SectionContents.substr(SectionOffset));
110   }
111 
112   if (BC.HasRelocations && opts::HotData && isReordered())
113     Streamer.emitLabel(BC.Ctx->getOrCreateSymbol("__hot_data_end"));
114 }
115 
flushPendingRelocations(raw_pwrite_stream & OS,SymbolResolverFuncTy Resolver)116 void BinarySection::flushPendingRelocations(raw_pwrite_stream &OS,
117                                             SymbolResolverFuncTy Resolver) {
118   if (PendingRelocations.empty() && Patches.empty())
119     return;
120 
121   const uint64_t SectionAddress = getAddress();
122 
123   // We apply relocations to original section contents. For allocatable sections
124   // this means using their input file offsets, since the output file offset
125   // could change (e.g. for new instance of .text). For non-allocatable
126   // sections, the output offset should always be a valid one.
127   const uint64_t SectionFileOffset =
128       isAllocatable() ? getInputFileOffset() : getOutputFileOffset();
129   LLVM_DEBUG(
130       dbgs() << "BOLT-DEBUG: flushing pending relocations for section "
131              << getName() << '\n'
132              << "  address: 0x" << Twine::utohexstr(SectionAddress) << '\n'
133              << "  offset: 0x" << Twine::utohexstr(SectionFileOffset) << '\n');
134 
135   for (BinaryPatch &Patch : Patches)
136     OS.pwrite(Patch.Bytes.data(), Patch.Bytes.size(),
137               SectionFileOffset + Patch.Offset);
138 
139   for (Relocation &Reloc : PendingRelocations) {
140     uint64_t Value = Reloc.Addend;
141     if (Reloc.Symbol)
142       Value += Resolver(Reloc.Symbol);
143 
144     Value = Relocation::adjustValue(Reloc.Type, Value,
145                                     SectionAddress + Reloc.Offset);
146 
147     OS.pwrite(reinterpret_cast<const char *>(&Value),
148               Relocation::getSizeForType(Reloc.Type),
149               SectionFileOffset + Reloc.Offset);
150 
151     LLVM_DEBUG(
152         dbgs() << "BOLT-DEBUG: writing value 0x" << Twine::utohexstr(Value)
153                << " of size " << Relocation::getSizeForType(Reloc.Type)
154                << " at section offset 0x" << Twine::utohexstr(Reloc.Offset)
155                << " address 0x"
156                << Twine::utohexstr(SectionAddress + Reloc.Offset)
157                << " file offset 0x"
158                << Twine::utohexstr(SectionFileOffset + Reloc.Offset) << '\n';);
159   }
160 
161   clearList(PendingRelocations);
162 }
163 
~BinarySection()164 BinarySection::~BinarySection() {
165   if (isReordered()) {
166     delete[] getData();
167     return;
168   }
169 
170   if (!isAllocatable() &&
171       (!hasSectionRef() ||
172        OutputContents.data() != getContents(Section).data())) {
173     delete[] getOutputData();
174   }
175 }
176 
clearRelocations()177 void BinarySection::clearRelocations() { clearList(Relocations); }
178 
print(raw_ostream & OS) const179 void BinarySection::print(raw_ostream &OS) const {
180   OS << getName() << ", "
181      << "0x" << Twine::utohexstr(getAddress()) << ", " << getSize() << " (0x"
182      << Twine::utohexstr(getOutputAddress()) << ", " << getOutputSize() << ")"
183      << ", data = " << getData() << ", output data = " << getOutputData();
184 
185   if (isAllocatable())
186     OS << " (allocatable)";
187 
188   if (isVirtual())
189     OS << " (virtual)";
190 
191   if (isTLS())
192     OS << " (tls)";
193 
194   if (opts::PrintRelocations)
195     for (const Relocation &R : relocations())
196       OS << "\n  " << R;
197 }
198 
199 BinarySection::RelocationSetType
reorderRelocations(bool Inplace) const200 BinarySection::reorderRelocations(bool Inplace) const {
201   assert(PendingRelocations.empty() &&
202          "reodering pending relocations not supported");
203   RelocationSetType NewRelocations;
204   for (const Relocation &Rel : relocations()) {
205     uint64_t RelAddr = Rel.Offset + getAddress();
206     BinaryData *BD = BC.getBinaryDataContainingAddress(RelAddr);
207     BD = BD->getAtomicRoot();
208     assert(BD);
209 
210     if ((!BD->isMoved() && !Inplace) || BD->isJumpTable())
211       continue;
212 
213     Relocation NewRel(Rel);
214     uint64_t RelOffset = RelAddr - BD->getAddress();
215     NewRel.Offset = BD->getOutputOffset() + RelOffset;
216     assert(NewRel.Offset < getSize());
217     LLVM_DEBUG(dbgs() << "BOLT-DEBUG: moving " << Rel << " -> " << NewRel
218                       << "\n");
219     auto Res = NewRelocations.emplace(std::move(NewRel));
220     (void)Res;
221     assert(Res.second && "Can't overwrite existing relocation");
222   }
223   return NewRelocations;
224 }
225 
reorderContents(const std::vector<BinaryData * > & Order,bool Inplace)226 void BinarySection::reorderContents(const std::vector<BinaryData *> &Order,
227                                     bool Inplace) {
228   IsReordered = true;
229 
230   Relocations = reorderRelocations(Inplace);
231 
232   std::string Str;
233   raw_string_ostream OS(Str);
234   const char *Src = Contents.data();
235   LLVM_DEBUG(dbgs() << "BOLT-DEBUG: reorderContents for " << Name << "\n");
236   for (BinaryData *BD : Order) {
237     assert((BD->isMoved() || !Inplace) && !BD->isJumpTable());
238     assert(BD->isAtomic() && BD->isMoveable());
239     const uint64_t SrcOffset = BD->getAddress() - getAddress();
240     assert(SrcOffset < Contents.size());
241     assert(SrcOffset == BD->getOffset());
242     while (OS.tell() < BD->getOutputOffset())
243       OS.write((unsigned char)0);
244     LLVM_DEBUG(dbgs() << "BOLT-DEBUG: " << BD->getName() << " @ " << OS.tell()
245                       << "\n");
246     OS.write(&Src[SrcOffset], BD->getOutputSize());
247   }
248   if (Relocations.empty()) {
249     // If there are no existing relocations, tack a phony one at the end
250     // of the reordered segment to force LLVM to recognize and map this
251     // section.
252     MCSymbol *ZeroSym = BC.registerNameAtAddress("Zero", 0, 0, 0);
253     addRelocation(OS.tell(), ZeroSym, ELF::R_X86_64_64, 0xdeadbeef);
254 
255     uint64_t Zero = 0;
256     OS.write(reinterpret_cast<const char *>(&Zero), sizeof(Zero));
257   }
258   auto *NewData = reinterpret_cast<char *>(copyByteArray(OS.str()));
259   Contents = OutputContents = StringRef(NewData, OS.str().size());
260   OutputSize = Contents.size();
261 }
262 
encodeELFNote(StringRef NameStr,StringRef DescStr,uint32_t Type)263 std::string BinarySection::encodeELFNote(StringRef NameStr, StringRef DescStr,
264                                          uint32_t Type) {
265   std::string Str;
266   raw_string_ostream OS(Str);
267   const uint32_t NameSz = NameStr.size() + 1;
268   const uint32_t DescSz = DescStr.size();
269   OS.write(reinterpret_cast<const char *>(&(NameSz)), 4);
270   OS.write(reinterpret_cast<const char *>(&(DescSz)), 4);
271   OS.write(reinterpret_cast<const char *>(&(Type)), 4);
272   OS << NameStr << '\0';
273   for (uint64_t I = NameSz; I < alignTo(NameSz, 4); ++I)
274     OS << '\0';
275   OS << DescStr;
276   for (uint64_t I = DescStr.size(); I < alignTo(DescStr.size(), 4); ++I)
277     OS << '\0';
278   return OS.str();
279 }
280