xref: /llvm-project/llvm/tools/llvm-objdump/llvm-objdump.cpp (revision 07f0f7762972cd1fc3a58b24f56df52930da4d7b) 
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         Address -= SectionAddr;
193 
194         StringRef Name;
195         if (error(si->getName(Name))) break;
196         Symbols.push_back(std::make_pair(Address, Name));
197       }
198     }
199 
200     // Sort the symbols by address, just in case they didn't come in that way.
201     array_pod_sort(Symbols.begin(), Symbols.end());
202 
203     // Make a list of all the relocations for this section.
204     std::vector<RelocationRef> Rels;
205     if (InlineRelocs) {
206       for (relocation_iterator ri = i->begin_relocations(),
207                                re = i->end_relocations();
208                               ri != re; ri.increment(ec)) {
209         if (error(ec)) break;
210         Rels.push_back(*ri);
211       }
212     }
213 
214     // Sort relocations by address.
215     std::sort(Rels.begin(), Rels.end(), RelocAddressLess);
216 
217     StringRef name;
218     if (error(i->getName(name))) break;
219     outs() << "Disassembly of section " << name << ':';
220 
221     // If the section has no symbols just insert a dummy one and disassemble
222     // the whole section.
223     if (Symbols.empty())
224       Symbols.push_back(std::make_pair(0, name));
225 
226     // Set up disassembler.
227     OwningPtr<const MCAsmInfo> AsmInfo(TheTarget->createMCAsmInfo(TripleName));
228 
229     if (!AsmInfo) {
230       errs() << "error: no assembly info for target " << TripleName << "\n";
231       return;
232     }
233 
234     OwningPtr<const MCSubtargetInfo> STI(
235       TheTarget->createMCSubtargetInfo(TripleName, "", ""));
236 
237     if (!STI) {
238       errs() << "error: no subtarget info for target " << TripleName << "\n";
239       return;
240     }
241 
242     OwningPtr<const MCDisassembler> DisAsm(
243       TheTarget->createMCDisassembler(*STI));
244     if (!DisAsm) {
245       errs() << "error: no disassembler for target " << TripleName << "\n";
246       return;
247     }
248 
249     int AsmPrinterVariant = AsmInfo->getAssemblerDialect();
250     OwningPtr<MCInstPrinter> IP(TheTarget->createMCInstPrinter(
251                                 AsmPrinterVariant, *AsmInfo, *STI));
252     if (!IP) {
253       errs() << "error: no instruction printer for target " << TripleName
254              << '\n';
255       return;
256     }
257 
258     StringRef Bytes;
259     if (error(i->getContents(Bytes))) break;
260     StringRefMemoryObject memoryObject(Bytes);
261     uint64_t Size;
262     uint64_t Index;
263     uint64_t SectSize;
264     if (error(i->getSize(SectSize))) break;
265 
266     std::vector<RelocationRef>::const_iterator rel_cur = Rels.begin();
267     std::vector<RelocationRef>::const_iterator rel_end = Rels.end();
268     // Disassemble symbol by symbol.
269     for (unsigned si = 0, se = Symbols.size(); si != se; ++si) {
270       uint64_t Start = Symbols[si].first;
271       uint64_t End;
272       // The end is either the size of the section or the beginning of the next
273       // symbol.
274       if (si == se - 1)
275         End = SectSize;
276       // Make sure this symbol takes up space.
277       else if (Symbols[si + 1].first != Start)
278         End = Symbols[si + 1].first - 1;
279       else
280         // This symbol has the same address as the next symbol. Skip it.
281         continue;
282 
283       outs() << '\n' << Symbols[si].second << ":\n";
284 
285 #ifndef NDEBUG
286         raw_ostream &DebugOut = DebugFlag ? dbgs() : nulls();
287 #else
288         raw_ostream &DebugOut = nulls();
289 #endif
290 
291       for (Index = Start; Index < End; Index += Size) {
292         MCInst Inst;
293 
294         if (DisAsm->getInstruction(Inst, Size, memoryObject, Index,
295                                    DebugOut, nulls())) {
296           outs() << format("%8"PRIx64":\t", SectionAddr + Index);
297           DumpBytes(StringRef(Bytes.data() + Index, Size));
298           IP->printInst(&Inst, outs(), "");
299           outs() << "\n";
300         } else {
301           errs() << ToolName << ": warning: invalid instruction encoding\n";
302           if (Size == 0)
303             Size = 1; // skip illegible bytes
304         }
305 
306         // Print relocation for instruction.
307         while (rel_cur != rel_end) {
308           bool hidden = false;
309           uint64_t addr;
310           SmallString<16> name;
311           SmallString<32> val;
312 
313           // If this relocation is hidden, skip it.
314           if (error(rel_cur->getHidden(hidden))) goto skip_print_rel;
315           if (hidden) goto skip_print_rel;
316 
317           if (error(rel_cur->getAddress(addr))) goto skip_print_rel;
318           // Stop when rel_cur's address is past the current instruction.
319           if (addr >= Index + Size) break;
320           if (error(rel_cur->getTypeName(name))) goto skip_print_rel;
321           if (error(rel_cur->getValueString(val))) goto skip_print_rel;
322 
323           outs() << format("\t\t\t%8"PRIx64": ", SectionAddr + addr) << name << "\t"
324                  << val << "\n";
325 
326         skip_print_rel:
327           ++rel_cur;
328         }
329       }
330     }
331   }
332 }
333 
334 static void PrintRelocations(const ObjectFile *o) {
335   error_code ec;
336   for (section_iterator si = o->begin_sections(), se = o->end_sections();
337                                                   si != se; si.increment(ec)){
338     if (error(ec)) return;
339     if (si->begin_relocations() == si->end_relocations())
340       continue;
341     StringRef secname;
342     if (error(si->getName(secname))) continue;
343     outs() << "RELOCATION RECORDS FOR [" << secname << "]:\n";
344     for (relocation_iterator ri = si->begin_relocations(),
345                              re = si->end_relocations();
346                              ri != re; ri.increment(ec)) {
347       if (error(ec)) return;
348 
349       bool hidden;
350       uint64_t address;
351       SmallString<32> relocname;
352       SmallString<32> valuestr;
353       if (error(ri->getHidden(hidden))) continue;
354       if (hidden) continue;
355       if (error(ri->getTypeName(relocname))) continue;
356       if (error(ri->getAddress(address))) continue;
357       if (error(ri->getValueString(valuestr))) continue;
358       outs() << address << " " << relocname << " " << valuestr << "\n";
359     }
360     outs() << "\n";
361   }
362 }
363 
364 static void PrintSectionHeaders(const ObjectFile *o) {
365   outs() << "Sections:\n"
366             "Idx Name          Size      Address          Type\n";
367   error_code ec;
368   unsigned i = 0;
369   for (section_iterator si = o->begin_sections(), se = o->end_sections();
370                                                   si != se; si.increment(ec)) {
371     if (error(ec)) return;
372     StringRef Name;
373     if (error(si->getName(Name))) return;
374     uint64_t Address;
375     if (error(si->getAddress(Address))) return;
376     uint64_t Size;
377     if (error(si->getSize(Size))) return;
378     bool Text, Data, BSS;
379     if (error(si->isText(Text))) return;
380     if (error(si->isData(Data))) return;
381     if (error(si->isBSS(BSS))) return;
382     std::string Type = (std::string(Text ? "TEXT " : "") +
383                         (Data ? "DATA " : "") + (BSS ? "BSS" : ""));
384     outs() << format("%3d %-13s %09"PRIx64" %017"PRIx64" %s\n", i, Name.str().c_str(), Size,
385                      Address, Type.c_str());
386     ++i;
387   }
388 }
389 
390 static void PrintSectionContents(const ObjectFile *o) {
391   error_code ec;
392   for (section_iterator si = o->begin_sections(),
393                         se = o->end_sections();
394                         si != se; si.increment(ec)) {
395     if (error(ec)) return;
396     StringRef Name;
397     StringRef Contents;
398     uint64_t BaseAddr;
399     if (error(si->getName(Name))) continue;
400     if (error(si->getContents(Contents))) continue;
401     if (error(si->getAddress(BaseAddr))) continue;
402 
403     outs() << "Contents of section " << Name << ":\n";
404 
405     // Dump out the content as hex and printable ascii characters.
406     for (std::size_t addr = 0, end = Contents.size(); addr < end; addr += 16) {
407       outs() << format(" %04"PRIx64" ", BaseAddr + addr);
408       // Dump line of hex.
409       for (std::size_t i = 0; i < 16; ++i) {
410         if (i != 0 && i % 4 == 0)
411           outs() << ' ';
412         if (addr + i < end)
413           outs() << hexdigit((Contents[addr + i] >> 4) & 0xF, true)
414                  << hexdigit(Contents[addr + i] & 0xF, true);
415         else
416           outs() << "  ";
417       }
418       // Print ascii.
419       outs() << "  ";
420       for (std::size_t i = 0; i < 16 && addr + i < end; ++i) {
421         if (std::isprint(Contents[addr + i] & 0xFF))
422           outs() << Contents[addr + i];
423         else
424           outs() << ".";
425       }
426       outs() << "\n";
427     }
428   }
429 }
430 
431 static void PrintCOFFSymbolTable(const COFFObjectFile *coff) {
432   const coff_file_header *header;
433   if (error(coff->getHeader(header))) return;
434   int aux_count = 0;
435   const coff_symbol *symbol = 0;
436   for (int i = 0, e = header->NumberOfSymbols; i != e; ++i) {
437     if (aux_count--) {
438       // Figure out which type of aux this is.
439       if (symbol->StorageClass == COFF::IMAGE_SYM_CLASS_STATIC
440           && symbol->Value == 0) { // Section definition.
441         const coff_aux_section_definition *asd;
442         if (error(coff->getAuxSymbol<coff_aux_section_definition>(i, asd)))
443           return;
444         outs() << "AUX "
445                << format("scnlen 0x%x nreloc %d nlnno %d checksum 0x%x "
446                          , unsigned(asd->Length)
447                          , unsigned(asd->NumberOfRelocations)
448                          , unsigned(asd->NumberOfLinenumbers)
449                          , unsigned(asd->CheckSum))
450                << format("assoc %d comdat %d\n"
451                          , unsigned(asd->Number)
452                          , unsigned(asd->Selection));
453       } else {
454         outs() << "AUX Unknown\n";
455       }
456     } else {
457       StringRef name;
458       if (error(coff->getSymbol(i, symbol))) return;
459       if (error(coff->getSymbolName(symbol, name))) return;
460       outs() << "[" << format("%2d", i) << "]"
461              << "(sec " << format("%2d", int(symbol->SectionNumber)) << ")"
462              << "(fl 0x00)" // Flag bits, which COFF doesn't have.
463              << "(ty " << format("%3x", unsigned(symbol->Type)) << ")"
464              << "(scl " << format("%3x", unsigned(symbol->StorageClass)) << ") "
465              << "(nx " << unsigned(symbol->NumberOfAuxSymbols) << ") "
466              << "0x" << format("%08x", unsigned(symbol->Value)) << " "
467              << name << "\n";
468       aux_count = symbol->NumberOfAuxSymbols;
469     }
470   }
471 }
472 
473 static void PrintSymbolTable(const ObjectFile *o) {
474   outs() << "SYMBOL TABLE:\n";
475 
476   if (const COFFObjectFile *coff = dyn_cast<const COFFObjectFile>(o))
477     PrintCOFFSymbolTable(coff);
478   else {
479     error_code ec;
480     for (symbol_iterator si = o->begin_symbols(),
481                          se = o->end_symbols(); si != se; si.increment(ec)) {
482       if (error(ec)) return;
483       StringRef Name;
484       uint64_t Address;
485       bool Global;
486       SymbolRef::Type Type;
487       bool Weak;
488       bool Absolute;
489       uint64_t Size;
490       section_iterator Section = o->end_sections();
491       if (error(si->getName(Name))) continue;
492       if (error(si->getAddress(Address))) continue;
493       if (error(si->isGlobal(Global))) continue;
494       if (error(si->getType(Type))) continue;
495       if (error(si->isWeak(Weak))) continue;
496       if (error(si->isAbsolute(Absolute))) continue;
497       if (error(si->getSize(Size))) continue;
498       if (error(si->getSection(Section))) continue;
499 
500       if (Address == UnknownAddressOrSize)
501         Address = 0;
502       if (Size == UnknownAddressOrSize)
503         Size = 0;
504       char GlobLoc = ' ';
505       if (Type != SymbolRef::ST_External)
506         GlobLoc = Global ? 'g' : 'l';
507       char Debug = (Type == SymbolRef::ST_Debug || Type == SymbolRef::ST_File)
508                    ? 'd' : ' ';
509       char FileFunc = ' ';
510       if (Type == SymbolRef::ST_File)
511         FileFunc = 'f';
512       else if (Type == SymbolRef::ST_Function)
513         FileFunc = 'F';
514 
515       outs() << format("%08"PRIx64, Address) << " "
516              << GlobLoc // Local -> 'l', Global -> 'g', Neither -> ' '
517              << (Weak ? 'w' : ' ') // Weak?
518              << ' ' // Constructor. Not supported yet.
519              << ' ' // Warning. Not supported yet.
520              << ' ' // Indirect reference to another symbol.
521              << Debug // Debugging (d) or dynamic (D) symbol.
522              << FileFunc // Name of function (F), file (f) or object (O).
523              << ' ';
524       if (Absolute)
525         outs() << "*ABS*";
526       else if (Section == o->end_sections())
527         outs() << "*UND*";
528       else {
529         StringRef SectionName;
530         if (error(Section->getName(SectionName)))
531           SectionName = "";
532         outs() << SectionName;
533       }
534       outs() << '\t'
535              << format("%08"PRIx64" ", Size)
536              << Name
537              << '\n';
538     }
539   }
540 }
541 
542 static void DumpObject(const ObjectFile *o) {
543   outs() << '\n';
544   outs() << o->getFileName()
545          << ":\tfile format " << o->getFileFormatName() << "\n\n";
546 
547   if (Disassemble)
548     DisassembleObject(o, Relocations);
549   if (Relocations && !Disassemble)
550     PrintRelocations(o);
551   if (SectionHeaders)
552     PrintSectionHeaders(o);
553   if (SectionContents)
554     PrintSectionContents(o);
555   if (SymbolTable)
556     PrintSymbolTable(o);
557 }
558 
559 /// @brief Dump each object file in \a a;
560 static void DumpArchive(const Archive *a) {
561   for (Archive::child_iterator i = a->begin_children(),
562                                e = a->end_children(); i != e; ++i) {
563     OwningPtr<Binary> child;
564     if (error_code ec = i->getAsBinary(child)) {
565       // Ignore non-object files.
566       if (ec != object_error::invalid_file_type)
567         errs() << ToolName << ": '" << a->getFileName() << "': " << ec.message()
568                << ".\n";
569       continue;
570     }
571     if (ObjectFile *o = dyn_cast<ObjectFile>(child.get()))
572       DumpObject(o);
573     else
574       errs() << ToolName << ": '" << a->getFileName() << "': "
575               << "Unrecognized file type.\n";
576   }
577 }
578 
579 /// @brief Open file and figure out how to dump it.
580 static void DumpInput(StringRef file) {
581   // If file isn't stdin, check that it exists.
582   if (file != "-" && !sys::fs::exists(file)) {
583     errs() << ToolName << ": '" << file << "': " << "No such file\n";
584     return;
585   }
586 
587   if (MachO && Disassemble) {
588     DisassembleInputMachO(file);
589     return;
590   }
591 
592   // Attempt to open the binary.
593   OwningPtr<Binary> binary;
594   if (error_code ec = createBinary(file, binary)) {
595     errs() << ToolName << ": '" << file << "': " << ec.message() << ".\n";
596     return;
597   }
598 
599   if (Archive *a = dyn_cast<Archive>(binary.get())) {
600     DumpArchive(a);
601   } else if (ObjectFile *o = dyn_cast<ObjectFile>(binary.get())) {
602     DumpObject(o);
603   } else {
604     errs() << ToolName << ": '" << file << "': " << "Unrecognized file type.\n";
605   }
606 }
607 
608 int main(int argc, char **argv) {
609   // Print a stack trace if we signal out.
610   sys::PrintStackTraceOnErrorSignal();
611   PrettyStackTraceProgram X(argc, argv);
612   llvm_shutdown_obj Y;  // Call llvm_shutdown() on exit.
613 
614   // Initialize targets and assembly printers/parsers.
615   llvm::InitializeAllTargetInfos();
616   llvm::InitializeAllTargetMCs();
617   llvm::InitializeAllAsmParsers();
618   llvm::InitializeAllDisassemblers();
619 
620   cl::ParseCommandLineOptions(argc, argv, "llvm object file dumper\n");
621   TripleName = Triple::normalize(TripleName);
622 
623   ToolName = argv[0];
624 
625   // Defaults to a.out if no filenames specified.
626   if (InputFilenames.size() == 0)
627     InputFilenames.push_back("a.out");
628 
629   if (!Disassemble
630       && !Relocations
631       && !SectionHeaders
632       && !SectionContents
633       && !SymbolTable) {
634     cl::PrintHelpMessage();
635     return 2;
636   }
637 
638   std::for_each(InputFilenames.begin(), InputFilenames.end(),
639                 DumpInput);
640 
641   return 0;
642 }
643