1 //===-- llvm-objdump.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 program is a utility that works like binutils "objdump", that is, it
11 // dumps out a plethora of information about an object file depending on the
12 // flags.
13 //
14 //===----------------------------------------------------------------------===//
15 
16 #include "MCFunction.h"
17 #include "llvm/Object/ObjectFile.h"
18 #include "llvm/ADT/OwningPtr.h"
19 #include "llvm/ADT/Triple.h"
20 #include "llvm/ADT/STLExtras.h"
21 #include "llvm/MC/MCAsmInfo.h"
22 #include "llvm/MC/MCDisassembler.h"
23 #include "llvm/MC/MCInst.h"
24 #include "llvm/MC/MCInstPrinter.h"
25 #include "llvm/MC/MCInstrDesc.h"
26 #include "llvm/MC/MCInstrInfo.h"
27 #include "llvm/Support/CommandLine.h"
28 #include "llvm/Support/Debug.h"
29 #include "llvm/Support/Format.h"
30 #include "llvm/Support/GraphWriter.h"
31 #include "llvm/Support/Host.h"
32 #include "llvm/Support/ManagedStatic.h"
33 #include "llvm/Support/MemoryBuffer.h"
34 #include "llvm/Support/MemoryObject.h"
35 #include "llvm/Support/PrettyStackTrace.h"
36 #include "llvm/Support/Signals.h"
37 #include "llvm/Support/SourceMgr.h"
38 #include "llvm/Support/raw_ostream.h"
39 #include "llvm/Support/system_error.h"
40 #include "llvm/Target/TargetRegistry.h"
41 #include "llvm/Target/TargetSelect.h"
42 #include <algorithm>
43 #include <cstring>
44 using namespace llvm;
45 using namespace object;
46 
47 namespace {
48   cl::list<std::string>
49   InputFilenames(cl::Positional, cl::desc("<input object files>"),
50                  cl::ZeroOrMore);
51 
52   cl::opt<bool>
53   Disassemble("disassemble",
54     cl::desc("Display assembler mnemonics for the machine instructions"));
55   cl::alias
56   Disassembled("d", cl::desc("Alias for --disassemble"),
57                cl::aliasopt(Disassemble));
58 
59   cl::opt<bool>
60   CFG("cfg", cl::desc("Create a CFG for every symbol in the object file and"
61                       "write it to a graphviz file"));
62 
63   cl::opt<std::string>
64   TripleName("triple", cl::desc("Target triple to disassemble for, "
65                                 "see -version for available targets"));
66 
67   cl::opt<std::string>
68   ArchName("arch", cl::desc("Target arch to disassemble for, "
69                             "see -version for available targets"));
70 
71   StringRef ToolName;
72 
73   bool error(error_code ec) {
74     if (!ec) return false;
75 
76     outs() << ToolName << ": error reading file: " << ec.message() << ".\n";
77     outs().flush();
78     return true;
79   }
80 }
81 
82 static const Target *GetTarget(const ObjectFile *Obj = NULL) {
83   // Figure out the target triple.
84   llvm::Triple TT("unknown-unknown-unknown");
85   if (TripleName.empty()) {
86     if (Obj)
87       TT.setArch(Triple::ArchType(Obj->getArch()));
88   } else
89     TT.setTriple(Triple::normalize(TripleName));
90 
91   if (!ArchName.empty())
92     TT.setArchName(ArchName);
93 
94   TripleName = TT.str();
95 
96   // Get the target specific parser.
97   std::string Error;
98   const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error);
99   if (TheTarget)
100     return TheTarget;
101 
102   errs() << ToolName << ": error: unable to get target for '" << TripleName
103          << "', see --version and --triple.\n";
104   return 0;
105 }
106 
107 namespace {
108 class StringRefMemoryObject : public MemoryObject {
109 private:
110   StringRef Bytes;
111 public:
112   StringRefMemoryObject(StringRef bytes) : Bytes(bytes) {}
113 
114   uint64_t getBase() const { return 0; }
115   uint64_t getExtent() const { return Bytes.size(); }
116 
117   int readByte(uint64_t Addr, uint8_t *Byte) const {
118     if (Addr >= getExtent())
119       return -1;
120     *Byte = Bytes[Addr];
121     return 0;
122   }
123 };
124 }
125 
126 static void DumpBytes(StringRef bytes) {
127   static char hex_rep[] = "0123456789abcdef";
128   // FIXME: The real way to do this is to figure out the longest instruction
129   //        and align to that size before printing. I'll fix this when I get
130   //        around to outputting relocations.
131   // 15 is the longest x86 instruction
132   // 3 is for the hex rep of a byte + a space.
133   // 1 is for the null terminator.
134   enum { OutputSize = (15 * 3) + 1 };
135   char output[OutputSize];
136 
137   assert(bytes.size() <= 15
138     && "DumpBytes only supports instructions of up to 15 bytes");
139   memset(output, ' ', sizeof(output));
140   unsigned index = 0;
141   for (StringRef::iterator i = bytes.begin(),
142                            e = bytes.end(); i != e; ++i) {
143     output[index] = hex_rep[(*i & 0xF0) >> 4];
144     output[index + 1] = hex_rep[*i & 0xF];
145     index += 3;
146   }
147 
148   output[sizeof(output) - 1] = 0;
149   outs() << output;
150 }
151 
152 static void DisassembleInput(const StringRef &Filename) {
153   OwningPtr<MemoryBuffer> Buff;
154 
155   if (error_code ec = MemoryBuffer::getFileOrSTDIN(Filename, Buff)) {
156     errs() << ToolName << ": " << Filename << ": " << ec.message() << "\n";
157     return;
158   }
159 
160   OwningPtr<ObjectFile> Obj(ObjectFile::createObjectFile(Buff.take()));
161 
162   const Target *TheTarget = GetTarget(Obj.get());
163   if (!TheTarget) {
164     // GetTarget prints out stuff.
165     return;
166   }
167   const MCInstrInfo *InstrInfo = TheTarget->createMCInstrInfo();
168   OwningPtr<MCInstrAnalysis>
169     InstrAnalysis(TheTarget->createMCInstrAnalysis(InstrInfo));
170 
171   outs() << '\n';
172   outs() << Filename
173          << ":\tfile format " << Obj->getFileFormatName() << "\n\n";
174 
175   error_code ec;
176   for (ObjectFile::section_iterator i = Obj->begin_sections(),
177                                     e = Obj->end_sections();
178                                     i != e; i.increment(ec)) {
179     if (error(ec)) break;
180     bool text;
181     if (error(i->isText(text))) break;
182     if (!text) continue;
183 
184     // Make a list of all the symbols in this section.
185     std::vector<std::pair<uint64_t, StringRef> > Symbols;
186     for (ObjectFile::symbol_iterator si = Obj->begin_symbols(),
187                                      se = Obj->end_symbols();
188                                      si != se; si.increment(ec)) {
189       bool contains;
190       if (!error(i->containsSymbol(*si, contains)) && contains) {
191         uint64_t Address;
192         if (error(si->getAddress(Address))) break;
193         StringRef Name;
194         if (error(si->getName(Name))) break;
195         Symbols.push_back(std::make_pair(Address, Name));
196       }
197     }
198 
199     // Sort the symbols by address, just in case they didn't come in that way.
200     array_pod_sort(Symbols.begin(), Symbols.end());
201 
202     StringRef name;
203     if (error(i->getName(name))) break;
204     outs() << "Disassembly of section " << name << ':';
205 
206     // If the section has no symbols just insert a dummy one and disassemble
207     // the whole section.
208     if (Symbols.empty())
209       Symbols.push_back(std::make_pair(0, name));
210 
211     // Set up disassembler.
212     OwningPtr<const MCAsmInfo> AsmInfo(TheTarget->createMCAsmInfo(TripleName));
213 
214     if (!AsmInfo) {
215       errs() << "error: no assembly info for target " << TripleName << "\n";
216       return;
217     }
218 
219     OwningPtr<const MCDisassembler> DisAsm(TheTarget->createMCDisassembler());
220     if (!DisAsm) {
221       errs() << "error: no disassembler for target " << TripleName << "\n";
222       return;
223     }
224 
225     int AsmPrinterVariant = AsmInfo->getAssemblerDialect();
226     OwningPtr<MCInstPrinter> IP(TheTarget->createMCInstPrinter(
227                                   AsmPrinterVariant, *AsmInfo));
228     if (!IP) {
229       errs() << "error: no instruction printer for target " << TripleName << '\n';
230       return;
231     }
232 
233     StringRef Bytes;
234     if (error(i->getContents(Bytes))) break;
235     StringRefMemoryObject memoryObject(Bytes);
236     uint64_t Size;
237     uint64_t Index;
238     uint64_t SectSize;
239     if (error(i->getSize(SectSize))) break;
240 
241     // Disassemble symbol by symbol.
242     for (unsigned si = 0, se = Symbols.size(); si != se; ++si) {
243       uint64_t Start = Symbols[si].first;
244       uint64_t End = si == se-1 ? SectSize : Symbols[si + 1].first - 1;
245       outs() << '\n' << Symbols[si].second << ":\n";
246 
247 #ifndef NDEBUG
248         raw_ostream &DebugOut = DebugFlag ? dbgs() : nulls();
249 #else
250         raw_ostream &DebugOut = nulls();
251 #endif
252 
253       if (!CFG) {
254         for (Index = Start; Index < End; Index += Size) {
255           MCInst Inst;
256           if (DisAsm->getInstruction(Inst, Size, memoryObject, Index,
257                                      DebugOut)) {
258             uint64_t addr;
259             if (error(i->getAddress(addr))) break;
260             outs() << format("%8x:\t", addr + Index);
261             DumpBytes(StringRef(Bytes.data() + Index, Size));
262             IP->printInst(&Inst, outs());
263             outs() << "\n";
264           } else {
265             errs() << ToolName << ": warning: invalid instruction encoding\n";
266             if (Size == 0)
267               Size = 1; // skip illegible bytes
268           }
269         }
270 
271       } else {
272         // Create CFG and use it for disassembly.
273         MCFunction f =
274           MCFunction::createFunctionFromMC(Symbols[si].second, DisAsm.get(),
275                                            memoryObject, Start, End,
276                                            InstrAnalysis.get(), DebugOut);
277 
278         for (MCFunction::iterator fi = f.begin(), fe = f.end(); fi != fe; ++fi){
279           bool hasPreds = false;
280           // Only print blocks that have predecessors.
281           // FIXME: Slow.
282           for (MCFunction::iterator pi = f.begin(), pe = f.end(); pi != pe;
283               ++pi)
284             if (pi->second.contains(&fi->second)) {
285               hasPreds = true;
286               break;
287             }
288 
289           // Data block.
290           if (!hasPreds && fi != f.begin()) {
291             uint64_t End = llvm::next(fi) == fe ? SectSize :
292                                                   llvm::next(fi)->first;
293             uint64_t addr;
294             if (error(i->getAddress(addr))) break;
295             outs() << "# " << End-fi->first << " bytes of data:\n";
296             for (unsigned pos = fi->first; pos != End; ++pos) {
297               outs() << format("%8x:\t", addr + pos);
298               DumpBytes(StringRef(Bytes.data() + pos, 1));
299               outs() << format("\t.byte 0x%02x\n", (uint8_t)Bytes[pos]);
300             }
301             continue;
302           }
303 
304           if (fi->second.contains(&fi->second))
305             outs() << "# Loop begin:\n";
306 
307           for (unsigned ii = 0, ie = fi->second.getInsts().size(); ii != ie;
308                ++ii) {
309             uint64_t addr;
310             if (error(i->getAddress(addr))) break;
311             const MCDecodedInst &Inst = fi->second.getInsts()[ii];
312             outs() << format("%8x:\t", addr + Inst.Address);
313             DumpBytes(StringRef(Bytes.data() + Inst.Address, Inst.Size));
314             // Simple loops.
315             if (fi->second.contains(&fi->second))
316               outs() << '\t';
317             IP->printInst(&Inst.Inst, outs());
318             outs() << '\n';
319           }
320         }
321 
322         // Start a new dot file.
323         std::string Error;
324         raw_fd_ostream Out((f.getName().str() + ".dot").c_str(), Error);
325         if (!Error.empty()) {
326           errs() << ToolName << ": warning: " << Error << '\n';
327           continue;
328         }
329 
330         Out << "digraph " << f.getName() << " {\n";
331         Out << "graph [ rankdir = \"LR\" ];\n";
332         for (MCFunction::iterator i = f.begin(), e = f.end(); i != e; ++i) {
333           bool hasPreds = false;
334           // Only print blocks that have predecessors.
335           // FIXME: Slow.
336           for (MCFunction::iterator pi = f.begin(), pe = f.end(); pi != pe;
337                ++pi)
338             if (pi->second.contains(&i->second)) {
339               hasPreds = true;
340               break;
341             }
342 
343           if (!hasPreds && i != f.begin())
344             continue;
345 
346           Out << '"' << (uintptr_t)&i->second << "\" [ label=\"<a>";
347           // Print instructions.
348           for (unsigned ii = 0, ie = i->second.getInsts().size(); ii != ie;
349                ++ii) {
350             // Escape special chars and print the instruction in mnemonic form.
351             std::string Str;
352             raw_string_ostream OS(Str);
353             IP->printInst(&i->second.getInsts()[ii].Inst, OS);
354             Out << DOT::EscapeString(OS.str()) << '|';
355           }
356           Out << "<o>\" shape=\"record\" ];\n";
357 
358           // Add edges.
359           for (MCBasicBlock::succ_iterator si = i->second.succ_begin(),
360               se = i->second.succ_end(); si != se; ++si)
361             Out << (uintptr_t)&i->second << ":o -> " << (uintptr_t)*si <<":a\n";
362         }
363         Out << "}\n";
364       }
365     }
366   }
367 }
368 
369 int main(int argc, char **argv) {
370   // Print a stack trace if we signal out.
371   sys::PrintStackTraceOnErrorSignal();
372   PrettyStackTraceProgram X(argc, argv);
373   llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
374 
375   // Initialize targets and assembly printers/parsers.
376   llvm::InitializeAllTargetInfos();
377   llvm::InitializeAllTargetMCs();
378   llvm::InitializeAllAsmParsers();
379   llvm::InitializeAllDisassemblers();
380 
381   cl::ParseCommandLineOptions(argc, argv, "llvm object file dumper\n");
382   TripleName = Triple::normalize(TripleName);
383 
384   ToolName = argv[0];
385 
386   // Defaults to a.out if no filenames specified.
387   if (InputFilenames.size() == 0)
388     InputFilenames.push_back("a.out");
389 
390   // -d is the only flag that is currently implemented, so just print help if
391   // it is not set.
392   if (!Disassemble) {
393     cl::PrintHelpMessage();
394     return 2;
395   }
396 
397   std::for_each(InputFilenames.begin(), InputFilenames.end(),
398                 DisassembleInput);
399 
400   return 0;
401 }
402