lib/CodeGen/MachOWriter.cpp

   1 //===-- MachOWriter.cpp - Target-independent Mach-O Writer code -----------===//
   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 target-independent Mach-O writer.  This file writes
  11 // out the Mach-O file in the following order:
  12 //
  13 //  #1 FatHeader (universal-only)
  14 //  #2 FatArch (universal-only, 1 per universal arch)
  15 //  Per arch:
  16 //    #3 Header
  17 //    #4 Load Commands
  18 //    #5 Sections
  19 //    #6 Relocations
  20 //    #7 Symbols
  21 //    #8 Strings
  22 //
  23 //===----------------------------------------------------------------------===//
  24
  25 #include "MachOWriter.h"
  26 #include "llvm/Constants.h"
  27 #include "llvm/DerivedTypes.h"
  28 #include "llvm/Module.h"
  29 #include "llvm/PassManager.h"
  30 #include "llvm/CodeGen/FileWriters.h"
  31 #include "llvm/CodeGen/MachineCodeEmitter.h"
  32 #include "llvm/CodeGen/MachineConstantPool.h"
  33 #include "llvm/CodeGen/MachineJumpTableInfo.h"
  34 #include "llvm/Target/TargetAsmInfo.h"
  35 #include "llvm/Target/TargetJITInfo.h"
  36 #include "llvm/Support/Mangler.h"
  37 #include "llvm/Support/MathExtras.h"
  38 #include "llvm/Support/OutputBuffer.h"
  39 #include "llvm/Support/Streams.h"
  40 #include "llvm/Support/raw_ostream.h"
  41 #include <algorithm>
  42 #include <cstring>
  43 using namespace llvm;
  44
  45 /// AddMachOWriter - Concrete function to add the Mach-O writer to the function
  46 /// pass manager.
  47 MachineCodeEmitter *llvm::AddMachOWriter(PassManagerBase &PM,
  48                                          raw_ostream &O,
  49                                          TargetMachine &TM) {
  50   MachOWriter *MOW = new MachOWriter(O, TM);
  51   PM.add(MOW);
  52   return &MOW->getMachineCodeEmitter();
  53 }
  54
  55 //===----------------------------------------------------------------------===//
  56 //                       MachOCodeEmitter Implementation
  57 //===----------------------------------------------------------------------===//
  58
  59 namespace llvm {
  60   /// MachOCodeEmitter - This class is used by the MachOWriter to emit the code
  61   /// for functions to the Mach-O file.
  62   class MachOCodeEmitter : public MachineCodeEmitter {
  63     MachOWriter &MOW;
  64
  65     /// Target machine description.
  66     TargetMachine &TM;
  67
  68     /// is64Bit/isLittleEndian - This information is inferred from the target
  69     /// machine directly, indicating what header values and flags to set.
  70     bool is64Bit, isLittleEndian;
  71
  72     /// Relocations - These are the relocations that the function needs, as
  73     /// emitted.
  74     std::vector<MachineRelocation> Relocations;
  75
  76     /// CPLocations - This is a map of constant pool indices to offsets from the
  77     /// start of the section for that constant pool index.
  78     std::vector<uintptr_t> CPLocations;
  79
  80     /// CPSections - This is a map of constant pool indices to the MachOSection
  81     /// containing the constant pool entry for that index.
  82     std::vector<unsigned> CPSections;
  83
  84     /// JTLocations - This is a map of jump table indices to offsets from the
  85     /// start of the section for that jump table index.
  86     std::vector<uintptr_t> JTLocations;
  87
  88     /// MBBLocations - This vector is a mapping from MBB ID's to their address.
  89     /// It is filled in by the StartMachineBasicBlock callback and queried by
  90     /// the getMachineBasicBlockAddress callback.
  91     std::vector<uintptr_t> MBBLocations;
  92
  93   public:
  94     MachOCodeEmitter(MachOWriter &mow) : MOW(mow), TM(MOW.TM) {
  95       is64Bit = TM.getTargetData()->getPointerSizeInBits() == 64;
  96       isLittleEndian = TM.getTargetData()->isLittleEndian();
  97     }
  98
  99     virtual void startFunction(MachineFunction &MF);
 100     virtual bool finishFunction(MachineFunction &MF);
 101
 102     virtual void addRelocation(const MachineRelocation &MR) {
 103       Relocations.push_back(MR);
 104     }
 105
 106     void emitConstantPool(MachineConstantPool *MCP);
 107     void emitJumpTables(MachineJumpTableInfo *MJTI);
 108
 109     virtual uintptr_t getConstantPoolEntryAddress(unsigned Index) const {
 110       assert(CPLocations.size() > Index && "CP not emitted!");
 111       return CPLocations[Index];
 112     }
 113     virtual uintptr_t getJumpTableEntryAddress(unsigned Index) const {
 114       assert(JTLocations.size() > Index && "JT not emitted!");
 115       return JTLocations[Index];
 116     }
 117
 118     virtual void StartMachineBasicBlock(MachineBasicBlock *MBB) {
 119       if (MBBLocations.size() <= (unsigned)MBB->getNumber())
 120         MBBLocations.resize((MBB->getNumber()+1)*2);
 121       MBBLocations[MBB->getNumber()] = getCurrentPCOffset();
 122     }
 123
 124     virtual uintptr_t getMachineBasicBlockAddress(MachineBasicBlock *MBB) const {
 125       assert(MBBLocations.size() > (unsigned)MBB->getNumber() &&
 126              MBBLocations[MBB->getNumber()] && "MBB not emitted!");
 127       return MBBLocations[MBB->getNumber()];
 128     }
 129
 130     virtual uintptr_t getLabelAddress(uint64_t Label) const {
 131       assert(0 && "get Label not implemented");
 132       abort();
 133       return 0;
 134     }
 135
 136     virtual void emitLabel(uint64_t LabelID) {
 137       assert(0 && "emit Label not implemented");
 138       abort();
 139     }
 140
 141
 142     virtual void setModuleInfo(llvm::MachineModuleInfo* MMI) { }
 143
 144     /// JIT SPECIFIC FUNCTIONS - DO NOT IMPLEMENT THESE HERE!
 145     virtual void startGVStub(const GlobalValue* F, unsigned StubSize,
 146                              unsigned Alignment = 1) {
 147       assert(0 && "JIT specific function called!");
 148       abort();
 149     }
 150     virtual void startGVStub(const GlobalValue* F, void *Buffer,
 151                              unsigned StubSize) {
 152       assert(0 && "JIT specific function called!");
 153       abort();
 154     }
 155     virtual void *finishGVStub(const GlobalValue* F) {
 156       assert(0 && "JIT specific function called!");
 157       abort();
 158       return 0;
 159     }
 160   };
 161 }
 162
 163 /// startFunction - This callback is invoked when a new machine function is
 164 /// about to be emitted.
 165 void MachOCodeEmitter::startFunction(MachineFunction &MF) {
 166   const TargetData *TD = TM.getTargetData();
 167   const Function *F = MF.getFunction();
 168
 169   // Align the output buffer to the appropriate alignment, power of 2.
 170   unsigned FnAlign = F->getAlignment();
 171   unsigned TDAlign = TD->getPrefTypeAlignment(F->getType());
 172   unsigned Align = Log2_32(std::max(FnAlign, TDAlign));
 173   assert(!(Align & (Align-1)) && "Alignment is not a power of two!");
 174
 175   // Get the Mach-O Section that this function belongs in.
 176   MachOWriter::MachOSection *MOS = MOW.getTextSection();
 177
 178   // FIXME: better memory management
 179   MOS->SectionData.reserve(4096);
 180   BufferBegin = &MOS->SectionData[0];
 181   BufferEnd = BufferBegin + MOS->SectionData.capacity();
 182
 183   // Upgrade the section alignment if required.
 184   if (MOS->align < Align) MOS->align = Align;
 185
 186   // Round the size up to the correct alignment for starting the new function.
 187   if ((MOS->size & ((1 << Align) - 1)) != 0) {
 188     MOS->size += (1 << Align);
 189     MOS->size &= ~((1 << Align) - 1);
 190   }
 191
 192   // FIXME: Using MOS->size directly here instead of calculating it from the
 193   // output buffer size (impossible because the code emitter deals only in raw
 194   // bytes) forces us to manually synchronize size and write padding zero bytes
 195   // to the output buffer for all non-text sections.  For text sections, we do
 196   // not synchonize the output buffer, and we just blow up if anyone tries to
 197   // write non-code to it.  An assert should probably be added to
 198   // AddSymbolToSection to prevent calling it on the text section.
 199   CurBufferPtr = BufferBegin + MOS->size;
 200
 201   // Clear per-function data structures.
 202   CPLocations.clear();
 203   CPSections.clear();
 204   JTLocations.clear();
 205   MBBLocations.clear();
 206 }
 207
 208 /// finishFunction - This callback is invoked after the function is completely
 209 /// finished.
 210 bool MachOCodeEmitter::finishFunction(MachineFunction &MF) {
 211   // Get the Mach-O Section that this function belongs in.
 212   MachOWriter::MachOSection *MOS = MOW.getTextSection();
 213
 214   // Get a symbol for the function to add to the symbol table
 215   // FIXME: it seems like we should call something like AddSymbolToSection
 216   // in startFunction rather than changing the section size and symbol n_value
 217   // here.
 218   const GlobalValue *FuncV = MF.getFunction();
 219   MachOSym FnSym(FuncV, MOW.Mang->getValueName(FuncV), MOS->Index, TM);
 220   FnSym.n_value = MOS->size;
 221   MOS->size = CurBufferPtr - BufferBegin;
 222
 223   // Emit constant pool to appropriate section(s)
 224   emitConstantPool(MF.getConstantPool());
 225
 226   // Emit jump tables to appropriate section
 227   emitJumpTables(MF.getJumpTableInfo());
 228
 229   // If we have emitted any relocations to function-specific objects such as
 230   // basic blocks, constant pools entries, or jump tables, record their
 231   // addresses now so that we can rewrite them with the correct addresses
 232   // later.
 233   for (unsigned i = 0, e = Relocations.size(); i != e; ++i) {
 234     MachineRelocation &MR = Relocations[i];
 235     intptr_t Addr;
 236
 237     if (MR.isBasicBlock()) {
 238       Addr = getMachineBasicBlockAddress(MR.getBasicBlock());
 239       MR.setConstantVal(MOS->Index);
 240       MR.setResultPointer((void*)Addr);
 241     } else if (MR.isJumpTableIndex()) {
 242       Addr = getJumpTableEntryAddress(MR.getJumpTableIndex());
 243       MR.setConstantVal(MOW.getJumpTableSection()->Index);
 244       MR.setResultPointer((void*)Addr);
 245     } else if (MR.isConstantPoolIndex()) {
 246       Addr = getConstantPoolEntryAddress(MR.getConstantPoolIndex());
 247       MR.setConstantVal(CPSections[MR.getConstantPoolIndex()]);
 248       MR.setResultPointer((void*)Addr);
 249     } else if (MR.isGlobalValue()) {
 250       // FIXME: This should be a set or something that uniques
 251       MOW.PendingGlobals.push_back(MR.getGlobalValue());
 252     } else {
 253       assert(0 && "Unhandled relocation type");
 254     }
 255     MOS->Relocations.push_back(MR);
 256   }
 257   Relocations.clear();
 258
 259   // Finally, add it to the symtab.
 260   MOW.SymbolTable.push_back(FnSym);
 261   return false;
 262 }
 263
 264 /// emitConstantPool - For each constant pool entry, figure out which section
 265 /// the constant should live in, allocate space for it, and emit it to the
 266 /// Section data buffer.
 267 void MachOCodeEmitter::emitConstantPool(MachineConstantPool *MCP) {
 268   const std::vector<MachineConstantPoolEntry> &CP = MCP->getConstants();
 269   if (CP.empty()) return;
 270
 271   // FIXME: handle PIC codegen
 272   assert(TM.getRelocationModel() != Reloc::PIC_ &&
 273          "PIC codegen not yet handled for mach-o jump tables!");
 274
 275   // Although there is no strict necessity that I am aware of, we will do what
 276   // gcc for OS X does and put each constant pool entry in a section of constant
 277   // objects of a certain size.  That means that float constants go in the
 278   // literal4 section, and double objects go in literal8, etc.
 279   //
 280   // FIXME: revisit this decision if we ever do the "stick everything into one
 281   // "giant object for PIC" optimization.
 282   for (unsigned i = 0, e = CP.size(); i != e; ++i) {
 283     const Type *Ty = CP[i].getType();
 284     unsigned Size = TM.getTargetData()->getTypeAllocSize(Ty);
 285
 286     MachOWriter::MachOSection *Sec = MOW.getConstSection(CP[i].Val.ConstVal);
 287     OutputBuffer SecDataOut(Sec->SectionData, is64Bit, isLittleEndian);
 288
 289     CPLocations.push_back(Sec->SectionData.size());
 290     CPSections.push_back(Sec->Index);
 291
 292     // FIXME: remove when we have unified size + output buffer
 293     Sec->size += Size;
 294
 295     // Allocate space in the section for the global.
 296     // FIXME: need alignment?
 297     // FIXME: share between here and AddSymbolToSection?
 298     for (unsigned j = 0; j < Size; ++j)
 299       SecDataOut.outbyte(0);
 300
 301     MOW.InitMem(CP[i].Val.ConstVal, &Sec->SectionData[0], CPLocations[i],
 302                 TM.getTargetData(), Sec->Relocations);
 303   }
 304 }
 305
 306 /// emitJumpTables - Emit all the jump tables for a given jump table info
 307 /// record to the appropriate section.
 308 void MachOCodeEmitter::emitJumpTables(MachineJumpTableInfo *MJTI) {
 309   const std::vector<MachineJumpTableEntry> &JT = MJTI->getJumpTables();
 310   if (JT.empty()) return;
 311
 312   // FIXME: handle PIC codegen
 313   assert(TM.getRelocationModel() != Reloc::PIC_ &&
 314          "PIC codegen not yet handled for mach-o jump tables!");
 315
 316   MachOWriter::MachOSection *Sec = MOW.getJumpTableSection();
 317   unsigned TextSecIndex = MOW.getTextSection()->Index;
 318   OutputBuffer SecDataOut(Sec->SectionData, is64Bit, isLittleEndian);
 319
 320   for (unsigned i = 0, e = JT.size(); i != e; ++i) {
 321     // For each jump table, record its offset from the start of the section,
 322     // reserve space for the relocations to the MBBs, and add the relocations.
 323     const std::vector<MachineBasicBlock*> &MBBs = JT[i].MBBs;
 324     JTLocations.push_back(Sec->SectionData.size());
 325     for (unsigned mi = 0, me = MBBs.size(); mi != me; ++mi) {
 326       MachineRelocation MR(MOW.GetJTRelocation(Sec->SectionData.size(),
 327                                                MBBs[mi]));
 328       MR.setResultPointer((void *)JTLocations[i]);
 329       MR.setConstantVal(TextSecIndex);
 330       Sec->Relocations.push_back(MR);
 331       SecDataOut.outaddr(0);
 332     }
 333   }
 334   // FIXME: remove when we have unified size + output buffer
 335   Sec->size = Sec->SectionData.size();
 336 }
 337
 338 //===----------------------------------------------------------------------===//
 339 //                          MachOWriter Implementation
 340 //===----------------------------------------------------------------------===//
 341
 342 char MachOWriter::ID = 0;
 343 MachOWriter::MachOWriter(raw_ostream &o, TargetMachine &tm)
 344   : MachineFunctionPass(&ID), O(o), TM(tm) {
 345   is64Bit = TM.getTargetData()->getPointerSizeInBits() == 64;
 346   isLittleEndian = TM.getTargetData()->isLittleEndian();
 347
 348   // Create the machine code emitter object for this target.
 349   MCE = new MachOCodeEmitter(*this);
 350 }
 351
 352 MachOWriter::~MachOWriter() {
 353   delete MCE;
 354 }
 355
 356 void MachOWriter::AddSymbolToSection(MachOSection *Sec, GlobalVariable *GV) {
 357   const Type *Ty = GV->getType()->getElementType();
 358   unsigned Size = TM.getTargetData()->getTypeAllocSize(Ty);
 359   unsigned Align = TM.getTargetData()->getPreferredAlignment(GV);
 360
 361   // Reserve space in the .bss section for this symbol while maintaining the
 362   // desired section alignment, which must be at least as much as required by
 363   // this symbol.
 364   OutputBuffer SecDataOut(Sec->SectionData, is64Bit, isLittleEndian);
 365
 366   if (Align) {
 367     uint64_t OrigSize = Sec->size;
 368     Align = Log2_32(Align);
 369     Sec->align = std::max(unsigned(Sec->align), Align);
 370     Sec->size = (Sec->size + Align - 1) & ~(Align-1);
 371
 372     // Add alignment padding to buffer as well.
 373     // FIXME: remove when we have unified size + output buffer
 374     unsigned AlignedSize = Sec->size - OrigSize;
 375     for (unsigned i = 0; i < AlignedSize; ++i)
 376       SecDataOut.outbyte(0);
 377   }
 378   // Globals without external linkage apparently do not go in the symbol table.
 379   if (!GV->hasLocalLinkage()) {
 380     MachOSym Sym(GV, Mang->getValueName(GV), Sec->Index, TM);
 381     Sym.n_value = Sec->size;
 382     SymbolTable.push_back(Sym);
 383   }
 384
 385   // Record the offset of the symbol, and then allocate space for it.
 386   // FIXME: remove when we have unified size + output buffer
 387   Sec->size += Size;
 388
 389   // Now that we know what section the GlovalVariable is going to be emitted
 390   // into, update our mappings.
 391   // FIXME: We may also need to update this when outputting non-GlobalVariable
 392   // GlobalValues such as functions.
 393   GVSection[GV] = Sec;
 394   GVOffset[GV] = Sec->SectionData.size();
 395
 396   // Allocate space in the section for the global.
 397   for (unsigned i = 0; i < Size; ++i)
 398     SecDataOut.outbyte(0);
 399 }
 400
 401 void MachOWriter::EmitGlobal(GlobalVariable *GV) {
 402   const Type *Ty = GV->getType()->getElementType();
 403   unsigned Size = TM.getTargetData()->getTypeAllocSize(Ty);
 404   bool NoInit = !GV->hasInitializer();
 405
 406   // If this global has a zero initializer, it is part of the .bss or common
 407   // section.
 408   if (NoInit || GV->getInitializer()->isNullValue()) {
 409     // If this global is part of the common block, add it now.  Variables are
 410     // part of the common block if they are zero initialized and allowed to be
 411     // merged with other symbols.
 412     if (NoInit || GV->hasLinkOnceLinkage() || GV->hasWeakLinkage() ||
 413         GV->hasCommonLinkage()) {
 414       MachOSym ExtOrCommonSym(GV, Mang->getValueName(GV), MachOSym::NO_SECT,TM);
 415       // For undefined (N_UNDF) external (N_EXT) types, n_value is the size in
 416       // bytes of the symbol.
 417       ExtOrCommonSym.n_value = Size;
 418       SymbolTable.push_back(ExtOrCommonSym);
 419       // Remember that we've seen this symbol
 420       GVOffset[GV] = Size;
 421       return;
 422     }
 423     // Otherwise, this symbol is part of the .bss section.
 424     MachOSection *BSS = getBSSSection();
 425     AddSymbolToSection(BSS, GV);
 426     return;
 427   }
 428
 429   // Scalar read-only data goes in a literal section if the scalar is 4, 8, or
 430   // 16 bytes, or a cstring.  Other read only data goes into a regular const
 431   // section.  Read-write data goes in the data section.
 432   MachOSection *Sec = GV->isConstant() ? getConstSection(GV->getInitializer()) :
 433                                          getDataSection();
 434   AddSymbolToSection(Sec, GV);
 435   InitMem(GV->getInitializer(), &Sec->SectionData[0], GVOffset[GV],
 436           TM.getTargetData(), Sec->Relocations);
 437 }
 438
 439
 440 bool MachOWriter::runOnMachineFunction(MachineFunction &MF) {
 441   // Nothing to do here, this is all done through the MCE object.
 442   return false;
 443 }
 444
 445 bool MachOWriter::doInitialization(Module &M) {
 446   // Set the magic value, now that we know the pointer size and endianness
 447   Header.setMagic(isLittleEndian, is64Bit);
 448
 449   // Set the file type
 450   // FIXME: this only works for object files, we do not support the creation
 451   //        of dynamic libraries or executables at this time.
 452   Header.filetype = MachOHeader::MH_OBJECT;
 453
 454   Mang = new Mangler(M);
 455   return false;
 456 }
 457
 458 /// doFinalization - Now that the module has been completely processed, emit
 459 /// the Mach-O file to 'O'.
 460 bool MachOWriter::doFinalization(Module &M) {
 461   // FIXME: we don't handle debug info yet, we should probably do that.
 462
 463   // Okay, the.text section has been completed, build the .data, .bss, and
 464   // "common" sections next.
 465   for (Module::global_iterator I = M.global_begin(), E = M.global_end();
 466        I != E; ++I)
 467     EmitGlobal(I);
 468
 469   // Emit the header and load commands.
 470   EmitHeaderAndLoadCommands();
 471
 472   // Emit the various sections and their relocation info.
 473   EmitSections();
 474
 475   // Write the symbol table and the string table to the end of the file.
 476   O.write((char*)&SymT[0], SymT.size());
 477   O.write((char*)&StrT[0], StrT.size());
 478
 479   // We are done with the abstract symbols.
 480   SectionList.clear();
 481   SymbolTable.clear();
 482   DynamicSymbolTable.clear();
 483
 484   // Release the name mangler object.
 485   delete Mang; Mang = 0;
 486   return false;
 487 }
 488
 489 void MachOWriter::EmitHeaderAndLoadCommands() {
 490   // Step #0: Fill in the segment load command size, since we need it to figure
 491   //          out the rest of the header fields
 492   MachOSegment SEG("", is64Bit);
 493   SEG.nsects  = SectionList.size();
 494   SEG.cmdsize = SEG.cmdSize(is64Bit) +
 495                 SEG.nsects * SectionList[0]->cmdSize(is64Bit);
 496
 497   // Step #1: calculate the number of load commands.  We always have at least
 498   //          one, for the LC_SEGMENT load command, plus two for the normal
 499   //          and dynamic symbol tables, if there are any symbols.
 500   Header.ncmds = SymbolTable.empty() ? 1 : 3;
 501
 502   // Step #2: calculate the size of the load commands
 503   Header.sizeofcmds = SEG.cmdsize;
 504   if (!SymbolTable.empty())
 505     Header.sizeofcmds += SymTab.cmdsize + DySymTab.cmdsize;
 506
 507   // Step #3: write the header to the file
 508   // Local alias to shortenify coming code.
 509   DataBuffer &FH = Header.HeaderData;
 510   OutputBuffer FHOut(FH, is64Bit, isLittleEndian);
 511
 512   FHOut.outword(Header.magic);
 513   FHOut.outword(TM.getMachOWriterInfo()->getCPUType());
 514   FHOut.outword(TM.getMachOWriterInfo()->getCPUSubType());
 515   FHOut.outword(Header.filetype);
 516   FHOut.outword(Header.ncmds);
 517   FHOut.outword(Header.sizeofcmds);
 518   FHOut.outword(Header.flags);
 519   if (is64Bit)
 520     FHOut.outword(Header.reserved);
 521
 522   // Step #4: Finish filling in the segment load command and write it out
 523   for (std::vector<MachOSection*>::iterator I = SectionList.begin(),
 524          E = SectionList.end(); I != E; ++I)
 525     SEG.filesize += (*I)->size;
 526
 527   SEG.vmsize = SEG.filesize;
 528   SEG.fileoff = Header.cmdSize(is64Bit) + Header.sizeofcmds;
 529
 530   FHOut.outword(SEG.cmd);
 531   FHOut.outword(SEG.cmdsize);
 532   FHOut.outstring(SEG.segname, 16);
 533   FHOut.outaddr(SEG.vmaddr);
 534   FHOut.outaddr(SEG.vmsize);
 535   FHOut.outaddr(SEG.fileoff);
 536   FHOut.outaddr(SEG.filesize);
 537   FHOut.outword(SEG.maxprot);
 538   FHOut.outword(SEG.initprot);
 539   FHOut.outword(SEG.nsects);
 540   FHOut.outword(SEG.flags);
 541
 542   // Step #5: Finish filling in the fields of the MachOSections
 543   uint64_t currentAddr = 0;
 544   for (std::vector<MachOSection*>::iterator I = SectionList.begin(),
 545          E = SectionList.end(); I != E; ++I) {
 546     MachOSection *MOS = *I;
 547     MOS->addr = currentAddr;
 548     MOS->offset = currentAddr + SEG.fileoff;
 549
 550     // FIXME: do we need to do something with alignment here?
 551     currentAddr += MOS->size;
 552   }
 553
 554   // Step #6: Emit the symbol table to temporary buffers, so that we know the
 555   // size of the string table when we write the next load command.  This also
 556   // sorts and assigns indices to each of the symbols, which is necessary for
 557   // emitting relocations to externally-defined objects.
 558   BufferSymbolAndStringTable();
 559
 560   // Step #7: Calculate the number of relocations for each section and write out
 561   // the section commands for each section
 562   currentAddr += SEG.fileoff;
 563   for (std::vector<MachOSection*>::iterator I = SectionList.begin(),
 564          E = SectionList.end(); I != E; ++I) {
 565     MachOSection *MOS = *I;
 566     // Convert the relocations to target-specific relocations, and fill in the
 567     // relocation offset for this section.
 568     CalculateRelocations(*MOS);
 569     MOS->reloff = MOS->nreloc ? currentAddr : 0;
 570     currentAddr += MOS->nreloc * 8;
 571
 572     // write the finalized section command to the output buffer
 573     FHOut.outstring(MOS->sectname, 16);
 574     FHOut.outstring(MOS->segname, 16);
 575     FHOut.outaddr(MOS->addr);
 576     FHOut.outaddr(MOS->size);
 577     FHOut.outword(MOS->offset);
 578     FHOut.outword(MOS->align);
 579     FHOut.outword(MOS->reloff);
 580     FHOut.outword(MOS->nreloc);
 581     FHOut.outword(MOS->flags);
 582     FHOut.outword(MOS->reserved1);
 583     FHOut.outword(MOS->reserved2);
 584     if (is64Bit)
 585       FHOut.outword(MOS->reserved3);
 586   }
 587
 588   // Step #8: Emit LC_SYMTAB/LC_DYSYMTAB load commands
 589   SymTab.symoff  = currentAddr;
 590   SymTab.nsyms   = SymbolTable.size();
 591   SymTab.stroff  = SymTab.symoff + SymT.size();
 592   SymTab.strsize = StrT.size();
 593   FHOut.outword(SymTab.cmd);
 594   FHOut.outword(SymTab.cmdsize);
 595   FHOut.outword(SymTab.symoff);
 596   FHOut.outword(SymTab.nsyms);
 597   FHOut.outword(SymTab.stroff);
 598   FHOut.outword(SymTab.strsize);
 599
 600   // FIXME: set DySymTab fields appropriately
 601   // We should probably just update these in BufferSymbolAndStringTable since
 602   // thats where we're partitioning up the different kinds of symbols.
 603   FHOut.outword(DySymTab.cmd);
 604   FHOut.outword(DySymTab.cmdsize);
 605   FHOut.outword(DySymTab.ilocalsym);
 606   FHOut.outword(DySymTab.nlocalsym);
 607   FHOut.outword(DySymTab.iextdefsym);
 608   FHOut.outword(DySymTab.nextdefsym);
 609   FHOut.outword(DySymTab.iundefsym);
 610   FHOut.outword(DySymTab.nundefsym);
 611   FHOut.outword(DySymTab.tocoff);
 612   FHOut.outword(DySymTab.ntoc);
 613   FHOut.outword(DySymTab.modtaboff);
 614   FHOut.outword(DySymTab.nmodtab);
 615   FHOut.outword(DySymTab.extrefsymoff);
 616   FHOut.outword(DySymTab.nextrefsyms);
 617   FHOut.outword(DySymTab.indirectsymoff);
 618   FHOut.outword(DySymTab.nindirectsyms);
 619   FHOut.outword(DySymTab.extreloff);
 620   FHOut.outword(DySymTab.nextrel);
 621   FHOut.outword(DySymTab.locreloff);
 622   FHOut.outword(DySymTab.nlocrel);
 623
 624   O.write((char*)&FH[0], FH.size());
 625 }
 626
 627 /// EmitSections - Now that we have constructed the file header and load
 628 /// commands, emit the data for each section to the file.
 629 void MachOWriter::EmitSections() {
 630   for (std::vector<MachOSection*>::iterator I = SectionList.begin(),
 631          E = SectionList.end(); I != E; ++I)
 632     // Emit the contents of each section
 633     O.write((char*)&(*I)->SectionData[0], (*I)->size);
 634   for (std::vector<MachOSection*>::iterator I = SectionList.begin(),
 635          E = SectionList.end(); I != E; ++I)
 636     // Emit the relocation entry data for each section.
 637     O.write((char*)&(*I)->RelocBuffer[0], (*I)->RelocBuffer.size());
 638 }
 639
 640 /// PartitionByLocal - Simple boolean predicate that returns true if Sym is
 641 /// a local symbol rather than an external symbol.
 642 bool MachOWriter::PartitionByLocal(const MachOSym &Sym) {
 643   return (Sym.n_type & (MachOSym::N_EXT | MachOSym::N_PEXT)) == 0;
 644 }
 645
 646 /// PartitionByDefined - Simple boolean predicate that returns true if Sym is
 647 /// defined in this module.
 648 bool MachOWriter::PartitionByDefined(const MachOSym &Sym) {
 649   // FIXME: Do N_ABS or N_INDR count as defined?
 650   return (Sym.n_type & MachOSym::N_SECT) == MachOSym::N_SECT;
 651 }
 652
 653 /// BufferSymbolAndStringTable - Sort the symbols we encountered and assign them
 654 /// each a string table index so that they appear in the correct order in the
 655 /// output file.
 656 void MachOWriter::BufferSymbolAndStringTable() {
 657   // The order of the symbol table is:
 658   // 1. local symbols
 659   // 2. defined external symbols (sorted by name)
 660   // 3. undefined external symbols (sorted by name)
 661
 662   // Before sorting the symbols, check the PendingGlobals for any undefined
 663   // globals that need to be put in the symbol table.
 664   for (std::vector<GlobalValue*>::iterator I = PendingGlobals.begin(),
 665          E = PendingGlobals.end(); I != E; ++I) {
 666     if (GVOffset[*I] == 0 && GVSection[*I] == 0) {
 667       MachOSym UndfSym(*I, Mang->getValueName(*I), MachOSym::NO_SECT, TM);
 668       SymbolTable.push_back(UndfSym);
 669       GVOffset[*I] = -1;
 670     }
 671   }
 672
 673   // Sort the symbols by name, so that when we partition the symbols by scope
 674   // of definition, we won't have to sort by name within each partition.
 675   std::sort(SymbolTable.begin(), SymbolTable.end(), MachOSymCmp());
 676
 677   // Parition the symbol table entries so that all local symbols come before
 678   // all symbols with external linkage. { 1 | 2 3 }
 679   std::partition(SymbolTable.begin(), SymbolTable.end(), PartitionByLocal);
 680
 681   // Advance iterator to beginning of external symbols and partition so that
 682   // all external symbols defined in this module come before all external
 683   // symbols defined elsewhere. { 1 | 2 | 3 }
 684   for (std::vector<MachOSym>::iterator I = SymbolTable.begin(),
 685          E = SymbolTable.end(); I != E; ++I) {
 686     if (!PartitionByLocal(*I)) {
 687       std::partition(I, E, PartitionByDefined);
 688       break;
 689     }
 690   }
 691
 692   // Calculate the starting index for each of the local, extern defined, and
 693   // undefined symbols, as well as the number of each to put in the LC_DYSYMTAB
 694   // load command.
 695   for (std::vector<MachOSym>::iterator I = SymbolTable.begin(),
 696          E = SymbolTable.end(); I != E; ++I) {
 697     if (PartitionByLocal(*I)) {
 698       ++DySymTab.nlocalsym;
 699       ++DySymTab.iextdefsym;
 700       ++DySymTab.iundefsym;
 701     } else if (PartitionByDefined(*I)) {
 702       ++DySymTab.nextdefsym;
 703       ++DySymTab.iundefsym;
 704     } else {
 705       ++DySymTab.nundefsym;
 706     }
 707   }
 708
 709   // Write out a leading zero byte when emitting string table, for n_strx == 0
 710   // which means an empty string.
 711   OutputBuffer StrTOut(StrT, is64Bit, isLittleEndian);
 712   StrTOut.outbyte(0);
 713
 714   // The order of the string table is:
 715   // 1. strings for external symbols
 716   // 2. strings for local symbols
 717   // Since this is the opposite order from the symbol table, which we have just
 718   // sorted, we can walk the symbol table backwards to output the string table.
 719   for (std::vector<MachOSym>::reverse_iterator I = SymbolTable.rbegin(),
 720         E = SymbolTable.rend(); I != E; ++I) {
 721     if (I->GVName == "") {
 722       I->n_strx = 0;
 723     } else {
 724       I->n_strx = StrT.size();
 725       StrTOut.outstring(I->GVName, I->GVName.length()+1);
 726     }
 727   }
 728
 729   OutputBuffer SymTOut(SymT, is64Bit, isLittleEndian);
 730
 731   unsigned index = 0;
 732   for (std::vector<MachOSym>::iterator I = SymbolTable.begin(),
 733          E = SymbolTable.end(); I != E; ++I, ++index) {
 734     // Add the section base address to the section offset in the n_value field
 735     // to calculate the full address.
 736     // FIXME: handle symbols where the n_value field is not the address
 737     GlobalValue *GV = const_cast<GlobalValue*>(I->GV);
 738     if (GV && GVSection[GV])
 739       I->n_value += GVSection[GV]->addr;
 740     if (GV && (GVOffset[GV] == -1))
 741       GVOffset[GV] = index;
 742
 743     // Emit nlist to buffer
 744     SymTOut.outword(I->n_strx);
 745     SymTOut.outbyte(I->n_type);
 746     SymTOut.outbyte(I->n_sect);
 747     SymTOut.outhalf(I->n_desc);
 748     SymTOut.outaddr(I->n_value);
 749   }
 750 }
 751
 752 /// CalculateRelocations - For each MachineRelocation in the current section,
 753 /// calculate the index of the section containing the object to be relocated,
 754 /// and the offset into that section.  From this information, create the
 755 /// appropriate target-specific MachORelocation type and add buffer it to be
 756 /// written out after we are finished writing out sections.
 757 void MachOWriter::CalculateRelocations(MachOSection &MOS) {
 758   for (unsigned i = 0, e = MOS.Relocations.size(); i != e; ++i) {
 759     MachineRelocation &MR = MOS.Relocations[i];
 760     unsigned TargetSection = MR.getConstantVal();
 761     unsigned TargetAddr = 0;
 762     unsigned TargetIndex = 0;
 763
 764     // This is a scattered relocation entry if it points to a global value with
 765     // a non-zero offset.
 766     bool Scattered = false;
 767     bool Extern = false;
 768
 769     // Since we may not have seen the GlobalValue we were interested in yet at
 770     // the time we emitted the relocation for it, fix it up now so that it
 771     // points to the offset into the correct section.
 772     if (MR.isGlobalValue()) {
 773       GlobalValue *GV = MR.getGlobalValue();
 774       MachOSection *MOSPtr = GVSection[GV];
 775       intptr_t Offset = GVOffset[GV];
 776
 777       // If we have never seen the global before, it must be to a symbol
 778       // defined in another module (N_UNDF).
 779       if (!MOSPtr) {
 780         // FIXME: need to append stub suffix
 781         Extern = true;
 782         TargetAddr = 0;
 783         TargetIndex = GVOffset[GV];
 784       } else {
 785         Scattered = TargetSection != 0;
 786         TargetSection = MOSPtr->Index;
 787       }
 788       MR.setResultPointer((void*)Offset);
 789     }
 790
 791     // If the symbol is locally defined, pass in the address of the section and
 792     // the section index to the code which will generate the target relocation.
 793     if (!Extern) {
 794         MachOSection &To = *SectionList[TargetSection - 1];
 795         TargetAddr = To.addr;
 796         TargetIndex = To.Index;
 797     }
 798
 799     OutputBuffer RelocOut(MOS.RelocBuffer, is64Bit, isLittleEndian);
 800     OutputBuffer SecOut(MOS.SectionData, is64Bit, isLittleEndian);
 801
 802     MOS.nreloc += GetTargetRelocation(MR, MOS.Index, TargetAddr, TargetIndex,
 803                                       RelocOut, SecOut, Scattered, Extern);
 804   }
 805 }
 806
 807 // InitMem - Write the value of a Constant to the specified memory location,
 808 // converting it into bytes and relocations.
 809 void MachOWriter::InitMem(const Constant *C, void *Addr, intptr_t Offset,
 810                           const TargetData *TD,
 811                           std::vector<MachineRelocation> &MRs) {
 812   typedef std::pair<const Constant*, intptr_t> CPair;
 813   std::vector<CPair> WorkList;
 814
 815   WorkList.push_back(CPair(C,(intptr_t)Addr + Offset));
 816
 817   intptr_t ScatteredOffset = 0;
 818
 819   while (!WorkList.empty()) {
 820     const Constant *PC = WorkList.back().first;
 821     intptr_t PA = WorkList.back().second;
 822     WorkList.pop_back();
 823
 824     if (isa<UndefValue>(PC)) {
 825       continue;
 826     } else if (const ConstantVector *CP = dyn_cast<ConstantVector>(PC)) {
 827       unsigned ElementSize =
 828         TD->getTypeAllocSize(CP->getType()->getElementType());
 829       for (unsigned i = 0, e = CP->getNumOperands(); i != e; ++i)
 830         WorkList.push_back(CPair(CP->getOperand(i), PA+i*ElementSize));
 831     } else if (const ConstantExpr *CE = dyn_cast<ConstantExpr>(PC)) {
 832       //
 833       // FIXME: Handle ConstantExpression.  See EE::getConstantValue()
 834       //
 835       switch (CE->getOpcode()) {
 836       case Instruction::GetElementPtr: {
 837         SmallVector<Value*, 8> Indices(CE->op_begin()+1, CE->op_end());
 838         ScatteredOffset = TD->getIndexedOffset(CE->getOperand(0)->getType(),
 839                                                &Indices[0], Indices.size());
 840         WorkList.push_back(CPair(CE->getOperand(0), PA));
 841         break;
 842       }
 843       case Instruction::Add:
 844       default:
 845         cerr << "ConstantExpr not handled as global var init: " << *CE << "\n";
 846         abort();
 847         break;
 848       }
 849     } else if (PC->getType()->isSingleValueType()) {
 850       unsigned char *ptr = (unsigned char *)PA;
 851       switch (PC->getType()->getTypeID()) {
 852       case Type::IntegerTyID: {
 853         unsigned NumBits = cast<IntegerType>(PC->getType())->getBitWidth();
 854         uint64_t val = cast<ConstantInt>(PC)->getZExtValue();
 855         if (NumBits <= 8)
 856           ptr[0] = val;
 857         else if (NumBits <= 16) {
 858           if (TD->isBigEndian())
 859             val = ByteSwap_16(val);
 860           ptr[0] = val;
 861           ptr[1] = val >> 8;
 862         } else if (NumBits <= 32) {
 863           if (TD->isBigEndian())
 864             val = ByteSwap_32(val);
 865           ptr[0] = val;
 866           ptr[1] = val >> 8;
 867           ptr[2] = val >> 16;
 868           ptr[3] = val >> 24;
 869         } else if (NumBits <= 64) {
 870           if (TD->isBigEndian())
 871             val = ByteSwap_64(val);
 872           ptr[0] = val;
 873           ptr[1] = val >> 8;
 874           ptr[2] = val >> 16;
 875           ptr[3] = val >> 24;
 876           ptr[4] = val >> 32;
 877           ptr[5] = val >> 40;
 878           ptr[6] = val >> 48;
 879           ptr[7] = val >> 56;
 880         } else {
 881           assert(0 && "Not implemented: bit widths > 64");
 882         }
 883         break;
 884       }
 885       case Type::FloatTyID: {
 886         uint32_t val = cast<ConstantFP>(PC)->getValueAPF().bitcastToAPInt().
 887                         getZExtValue();
 888         if (TD->isBigEndian())
 889           val = ByteSwap_32(val);
 890         ptr[0] = val;
 891         ptr[1] = val >> 8;
 892         ptr[2] = val >> 16;
 893         ptr[3] = val >> 24;
 894         break;
 895       }
 896       case Type::DoubleTyID: {
 897         uint64_t val = cast<ConstantFP>(PC)->getValueAPF().bitcastToAPInt().
 898                          getZExtValue();
 899         if (TD->isBigEndian())
 900           val = ByteSwap_64(val);
 901         ptr[0] = val;
 902         ptr[1] = val >> 8;
 903         ptr[2] = val >> 16;
 904         ptr[3] = val >> 24;
 905         ptr[4] = val >> 32;
 906         ptr[5] = val >> 40;
 907         ptr[6] = val >> 48;
 908         ptr[7] = val >> 56;
 909         break;
 910       }
 911       case Type::PointerTyID:
 912         if (isa<ConstantPointerNull>(PC))
 913           memset(ptr, 0, TD->getPointerSize());
 914         else if (const GlobalValue* GV = dyn_cast<GlobalValue>(PC)) {
 915           // FIXME: what about function stubs?
 916           MRs.push_back(MachineRelocation::getGV(PA-(intptr_t)Addr,
 917                                                  MachineRelocation::VANILLA,
 918                                                  const_cast<GlobalValue*>(GV),
 919                                                  ScatteredOffset));
 920           ScatteredOffset = 0;
 921         } else
 922           assert(0 && "Unknown constant pointer type!");
 923         break;
 924       default:
 925         cerr << "ERROR: Constant unimp for type: " << *PC->getType() << "\n";
 926         abort();
 927       }
 928     } else if (isa<ConstantAggregateZero>(PC)) {
 929       memset((void*)PA, 0, (size_t)TD->getTypeAllocSize(PC->getType()));
 930     } else if (const ConstantArray *CPA = dyn_cast<ConstantArray>(PC)) {
 931       unsigned ElementSize =
 932         TD->getTypeAllocSize(CPA->getType()->getElementType());
 933       for (unsigned i = 0, e = CPA->getNumOperands(); i != e; ++i)
 934         WorkList.push_back(CPair(CPA->getOperand(i), PA+i*ElementSize));
 935     } else if (const ConstantStruct *CPS = dyn_cast<ConstantStruct>(PC)) {
 936       const StructLayout *SL =
 937         TD->getStructLayout(cast<StructType>(CPS->getType()));
 938       for (unsigned i = 0, e = CPS->getNumOperands(); i != e; ++i)
 939         WorkList.push_back(CPair(CPS->getOperand(i),
 940                                  PA+SL->getElementOffset(i)));
 941     } else {
 942       cerr << "Bad Type: " << *PC->getType() << "\n";
 943       assert(0 && "Unknown constant type to initialize memory with!");
 944     }
 945   }
 946 }
 947
 948 MachOSym::MachOSym(const GlobalValue *gv, std::string name, uint8_t sect,
 949                    TargetMachine &TM) :
 950   GV(gv), n_strx(0), n_type(sect == NO_SECT ? N_UNDF : N_SECT), n_sect(sect),
 951   n_desc(0), n_value(0) {
 952
 953   const TargetAsmInfo *TAI = TM.getTargetAsmInfo();
 954
 955   switch (GV->getLinkage()) {
 956   default:
 957     assert(0 && "Unexpected linkage type!");
 958     break;
 959   case GlobalValue::WeakAnyLinkage:
 960   case GlobalValue::WeakODRLinkage:
 961   case GlobalValue::LinkOnceAnyLinkage:
 962   case GlobalValue::LinkOnceODRLinkage:
 963   case GlobalValue::CommonLinkage:
 964     assert(!isa<Function>(gv) && "Unexpected linkage type for Function!");
 965   case GlobalValue::ExternalLinkage:
 966     GVName = TAI->getGlobalPrefix() + name;
 967     n_type |= GV->hasHiddenVisibility() ? N_PEXT : N_EXT;
 968     break;
 969   case GlobalValue::PrivateLinkage:
 970     GVName = TAI->getPrivateGlobalPrefix() + name;
 971     break;
 972   case GlobalValue::InternalLinkage:
 973     GVName = TAI->getGlobalPrefix() + name;
 974     break;
 975   }
 976 }