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 "llvm-objdump.h"
17 #include "MCFunction.h"
18 #include "llvm/Object/Archive.h"
19 #include "llvm/Object/COFF.h"
20 #include "llvm/Object/ObjectFile.h"
21 #include "llvm/ADT/OwningPtr.h"
22 #include "llvm/ADT/StringExtras.h"
23 #include "llvm/ADT/Triple.h"
24 #include "llvm/ADT/STLExtras.h"
25 #include "llvm/MC/MCAsmInfo.h"
26 #include "llvm/MC/MCDisassembler.h"
27 #include "llvm/MC/MCInst.h"
28 #include "llvm/MC/MCInstPrinter.h"
29 #include "llvm/MC/MCSubtargetInfo.h"
30 #include "llvm/Support/Casting.h"
31 #include "llvm/Support/CommandLine.h"
32 #include "llvm/Support/Debug.h"
33 #include "llvm/Support/FileSystem.h"
34 #include "llvm/Support/Format.h"
35 #include "llvm/Support/GraphWriter.h"
36 #include "llvm/Support/Host.h"
37 #include "llvm/Support/ManagedStatic.h"
38 #include "llvm/Support/MemoryBuffer.h"
39 #include "llvm/Support/MemoryObject.h"
40 #include "llvm/Support/PrettyStackTrace.h"
41 #include "llvm/Support/Signals.h"
42 #include "llvm/Support/SourceMgr.h"
43 #include "llvm/Support/TargetRegistry.h"
44 #include "llvm/Support/TargetSelect.h"
45 #include "llvm/Support/raw_ostream.h"
46 #include "llvm/Support/system_error.h"
47 #include <algorithm>
48 #include <cstring>
49 using namespace llvm;
50 using namespace object;
51 
52 static cl::list<std::string>
53 InputFilenames(cl::Positional, cl::desc("<input object files>"),cl::ZeroOrMore);
54 
55 static cl::opt<bool>
56 Disassemble("disassemble",
57   cl::desc("Display assembler mnemonics for the machine instructions"));
58 static cl::alias
59 Disassembled("d", cl::desc("Alias for --disassemble"),
60              cl::aliasopt(Disassemble));
61 
62 static cl::opt<bool>
63 Relocations("r", cl::desc("Display the relocation entries in the file"));
64 
65 static cl::opt<bool>
66 SectionContents("s", cl::desc("Display the content of each section"));
67 
68 static cl::opt<bool>
69 SymbolTable("t", cl::desc("Display the symbol table"));
70 
71 static cl::opt<bool>
72 MachO("macho", cl::desc("Use MachO specific object file parser"));
73 static cl::alias
74 MachOm("m", cl::desc("Alias for --macho"), cl::aliasopt(MachO));
75 
76 cl::opt<std::string>
77 llvm::TripleName("triple", cl::desc("Target triple to disassemble for, "
78                                     "see -version for available targets"));
79 
80 cl::opt<std::string>
81 llvm::ArchName("arch", cl::desc("Target arch to disassemble for, "
82                                 "see -version for available targets"));
83 
84 static cl::opt<bool>
85 SectionHeaders("section-headers", cl::desc("Display summaries of the headers "
86                                            "for each section."));
87 static cl::alias
88 SectionHeadersShort("headers", cl::desc("Alias for --section-headers"),
89                     cl::aliasopt(SectionHeaders));
90 static cl::alias
91 SectionHeadersShorter("h", cl::desc("Alias for --section-headers"),
92                       cl::aliasopt(SectionHeaders));
93 
94 static StringRef ToolName;
95 
96 static bool error(error_code ec) {
97   if (!ec) return false;
98 
99   outs() << ToolName << ": error reading file: " << ec.message() << ".\n";
100   outs().flush();
101   return true;
102 }
103 
104 static const Target *GetTarget(const ObjectFile *Obj = NULL) {
105   // Figure out the target triple.
106   llvm::Triple TT("unknown-unknown-unknown");
107   if (TripleName.empty()) {
108     if (Obj)
109       TT.setArch(Triple::ArchType(Obj->getArch()));
110   } else
111     TT.setTriple(Triple::normalize(TripleName));
112 
113   if (!ArchName.empty())
114     TT.setArchName(ArchName);
115 
116   TripleName = TT.str();
117 
118   // Get the target specific parser.
119   std::string Error;
120   const Target *TheTarget = TargetRegistry::lookupTarget(TripleName, Error);
121   if (TheTarget)
122     return TheTarget;
123 
124   errs() << ToolName << ": error: unable to get target for '" << TripleName
125          << "', see --version and --triple.\n";
126   return 0;
127 }
128 
129 void llvm::StringRefMemoryObject::anchor() { }
130 
131 void llvm::DumpBytes(StringRef bytes) {
132   static const char hex_rep[] = "0123456789abcdef";
133   // FIXME: The real way to do this is to figure out the longest instruction
134   //        and align to that size before printing. I'll fix this when I get
135   //        around to outputting relocations.
136   // 15 is the longest x86 instruction
137   // 3 is for the hex rep of a byte + a space.
138   // 1 is for the null terminator.
139   enum { OutputSize = (15 * 3) + 1 };
140   char output[OutputSize];
141 
142   assert(bytes.size() <= 15
143     && "DumpBytes only supports instructions of up to 15 bytes");
144   memset(output, ' ', sizeof(output));
145   unsigned index = 0;
146   for (StringRef::iterator i = bytes.begin(),
147                            e = bytes.end(); i != e; ++i) {
148     output[index] = hex_rep[(*i & 0xF0) >> 4];
149     output[index + 1] = hex_rep[*i & 0xF];
150     index += 3;
151   }
152 
153   output[sizeof(output) - 1] = 0;
154   outs() << output;
155 }
156 
157 static bool RelocAddressLess(RelocationRef a, RelocationRef b) {
158   uint64_t a_addr, b_addr;
159   if (error(a.getAddress(a_addr))) return false;
160   if (error(b.getAddress(b_addr))) return false;
161   return a_addr < b_addr;
162 }
163 
164 static void DisassembleObject(const ObjectFile *Obj, bool InlineRelocs) {
165   const Target *TheTarget = GetTarget(Obj);
166   if (!TheTarget) {
167     // GetTarget prints out stuff.
168     return;
169   }
170 
171   error_code ec;
172   for (section_iterator i = Obj->begin_sections(),
173                         e = Obj->end_sections();
174                         i != e; i.increment(ec)) {
175     if (error(ec)) break;
176     bool text;
177     if (error(i->isText(text))) break;
178     if (!text) continue;
179 
180     uint64_t SectionAddr;
181     if (error(i->getAddress(SectionAddr))) break;
182 
183     // Make a list of all the symbols in this section.
184     std::vector<std::pair<uint64_t, StringRef> > Symbols;
185     for (symbol_iterator si = Obj->begin_symbols(),
186                          se = Obj->end_symbols();
187                          si != se; si.increment(ec)) {
188       bool contains;
189       if (!error(i->containsSymbol(*si, contains)) && contains) {
190         uint64_t Address;
191         if (error(si->getAddress(Address))) break;
192         StringRef Name;
193         if (error(si->getName(Name))) break;
194         Symbols.push_back(std::make_pair(Address, Name));
195       }
196     }
197 
198     // Sort the symbols by address, just in case they didn't come in that way.
199     array_pod_sort(Symbols.begin(), Symbols.end());
200 
201     // Make a list of all the relocations for this section.
202     std::vector<RelocationRef> Rels;
203     if (InlineRelocs) {
204       for (relocation_iterator ri = i->begin_relocations(),
205                                re = i->end_relocations();
206                               ri != re; ri.increment(ec)) {
207         if (error(ec)) break;
208         Rels.push_back(*ri);
209       }
210     }
211 
212     // Sort relocations by address.
213     std::sort(Rels.begin(), Rels.end(), RelocAddressLess);
214 
215     StringRef name;
216     if (error(i->getName(name))) break;
217     outs() << "Disassembly of section " << name << ':';
218 
219     // If the section has no symbols just insert a dummy one and disassemble
220     // the whole section.
221     if (Symbols.empty())
222       Symbols.push_back(std::make_pair(0, name));
223 
224     // Set up disassembler.
225     OwningPtr<const MCAsmInfo> AsmInfo(TheTarget->createMCAsmInfo(TripleName));
226 
227     if (!AsmInfo) {
228       errs() << "error: no assembly info for target " << TripleName << "\n";
229       return;
230     }
231 
232     OwningPtr<const MCSubtargetInfo> STI(
233       TheTarget->createMCSubtargetInfo(TripleName, "", ""));
234 
235     if (!STI) {
236       errs() << "error: no subtarget info for target " << TripleName << "\n";
237       return;
238     }
239 
240     OwningPtr<const MCDisassembler> DisAsm(
241       TheTarget->createMCDisassembler(*STI));
242     if (!DisAsm) {
243       errs() << "error: no disassembler for target " << TripleName << "\n";
244       return;
245     }
246 
247     int AsmPrinterVariant = AsmInfo->getAssemblerDialect();
248     OwningPtr<MCInstPrinter> IP(TheTarget->createMCInstPrinter(
249                                 AsmPrinterVariant, *AsmInfo, *STI));
250     if (!IP) {
251       errs() << "error: no instruction printer for target " << TripleName
252              << '\n';
253       return;
254     }
255 
256     StringRef Bytes;
257     if (error(i->getContents(Bytes))) break;
258     StringRefMemoryObject memoryObject(Bytes);
259     uint64_t Size;
260     uint64_t Index;
261     uint64_t SectSize;
262     if (error(i->getSize(SectSize))) break;
263 
264     std::vector<RelocationRef>::const_iterator rel_cur = Rels.begin();
265     std::vector<RelocationRef>::const_iterator rel_end = Rels.end();
266     // Disassemble symbol by symbol.
267     for (unsigned si = 0, se = Symbols.size(); si != se; ++si) {
268       uint64_t Start = Symbols[si].first;
269       uint64_t End;
270       // The end is either the size of the section or the beginning of the next
271       // symbol.
272       if (si == se - 1)
273         End = SectSize;
274       // Make sure this symbol takes up space.
275       else if (Symbols[si + 1].first != Start)
276         End = Symbols[si + 1].first - 1;
277       else
278         // This symbol has the same address as the next symbol. Skip it.
279         continue;
280 
281       outs() << '\n' << Symbols[si].second << ":\n";
282 
283 #ifndef NDEBUG
284         raw_ostream &DebugOut = DebugFlag ? dbgs() : nulls();
285 #else
286         raw_ostream &DebugOut = nulls();
287 #endif
288 
289       for (Index = Start; Index < End; Index += Size) {
290         MCInst Inst;
291 
292         if (DisAsm->getInstruction(Inst, Size, memoryObject, Index,
293                                    DebugOut, nulls())) {
294           outs() << format("%8"PRIx64":\t", SectionAddr + Index);
295           DumpBytes(StringRef(Bytes.data() + Index, Size));
296           IP->printInst(&Inst, outs(), "");
297           outs() << "\n";
298         } else {
299           errs() << ToolName << ": warning: invalid instruction encoding\n";
300           if (Size == 0)
301             Size = 1; // skip illegible bytes
302         }
303 
304         // Print relocation for instruction.
305         while (rel_cur != rel_end) {
306           bool hidden = false;
307           uint64_t addr;
308           SmallString<16> name;
309           SmallString<32> val;
310 
311           // If this relocation is hidden, skip it.
312           if (error(rel_cur->getHidden(hidden))) goto skip_print_rel;
313           if (hidden) goto skip_print_rel;
314 
315           if (error(rel_cur->getAddress(addr))) goto skip_print_rel;
316           // Stop when rel_cur's address is past the current instruction.
317           if (addr >= Index + Size) break;
318           if (error(rel_cur->getTypeName(name))) goto skip_print_rel;
319           if (error(rel_cur->getValueString(val))) goto skip_print_rel;
320 
321           outs() << format("\t\t\t%8"PRIx64": ", SectionAddr + addr) << name << "\t"
322                  << val << "\n";
323 
324         skip_print_rel:
325           ++rel_cur;
326         }
327       }
328     }
329   }
330 }
331 
332 static void PrintRelocations(const ObjectFile *o) {
333   error_code ec;
334   for (section_iterator si = o->begin_sections(), se = o->end_sections();
335                                                   si != se; si.increment(ec)){
336     if (error(ec)) return;
337     if (si->begin_relocations() == si->end_relocations())
338       continue;
339     StringRef secname;
340     if (error(si->getName(secname))) continue;
341     outs() << "RELOCATION RECORDS FOR [" << secname << "]:\n";
342     for (relocation_iterator ri = si->begin_relocations(),
343                              re = si->end_relocations();
344                              ri != re; ri.increment(ec)) {
345       if (error(ec)) return;
346 
347       bool hidden;
348       uint64_t address;
349       SmallString<32> relocname;
350       SmallString<32> valuestr;
351       if (error(ri->getHidden(hidden))) continue;
352       if (hidden) continue;
353       if (error(ri->getTypeName(relocname))) continue;
354       if (error(ri->getAddress(address))) continue;
355       if (error(ri->getValueString(valuestr))) continue;
356       outs() << address << " " << relocname << " " << valuestr << "\n";
357     }
358     outs() << "\n";
359   }
360 }
361 
362 static void PrintSectionHeaders(const ObjectFile *o) {
363   outs() << "Sections:\n"
364             "Idx Name          Size      Address          Type\n";
365   error_code ec;
366   unsigned i = 0;
367   for (section_iterator si = o->begin_sections(), se = o->end_sections();
368                                                   si != se; si.increment(ec)) {
369     if (error(ec)) return;
370     StringRef Name;
371     if (error(si->getName(Name))) return;
372     uint64_t Address;
373     if (error(si->getAddress(Address))) return;
374     uint64_t Size;
375     if (error(si->getSize(Size))) return;
376     bool Text, Data, BSS;
377     if (error(si->isText(Text))) return;
378     if (error(si->isData(Data))) return;
379     if (error(si->isBSS(BSS))) return;
380     std::string Type = (std::string(Text ? "TEXT " : "") +
381                         (Data ? "DATA " : "") + (BSS ? "BSS" : ""));
382     outs() << format("%3d %-13s %09"PRIx64" %017"PRIx64" %s\n", i, Name.str().c_str(), Size,
383                      Address, Type.c_str());
384     ++i;
385   }
386 }
387 
388 static void PrintSectionContents(const ObjectFile *o) {
389   error_code ec;
390   for (section_iterator si = o->begin_sections(),
391                         se = o->end_sections();
392                         si != se; si.increment(ec)) {
393     if (error(ec)) return;
394     StringRef Name;
395     StringRef Contents;
396     uint64_t BaseAddr;
397     if (error(si->getName(Name))) continue;
398     if (error(si->getContents(Contents))) continue;
399     if (error(si->getAddress(BaseAddr))) continue;
400 
401     outs() << "Contents of section " << Name << ":\n";
402 
403     // Dump out the content as hex and printable ascii characters.
404     for (std::size_t addr = 0, end = Contents.size(); addr < end; addr += 16) {
405       outs() << format(" %04"PRIx64" ", BaseAddr + addr);
406       // Dump line of hex.
407       for (std::size_t i = 0; i < 16; ++i) {
408         if (i != 0 && i % 4 == 0)
409           outs() << ' ';
410         if (addr + i < end)
411           outs() << hexdigit((Contents[addr + i] >> 4) & 0xF, true)
412                  << hexdigit(Contents[addr + i] & 0xF, true);
413         else
414           outs() << "  ";
415       }
416       // Print ascii.
417       outs() << "  ";
418       for (std::size_t i = 0; i < 16 && addr + i < end; ++i) {
419         if (std::isprint(Contents[addr + i] & 0xFF))
420           outs() << Contents[addr + i];
421         else
422           outs() << ".";
423       }
424       outs() << "\n";
425     }
426   }
427 }
428 
429 static void PrintCOFFSymbolTable(const COFFObjectFile *coff) {
430   const coff_file_header *header;
431   if (error(coff->getHeader(header))) return;
432   int aux_count = 0;
433   const coff_symbol *symbol = 0;
434   for (int i = 0, e = header->NumberOfSymbols; i != e; ++i) {
435     if (aux_count--) {
436       // Figure out which type of aux this is.
437       if (symbol->StorageClass == COFF::IMAGE_SYM_CLASS_STATIC
438           && symbol->Value == 0) { // Section definition.
439         const coff_aux_section_definition *asd;
440         if (error(coff->getAuxSymbol<coff_aux_section_definition>(i, asd)))
441           return;
442         outs() << "AUX "
443                << format("scnlen 0x%x nreloc %d nlnno %d checksum 0x%x "
444                          , unsigned(asd->Length)
445                          , unsigned(asd->NumberOfRelocations)
446                          , unsigned(asd->NumberOfLinenumbers)
447                          , unsigned(asd->CheckSum))
448                << format("assoc %d comdat %d\n"
449                          , unsigned(asd->Number)
450                          , unsigned(asd->Selection));
451       } else {
452         outs() << "AUX Unknown\n";
453       }
454     } else {
455       StringRef name;
456       if (error(coff->getSymbol(i, symbol))) return;
457       if (error(coff->getSymbolName(symbol, name))) return;
458       outs() << "[" << format("%2d", i) << "]"
459              << "(sec " << format("%2d", int(symbol->SectionNumber)) << ")"
460              << "(fl 0x00)" // Flag bits, which COFF doesn't have.
461              << "(ty " << format("%3x", unsigned(symbol->Type)) << ")"
462              << "(scl " << format("%3x", unsigned(symbol->StorageClass)) << ") "
463              << "(nx " << unsigned(symbol->NumberOfAuxSymbols) << ") "
464              << "0x" << format("%08x", unsigned(symbol->Value)) << " "
465              << name << "\n";
466       aux_count = symbol->NumberOfAuxSymbols;
467     }
468   }
469 }
470 
471 static void PrintSymbolTable(const ObjectFile *o) {
472   outs() << "SYMBOL TABLE:\n";
473 
474   if (const COFFObjectFile *coff = dyn_cast<const COFFObjectFile>(o))
475     PrintCOFFSymbolTable(coff);
476   else {
477     error_code ec;
478     for (symbol_iterator si = o->begin_symbols(),
479                          se = o->end_symbols(); si != se; si.increment(ec)) {
480       if (error(ec)) return;
481       StringRef Name;
482       uint64_t Address;
483       bool Global;
484       SymbolRef::Type Type;
485       bool Weak;
486       bool Absolute;
487       uint64_t Size;
488       section_iterator Section = o->end_sections();
489       if (error(si->getName(Name))) continue;
490       if (error(si->getAddress(Address))) continue;
491       if (error(si->isGlobal(Global))) continue;
492       if (error(si->getType(Type))) continue;
493       if (error(si->isWeak(Weak))) continue;
494       if (error(si->isAbsolute(Absolute))) continue;
495       if (error(si->getSize(Size))) continue;
496       if (error(si->getSection(Section))) continue;
497 
498       if (Address == UnknownAddressOrSize)
499         Address = 0;
500       if (Size == UnknownAddressOrSize)
501         Size = 0;
502       char GlobLoc = ' ';
503       if (Type != SymbolRef::ST_External)
504         GlobLoc = Global ? 'g' : 'l';
505       char Debug = (Type == SymbolRef::ST_Debug || Type == SymbolRef::ST_File)
506                    ? 'd' : ' ';
507       char FileFunc = ' ';
508       if (Type == SymbolRef::ST_File)
509         FileFunc = 'f';
510       else if (Type == SymbolRef::ST_Function)
511         FileFunc = 'F';
512 
513       outs() << format("%08"PRIx64, Address) << " "
514              << GlobLoc // Local -> 'l', Global -> 'g', Neither -> ' '
515              << (Weak ? 'w' : ' ') // Weak?
516              << ' ' // Constructor. Not supported yet.
517              << ' ' // Warning. Not supported yet.
518              << ' ' // Indirect reference to another symbol.
519              << Debug // Debugging (d) or dynamic (D) symbol.
520              << FileFunc // Name of function (F), file (f) or object (O).
521              << ' ';
522       if (Absolute)
523         outs() << "*ABS*";
524       else if (Section == o->end_sections())
525         outs() << "*UND*";
526       else {
527         StringRef SectionName;
528         if (error(Section->getName(SectionName)))
529           SectionName = "";
530         outs() << SectionName;
531       }
532       outs() << '\t'
533              << format("%08"PRIx64" ", Size)
534              << Name
535              << '\n';
536     }
537   }
538 }
539 
540 static void DumpObject(const ObjectFile *o) {
541   outs() << '\n';
542   outs() << o->getFileName()
543          << ":\tfile format " << o->getFileFormatName() << "\n\n";
544 
545   if (Disassemble)
546     DisassembleObject(o, Relocations);
547   if (Relocations && !Disassemble)
548     PrintRelocations(o);
549   if (SectionHeaders)
550     PrintSectionHeaders(o);
551   if (SectionContents)
552     PrintSectionContents(o);
553   if (SymbolTable)
554     PrintSymbolTable(o);
555 }
556 
557 /// @brief Dump each object file in \a a;
558 static void DumpArchive(const Archive *a) {
559   for (Archive::child_iterator i = a->begin_children(),
560                                e = a->end_children(); i != e; ++i) {
561     OwningPtr<Binary> child;
562     if (error_code ec = i->getAsBinary(child)) {
563       // Ignore non-object files.
564       if (ec != object_error::invalid_file_type)
565         errs() << ToolName << ": '" << a->getFileName() << "': " << ec.message()
566                << ".\n";
567       continue;
568     }
569     if (ObjectFile *o = dyn_cast<ObjectFile>(child.get()))
570       DumpObject(o);
571     else
572       errs() << ToolName << ": '" << a->getFileName() << "': "
573               << "Unrecognized file type.\n";
574   }
575 }
576 
577 /// @brief Open file and figure out how to dump it.
578 static void DumpInput(StringRef file) {
579   // If file isn't stdin, check that it exists.
580   if (file != "-" && !sys::fs::exists(file)) {
581     errs() << ToolName << ": '" << file << "': " << "No such file\n";
582     return;
583   }
584 
585   if (MachO && Disassemble) {
586     DisassembleInputMachO(file);
587     return;
588   }
589 
590   // Attempt to open the binary.
591   OwningPtr<Binary> binary;
592   if (error_code ec = createBinary(file, binary)) {
593     errs() << ToolName << ": '" << file << "': " << ec.message() << ".\n";
594     return;
595   }
596 
597   if (Archive *a = dyn_cast<Archive>(binary.get())) {
598     DumpArchive(a);
599   } else if (ObjectFile *o = dyn_cast<ObjectFile>(binary.get())) {
600     DumpObject(o);
601   } else {
602     errs() << ToolName << ": '" << file << "': " << "Unrecognized file type.\n";
603   }
604 }
605 
606 int main(int argc, char **argv) {
607   // Print a stack trace if we signal out.
608   sys::PrintStackTraceOnErrorSignal();
609   PrettyStackTraceProgram X(argc, argv);
610   llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
611 
612   // Initialize targets and assembly printers/parsers.
613   llvm::InitializeAllTargetInfos();
614   llvm::InitializeAllTargetMCs();
615   llvm::InitializeAllAsmParsers();
616   llvm::InitializeAllDisassemblers();
617 
618   cl::ParseCommandLineOptions(argc, argv, "llvm object file dumper\n");
619   TripleName = Triple::normalize(TripleName);
620 
621   ToolName = argv[0];
622 
623   // Defaults to a.out if no filenames specified.
624   if (InputFilenames.size() == 0)
625     InputFilenames.push_back("a.out");
626 
627   if (!Disassemble
628       && !Relocations
629       && !SectionHeaders
630       && !SectionContents
631       && !SymbolTable) {
632     cl::PrintHelpMessage();
633     return 2;
634   }
635 
636   std::for_each(InputFilenames.begin(), InputFilenames.end(),
637                 DumpInput);
638 
639   return 0;
640 }
641