MachODump.cpp revision ef99356dfebb96f6f90efb912c2877214bad060e 
1//===-- MachODump.cpp - Object file dumping utility for llvm --------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file implements the MachO-specific dumper for llvm-objdump.
11//
12//===----------------------------------------------------------------------===//
13
14#include "llvm-objdump.h"
15#include "llvm/ADT/OwningPtr.h"
16#include "llvm/ADT/STLExtras.h"
17#include "llvm/ADT/StringExtras.h"
18#include "llvm/ADT/Triple.h"
19#include "llvm/DebugInfo/DIContext.h"
20#include "llvm/MC/MCAsmInfo.h"
21#include "llvm/MC/MCDisassembler.h"
22#include "llvm/MC/MCInst.h"
23#include "llvm/MC/MCInstPrinter.h"
24#include "llvm/MC/MCInstrAnalysis.h"
25#include "llvm/MC/MCInstrDesc.h"
26#include "llvm/MC/MCInstrInfo.h"
27#include "llvm/MC/MCRegisterInfo.h"
28#include "llvm/MC/MCSubtargetInfo.h"
29#include "llvm/Object/MachO.h"
30#include "llvm/Support/Casting.h"
31#include "llvm/Support/CommandLine.h"
32#include "llvm/Support/Debug.h"
33#include "llvm/Support/Format.h"
34#include "llvm/Support/GraphWriter.h"
35#include "llvm/Support/MachO.h"
36#include "llvm/Support/MemoryBuffer.h"
37#include "llvm/Support/TargetRegistry.h"
38#include "llvm/Support/TargetSelect.h"
39#include "llvm/Support/raw_ostream.h"
40#include "llvm/Support/system_error.h"
41#include <algorithm>
42#include <cstring>
43using namespace llvm;
44using namespace object;
45
46static cl::opt<bool>
47  UseDbg("g", cl::desc("Print line information from debug info if available"));
48
49static cl::opt<std::string>
50  DSYMFile("dsym", cl::desc("Use .dSYM file for debug info"));
51
52static const Target *GetTarget(const MachOObjectFile *MachOObj) {
53  // Figure out the target triple.
54  if (TripleName.empty()) {
55    llvm::Triple TT("unknown-unknown-unknown");
56    TT.setArch(Triple::ArchType(MachOObj->getArch()));
57    TripleName = TT.str();
58  }
59
60  // Get the target specific parser.
61  std::string Error;
62  const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error);
63  if (TheTarget)
64    return TheTarget;
65
66  errs() << "llvm-objdump: error: unable to get target for '" << TripleName
67         << "', see --version and --triple.\n";
68  return 0;
69}
70
71struct SymbolSorter {
72  bool operator()(const SymbolRef &A, const SymbolRef &B) {
73    SymbolRef::Type AType, BType;
74    A.getType(AType);
75    B.getType(BType);
76
77    uint64_t AAddr, BAddr;
78    if (AType != SymbolRef::ST_Function)
79      AAddr = 0;
80    else
81      A.getAddress(AAddr);
82    if (BType != SymbolRef::ST_Function)
83      BAddr = 0;
84    else
85      B.getAddress(BAddr);
86    return AAddr < BAddr;
87  }
88};
89
90static void
91getSectionsAndSymbols(const macho::Header Header,
92                      MachOObjectFile *MachOObj,
93                      std::vector<SectionRef> &Sections,
94                      std::vector<SymbolRef> &Symbols,
95                      SmallVectorImpl<uint64_t> &FoundFns) {
96  error_code ec;
97  for (symbol_iterator SI = MachOObj->begin_symbols(),
98       SE = MachOObj->end_symbols(); SI != SE; SI.increment(ec))
99    Symbols.push_back(*SI);
100
101  for (section_iterator SI = MachOObj->begin_sections(),
102       SE = MachOObj->end_sections(); SI != SE; SI.increment(ec)) {
103    SectionRef SR = *SI;
104    StringRef SectName;
105    SR.getName(SectName);
106    Sections.push_back(*SI);
107  }
108
109  MachOObjectFile::LoadCommandInfo Command =
110    MachOObj->getFirstLoadCommandInfo();
111  for (unsigned i = 0; ; ++i) {
112    if (Command.C.Type == macho::LCT_FunctionStarts) {
113      // We found a function starts segment, parse the addresses for later
114      // consumption.
115      macho::LinkeditDataLoadCommand LLC =
116        MachOObj->getLinkeditDataLoadCommand(Command);
117
118      MachOObj->ReadULEB128s(LLC.DataOffset, FoundFns);
119    }
120
121    if (i == Header.NumLoadCommands - 1)
122      break;
123    else
124      Command = MachOObj->getNextLoadCommandInfo(Command);
125  }
126}
127
128static void DisassembleInputMachO2(StringRef Filename,
129                                   MachOObjectFile *MachOOF);
130
131void llvm::DisassembleInputMachO(StringRef Filename) {
132  OwningPtr<MemoryBuffer> Buff;
133
134  if (error_code ec = MemoryBuffer::getFileOrSTDIN(Filename, Buff)) {
135    errs() << "llvm-objdump: " << Filename << ": " << ec.message() << "\n";
136    return;
137  }
138
139  OwningPtr<MachOObjectFile> MachOOF(static_cast<MachOObjectFile*>(
140        ObjectFile::createMachOObjectFile(Buff.take())));
141
142  DisassembleInputMachO2(Filename, MachOOF.get());
143}
144
145static void DisassembleInputMachO2(StringRef Filename,
146                                   MachOObjectFile *MachOOF) {
147  const Target *TheTarget = GetTarget(MachOOF);
148  if (!TheTarget) {
149    // GetTarget prints out stuff.
150    return;
151  }
152  OwningPtr<const MCInstrInfo> InstrInfo(TheTarget->createMCInstrInfo());
153  OwningPtr<MCInstrAnalysis>
154    InstrAnalysis(TheTarget->createMCInstrAnalysis(InstrInfo.get()));
155
156  // Set up disassembler.
157  OwningPtr<const MCRegisterInfo> MRI(TheTarget->createMCRegInfo(TripleName));
158  OwningPtr<const MCAsmInfo> AsmInfo(
159      TheTarget->createMCAsmInfo(*MRI, TripleName));
160  OwningPtr<const MCSubtargetInfo>
161    STI(TheTarget->createMCSubtargetInfo(TripleName, "", ""));
162  OwningPtr<const MCDisassembler> DisAsm(TheTarget->createMCDisassembler(*STI));
163  int AsmPrinterVariant = AsmInfo->getAssemblerDialect();
164  OwningPtr<MCInstPrinter>
165    IP(TheTarget->createMCInstPrinter(AsmPrinterVariant, *AsmInfo, *InstrInfo,
166                                      *MRI, *STI));
167
168  if (!InstrAnalysis || !AsmInfo || !STI || !DisAsm || !IP) {
169    errs() << "error: couldn't initialize disassembler for target "
170           << TripleName << '\n';
171    return;
172  }
173
174  outs() << '\n' << Filename << ":\n\n";
175
176  macho::Header Header = MachOOF->getHeader();
177
178  // FIXME: FoundFns isn't used anymore. Using symbols/LC_FUNCTION_STARTS to
179  // determine function locations will eventually go in MCObjectDisassembler.
180  // FIXME: Using the -cfg command line option, this code used to be able to
181  // annotate relocations with the referenced symbol's name, and if this was
182  // inside a __[cf]string section, the data it points to. This is now replaced
183  // by the upcoming MCSymbolizer, which needs the appropriate setup done above.
184  std::vector<SectionRef> Sections;
185  std::vector<SymbolRef> Symbols;
186  SmallVector<uint64_t, 8> FoundFns;
187
188  getSectionsAndSymbols(Header, MachOOF, Sections, Symbols, FoundFns);
189
190  // Make a copy of the unsorted symbol list. FIXME: duplication
191  std::vector<SymbolRef> UnsortedSymbols(Symbols);
192  // Sort the symbols by address, just in case they didn't come in that way.
193  std::sort(Symbols.begin(), Symbols.end(), SymbolSorter());
194
195#ifndef NDEBUG
196  raw_ostream &DebugOut = DebugFlag ? dbgs() : nulls();
197#else
198  raw_ostream &DebugOut = nulls();
199#endif
200
201  OwningPtr<DIContext> diContext;
202  ObjectFile *DbgObj = MachOOF;
203  // Try to find debug info and set up the DIContext for it.
204  if (UseDbg) {
205    // A separate DSym file path was specified, parse it as a macho file,
206    // get the sections and supply it to the section name parsing machinery.
207    if (!DSYMFile.empty()) {
208      OwningPtr<MemoryBuffer> Buf;
209      if (error_code ec = MemoryBuffer::getFileOrSTDIN(DSYMFile.c_str(), Buf)) {
210        errs() << "llvm-objdump: " << Filename << ": " << ec.message() << '\n';
211        return;
212      }
213      DbgObj = ObjectFile::createMachOObjectFile(Buf.take());
214    }
215
216    // Setup the DIContext
217    diContext.reset(DIContext::getDWARFContext(DbgObj));
218  }
219
220  for (unsigned SectIdx = 0; SectIdx != Sections.size(); SectIdx++) {
221
222    bool SectIsText = false;
223    Sections[SectIdx].isText(SectIsText);
224    if (SectIsText == false)
225      continue;
226
227    StringRef SectName;
228    if (Sections[SectIdx].getName(SectName) ||
229        SectName != "__text")
230      continue; // Skip non-text sections
231
232    DataRefImpl DR = Sections[SectIdx].getRawDataRefImpl();
233
234    StringRef SegmentName = MachOOF->getSectionFinalSegmentName(DR);
235    if (SegmentName != "__TEXT")
236      continue;
237
238    StringRef Bytes;
239    Sections[SectIdx].getContents(Bytes);
240    StringRefMemoryObject memoryObject(Bytes);
241    bool symbolTableWorked = false;
242
243    // Parse relocations.
244    std::vector<std::pair<uint64_t, SymbolRef> > Relocs;
245    error_code ec;
246    for (relocation_iterator RI = Sections[SectIdx].begin_relocations(),
247         RE = Sections[SectIdx].end_relocations(); RI != RE; RI.increment(ec)) {
248      uint64_t RelocOffset, SectionAddress;
249      RI->getOffset(RelocOffset);
250      Sections[SectIdx].getAddress(SectionAddress);
251      RelocOffset -= SectionAddress;
252
253      SymbolRef RelocSym;
254      RI->getSymbol(RelocSym);
255
256      Relocs.push_back(std::make_pair(RelocOffset, RelocSym));
257    }
258    array_pod_sort(Relocs.begin(), Relocs.end());
259
260    // Disassemble symbol by symbol.
261    for (unsigned SymIdx = 0; SymIdx != Symbols.size(); SymIdx++) {
262      StringRef SymName;
263      Symbols[SymIdx].getName(SymName);
264
265      SymbolRef::Type ST;
266      Symbols[SymIdx].getType(ST);
267      if (ST != SymbolRef::ST_Function)
268        continue;
269
270      // Make sure the symbol is defined in this section.
271      bool containsSym = false;
272      Sections[SectIdx].containsSymbol(Symbols[SymIdx], containsSym);
273      if (!containsSym)
274        continue;
275
276      // Start at the address of the symbol relative to the section's address.
277      uint64_t SectionAddress = 0;
278      uint64_t Start = 0;
279      Sections[SectIdx].getAddress(SectionAddress);
280      Symbols[SymIdx].getAddress(Start);
281      Start -= SectionAddress;
282
283      // Stop disassembling either at the beginning of the next symbol or at
284      // the end of the section.
285      bool containsNextSym = false;
286      uint64_t NextSym = 0;
287      uint64_t NextSymIdx = SymIdx+1;
288      while (Symbols.size() > NextSymIdx) {
289        SymbolRef::Type NextSymType;
290        Symbols[NextSymIdx].getType(NextSymType);
291        if (NextSymType == SymbolRef::ST_Function) {
292          Sections[SectIdx].containsSymbol(Symbols[NextSymIdx],
293                                           containsNextSym);
294          Symbols[NextSymIdx].getAddress(NextSym);
295          NextSym -= SectionAddress;
296          break;
297        }
298        ++NextSymIdx;
299      }
300
301      uint64_t SectSize;
302      Sections[SectIdx].getSize(SectSize);
303      uint64_t End = containsNextSym ?  NextSym : SectSize;
304      uint64_t Size;
305
306      symbolTableWorked = true;
307
308      outs() << SymName << ":\n";
309      DILineInfo lastLine;
310      for (uint64_t Index = Start; Index < End; Index += Size) {
311        MCInst Inst;
312
313        if (DisAsm->getInstruction(Inst, Size, memoryObject, Index,
314                                   DebugOut, nulls())) {
315          uint64_t SectAddress = 0;
316          Sections[SectIdx].getAddress(SectAddress);
317          outs() << format("%8" PRIx64 ":\t", SectAddress + Index);
318
319          DumpBytes(StringRef(Bytes.data() + Index, Size));
320          IP->printInst(&Inst, outs(), "");
321
322          // Print debug info.
323          if (diContext) {
324            DILineInfo dli =
325              diContext->getLineInfoForAddress(SectAddress + Index);
326            // Print valid line info if it changed.
327            if (dli != lastLine && dli.getLine() != 0)
328              outs() << "\t## " << dli.getFileName() << ':'
329                << dli.getLine() << ':' << dli.getColumn();
330            lastLine = dli;
331          }
332          outs() << "\n";
333        } else {
334          errs() << "llvm-objdump: warning: invalid instruction encoding\n";
335          if (Size == 0)
336            Size = 1; // skip illegible bytes
337        }
338      }
339    }
340    if (!symbolTableWorked) {
341      // Reading the symbol table didn't work, disassemble the whole section.
342      uint64_t SectAddress;
343      Sections[SectIdx].getAddress(SectAddress);
344      uint64_t SectSize;
345      Sections[SectIdx].getSize(SectSize);
346      uint64_t InstSize;
347      for (uint64_t Index = 0; Index < SectSize; Index += InstSize) {
348        MCInst Inst;
349
350        if (DisAsm->getInstruction(Inst, InstSize, memoryObject, Index,
351                                   DebugOut, nulls())) {
352          outs() << format("%8" PRIx64 ":\t", SectAddress + Index);
353          DumpBytes(StringRef(Bytes.data() + Index, InstSize));
354          IP->printInst(&Inst, outs(), "");
355          outs() << "\n";
356        } else {
357          errs() << "llvm-objdump: warning: invalid instruction encoding\n";
358          if (InstSize == 0)
359            InstSize = 1; // skip illegible bytes
360        }
361      }
362    }
363  }
364}
365