MCMachOStreamer.cpp revision a86de1050889b61df2cd42cdc2acc5f9796b7f0c 
1//===- lib/MC/MCMachOStreamer.cpp - Mach-O Object Output ------------===//
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#include "llvm/MC/MCStreamer.h"
11
12#include "llvm/MC/MCAssembler.h"
13#include "llvm/MC/MCContext.h"
14#include "llvm/MC/MCCodeEmitter.h"
15#include "llvm/MC/MCExpr.h"
16#include "llvm/MC/MCInst.h"
17#include "llvm/MC/MCObjectStreamer.h"
18#include "llvm/MC/MCSection.h"
19#include "llvm/MC/MCSymbol.h"
20#include "llvm/MC/MCMachOSymbolFlags.h"
21#include "llvm/Support/ErrorHandling.h"
22#include "llvm/Support/raw_ostream.h"
23#include "llvm/Target/TargetAsmBackend.h"
24
25using namespace llvm;
26
27namespace {
28
29class MCMachOStreamer : public MCObjectStreamer {
30private:
31  MCFragment *getCurrentFragment() const {
32    assert(getCurrentSectionData() && "No current section!");
33
34    if (!getCurrentSectionData()->empty())
35      return &getCurrentSectionData()->getFragmentList().back();
36
37    return 0;
38  }
39
40  /// Get a data fragment to write into, creating a new one if the current
41  /// fragment is not a data fragment.
42  MCDataFragment *getOrCreateDataFragment() const {
43    MCDataFragment *F = dyn_cast_or_null<MCDataFragment>(getCurrentFragment());
44    if (!F)
45      F = new MCDataFragment(getCurrentSectionData());
46    return F;
47  }
48
49  void EmitInstToFragment(const MCInst &Inst);
50  void EmitInstToData(const MCInst &Inst);
51
52public:
53  MCMachOStreamer(MCContext &Context, TargetAsmBackend &TAB,
54                  raw_ostream &OS, MCCodeEmitter *Emitter)
55    : MCObjectStreamer(Context, TAB, OS, Emitter) {}
56
57  const MCExpr *AddValueSymbols(const MCExpr *Value) {
58    switch (Value->getKind()) {
59    case MCExpr::Target: assert(0 && "Can't handle target exprs yet!");
60    case MCExpr::Constant:
61      break;
62
63    case MCExpr::Binary: {
64      const MCBinaryExpr *BE = cast<MCBinaryExpr>(Value);
65      AddValueSymbols(BE->getLHS());
66      AddValueSymbols(BE->getRHS());
67      break;
68    }
69
70    case MCExpr::SymbolRef:
71      getAssembler().getOrCreateSymbolData(
72        cast<MCSymbolRefExpr>(Value)->getSymbol());
73      break;
74
75    case MCExpr::Unary:
76      AddValueSymbols(cast<MCUnaryExpr>(Value)->getSubExpr());
77      break;
78    }
79
80    return Value;
81  }
82
83  /// @name MCStreamer Interface
84  /// @{
85
86  virtual void EmitLabel(MCSymbol *Symbol);
87  virtual void EmitAssemblerFlag(MCAssemblerFlag Flag);
88  virtual void EmitAssignment(MCSymbol *Symbol, const MCExpr *Value);
89  virtual void EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute);
90  virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue);
91  virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size,
92                                unsigned ByteAlignment);
93  virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) {
94    assert(0 && "macho doesn't support this directive");
95  }
96  virtual void EmitCOFFSymbolStorageClass(int StorageClass) {
97    assert(0 && "macho doesn't support this directive");
98  }
99  virtual void EmitCOFFSymbolType(int Type) {
100    assert(0 && "macho doesn't support this directive");
101  }
102  virtual void EndCOFFSymbolDef() {
103    assert(0 && "macho doesn't support this directive");
104  }
105  virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) {
106    assert(0 && "macho doesn't support this directive");
107  }
108  virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size) {
109    assert(0 && "macho doesn't support this directive");
110  }
111  virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0,
112                            unsigned Size = 0, unsigned ByteAlignment = 0);
113  virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol,
114                              uint64_t Size, unsigned ByteAlignment = 0);
115  virtual void EmitBytes(StringRef Data, unsigned AddrSpace);
116  virtual void EmitValue(const MCExpr *Value, unsigned Size,unsigned AddrSpace);
117  virtual void EmitGPRel32Value(const MCExpr *Value) {
118    assert(0 && "macho doesn't support this directive");
119  }
120  virtual void EmitValueToAlignment(unsigned ByteAlignment, int64_t Value = 0,
121                                    unsigned ValueSize = 1,
122                                    unsigned MaxBytesToEmit = 0);
123  virtual void EmitCodeAlignment(unsigned ByteAlignment,
124                                 unsigned MaxBytesToEmit = 0);
125  virtual void EmitValueToOffset(const MCExpr *Offset,
126                                 unsigned char Value = 0);
127
128  virtual void EmitFileDirective(StringRef Filename) {
129    report_fatal_error("unsupported directive: '.file'");
130  }
131  virtual void EmitDwarfFileDirective(unsigned FileNo, StringRef Filename) {
132    report_fatal_error("unsupported directive: '.file'");
133  }
134
135  virtual void EmitInstruction(const MCInst &Inst);
136
137  virtual void Finish();
138
139  /// @}
140};
141
142} // end anonymous namespace.
143
144void MCMachOStreamer::EmitLabel(MCSymbol *Symbol) {
145  assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
146  assert(!Symbol->isVariable() && "Cannot emit a variable symbol!");
147  assert(CurSection && "Cannot emit before setting section!");
148
149  Symbol->setSection(*CurSection);
150
151  MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
152
153  // We have to create a new fragment if this is an atom defining symbol,
154  // fragments cannot span atoms.
155  if (getAssembler().isSymbolLinkerVisible(SD.getSymbol()))
156    new MCDataFragment(getCurrentSectionData());
157
158  // FIXME: This is wasteful, we don't necessarily need to create a data
159  // fragment. Instead, we should mark the symbol as pointing into the data
160  // fragment if it exists, otherwise we should just queue the label and set its
161  // fragment pointer when we emit the next fragment.
162  MCDataFragment *F = getOrCreateDataFragment();
163  assert(!SD.getFragment() && "Unexpected fragment on symbol data!");
164  SD.setFragment(F);
165  SD.setOffset(F->getContents().size());
166
167  // This causes the reference type flag to be cleared. Darwin 'as' was "trying"
168  // to clear the weak reference and weak definition bits too, but the
169  // implementation was buggy. For now we just try to match 'as', for
170  // diffability.
171  //
172  // FIXME: Cleanup this code, these bits should be emitted based on semantic
173  // properties, not on the order of definition, etc.
174  SD.setFlags(SD.getFlags() & ~SF_ReferenceTypeMask);
175}
176
177void MCMachOStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) {
178  switch (Flag) {
179  case MCAF_SubsectionsViaSymbols:
180    getAssembler().setSubsectionsViaSymbols(true);
181    return;
182  }
183
184  assert(0 && "invalid assembler flag!");
185}
186
187void MCMachOStreamer::EmitAssignment(MCSymbol *Symbol, const MCExpr *Value) {
188  // FIXME: Lift context changes into super class.
189  getAssembler().getOrCreateSymbolData(*Symbol);
190  Symbol->setVariableValue(AddValueSymbols(Value));
191}
192
193void MCMachOStreamer::EmitSymbolAttribute(MCSymbol *Symbol,
194                                          MCSymbolAttr Attribute) {
195  // Indirect symbols are handled differently, to match how 'as' handles
196  // them. This makes writing matching .o files easier.
197  if (Attribute == MCSA_IndirectSymbol) {
198    // Note that we intentionally cannot use the symbol data here; this is
199    // important for matching the string table that 'as' generates.
200    IndirectSymbolData ISD;
201    ISD.Symbol = Symbol;
202    ISD.SectionData = getCurrentSectionData();
203    getAssembler().getIndirectSymbols().push_back(ISD);
204    return;
205  }
206
207  // Adding a symbol attribute always introduces the symbol, note that an
208  // important side effect of calling getOrCreateSymbolData here is to register
209  // the symbol with the assembler.
210  MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
211
212  // The implementation of symbol attributes is designed to match 'as', but it
213  // leaves much to desired. It doesn't really make sense to arbitrarily add and
214  // remove flags, but 'as' allows this (in particular, see .desc).
215  //
216  // In the future it might be worth trying to make these operations more well
217  // defined.
218  switch (Attribute) {
219  case MCSA_Invalid:
220  case MCSA_ELF_TypeFunction:
221  case MCSA_ELF_TypeIndFunction:
222  case MCSA_ELF_TypeObject:
223  case MCSA_ELF_TypeTLS:
224  case MCSA_ELF_TypeCommon:
225  case MCSA_ELF_TypeNoType:
226  case MCSA_IndirectSymbol:
227  case MCSA_Hidden:
228  case MCSA_Internal:
229  case MCSA_Protected:
230  case MCSA_Weak:
231  case MCSA_Local:
232    assert(0 && "Invalid symbol attribute for Mach-O!");
233    break;
234
235  case MCSA_Global:
236    SD.setExternal(true);
237    // This effectively clears the undefined lazy bit, in Darwin 'as', although
238    // it isn't very consistent because it implements this as part of symbol
239    // lookup.
240    //
241    // FIXME: Cleanup this code, these bits should be emitted based on semantic
242    // properties, not on the order of definition, etc.
243    SD.setFlags(SD.getFlags() & ~SF_ReferenceTypeUndefinedLazy);
244    break;
245
246  case MCSA_LazyReference:
247    // FIXME: This requires -dynamic.
248    SD.setFlags(SD.getFlags() | SF_NoDeadStrip);
249    if (Symbol->isUndefined())
250      SD.setFlags(SD.getFlags() | SF_ReferenceTypeUndefinedLazy);
251    break;
252
253    // Since .reference sets the no dead strip bit, it is equivalent to
254    // .no_dead_strip in practice.
255  case MCSA_Reference:
256  case MCSA_NoDeadStrip:
257    SD.setFlags(SD.getFlags() | SF_NoDeadStrip);
258    break;
259
260  case MCSA_PrivateExtern:
261    SD.setExternal(true);
262    SD.setPrivateExtern(true);
263    break;
264
265  case MCSA_WeakReference:
266    // FIXME: This requires -dynamic.
267    if (Symbol->isUndefined())
268      SD.setFlags(SD.getFlags() | SF_WeakReference);
269    break;
270
271  case MCSA_WeakDefinition:
272    // FIXME: 'as' enforces that this is defined and global. The manual claims
273    // it has to be in a coalesced section, but this isn't enforced.
274    SD.setFlags(SD.getFlags() | SF_WeakDefinition);
275    break;
276  }
277}
278
279void MCMachOStreamer::EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) {
280  // Encode the 'desc' value into the lowest implementation defined bits.
281  assert(DescValue == (DescValue & SF_DescFlagsMask) &&
282         "Invalid .desc value!");
283  getAssembler().getOrCreateSymbolData(*Symbol).setFlags(
284    DescValue & SF_DescFlagsMask);
285}
286
287void MCMachOStreamer::EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size,
288                                       unsigned ByteAlignment) {
289  // FIXME: Darwin 'as' does appear to allow redef of a .comm by itself.
290  assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
291
292  MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
293  SD.setExternal(true);
294  SD.setCommon(Size, ByteAlignment);
295}
296
297void MCMachOStreamer::EmitZerofill(const MCSection *Section, MCSymbol *Symbol,
298                                   unsigned Size, unsigned ByteAlignment) {
299  MCSectionData &SectData = getAssembler().getOrCreateSectionData(*Section);
300
301  // The symbol may not be present, which only creates the section.
302  if (!Symbol)
303    return;
304
305  // FIXME: Assert that this section has the zerofill type.
306
307  assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
308
309  MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
310
311  // Emit an align fragment if necessary.
312  if (ByteAlignment != 1)
313    new MCAlignFragment(ByteAlignment, 0, 0, ByteAlignment, &SectData);
314
315  MCFragment *F = new MCFillFragment(0, 0, Size, &SectData);
316  SD.setFragment(F);
317
318  Symbol->setSection(*Section);
319
320  // Update the maximum alignment on the zero fill section if necessary.
321  if (ByteAlignment > SectData.getAlignment())
322    SectData.setAlignment(ByteAlignment);
323}
324
325// This should always be called with the thread local bss section.  Like the
326// .zerofill directive this doesn't actually switch sections on us.
327void MCMachOStreamer::EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol,
328                                     uint64_t Size, unsigned ByteAlignment) {
329  EmitZerofill(Section, Symbol, Size, ByteAlignment);
330  return;
331}
332
333void MCMachOStreamer::EmitBytes(StringRef Data, unsigned AddrSpace) {
334  getOrCreateDataFragment()->getContents().append(Data.begin(), Data.end());
335}
336
337void MCMachOStreamer::EmitValue(const MCExpr *Value, unsigned Size,
338                                unsigned AddrSpace) {
339  MCDataFragment *DF = getOrCreateDataFragment();
340
341  // Avoid fixups when possible.
342  int64_t AbsValue;
343  if (AddValueSymbols(Value)->EvaluateAsAbsolute(AbsValue)) {
344    // FIXME: Endianness assumption.
345    for (unsigned i = 0; i != Size; ++i)
346      DF->getContents().push_back(uint8_t(AbsValue >> (i * 8)));
347  } else {
348    DF->addFixup(MCFixup::Create(DF->getContents().size(),
349                                 AddValueSymbols(Value),
350                                 MCFixup::getKindForSize(Size)));
351    DF->getContents().resize(DF->getContents().size() + Size, 0);
352  }
353}
354
355void MCMachOStreamer::EmitValueToAlignment(unsigned ByteAlignment,
356                                           int64_t Value, unsigned ValueSize,
357                                           unsigned MaxBytesToEmit) {
358  if (MaxBytesToEmit == 0)
359    MaxBytesToEmit = ByteAlignment;
360  new MCAlignFragment(ByteAlignment, Value, ValueSize, MaxBytesToEmit,
361                      getCurrentSectionData());
362
363  // Update the maximum alignment on the current section if necessary.
364  if (ByteAlignment > getCurrentSectionData()->getAlignment())
365    getCurrentSectionData()->setAlignment(ByteAlignment);
366}
367
368void MCMachOStreamer::EmitCodeAlignment(unsigned ByteAlignment,
369                                        unsigned MaxBytesToEmit) {
370  if (MaxBytesToEmit == 0)
371    MaxBytesToEmit = ByteAlignment;
372  MCAlignFragment *F = new MCAlignFragment(ByteAlignment, 0, 1, MaxBytesToEmit,
373                                           getCurrentSectionData());
374  F->setEmitNops(true);
375
376  // Update the maximum alignment on the current section if necessary.
377  if (ByteAlignment > getCurrentSectionData()->getAlignment())
378    getCurrentSectionData()->setAlignment(ByteAlignment);
379}
380
381void MCMachOStreamer::EmitValueToOffset(const MCExpr *Offset,
382                                        unsigned char Value) {
383  new MCOrgFragment(*Offset, Value, getCurrentSectionData());
384}
385
386void MCMachOStreamer::EmitInstToFragment(const MCInst &Inst) {
387  MCInstFragment *IF = new MCInstFragment(Inst, getCurrentSectionData());
388
389  // Add the fixups and data.
390  //
391  // FIXME: Revisit this design decision when relaxation is done, we may be
392  // able to get away with not storing any extra data in the MCInst.
393  SmallVector<MCFixup, 4> Fixups;
394  SmallString<256> Code;
395  raw_svector_ostream VecOS(Code);
396  getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups);
397  VecOS.flush();
398
399  IF->getCode() = Code;
400  IF->getFixups() = Fixups;
401}
402
403void MCMachOStreamer::EmitInstToData(const MCInst &Inst) {
404  MCDataFragment *DF = getOrCreateDataFragment();
405
406  SmallVector<MCFixup, 4> Fixups;
407  SmallString<256> Code;
408  raw_svector_ostream VecOS(Code);
409  getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups);
410  VecOS.flush();
411
412  // Add the fixups and data.
413  for (unsigned i = 0, e = Fixups.size(); i != e; ++i) {
414    Fixups[i].setOffset(Fixups[i].getOffset() + DF->getContents().size());
415    DF->addFixup(Fixups[i]);
416  }
417  DF->getContents().append(Code.begin(), Code.end());
418}
419
420void MCMachOStreamer::EmitInstruction(const MCInst &Inst) {
421  // Scan for values.
422  for (unsigned i = Inst.getNumOperands(); i--; )
423    if (Inst.getOperand(i).isExpr())
424      AddValueSymbols(Inst.getOperand(i).getExpr());
425
426  getCurrentSectionData()->setHasInstructions(true);
427
428  // If this instruction doesn't need relaxation, just emit it as data.
429  if (!getAssembler().getBackend().MayNeedRelaxation(Inst)) {
430    EmitInstToData(Inst);
431    return;
432  }
433
434  // Otherwise, if we are relaxing everything, relax the instruction as much as
435  // possible and emit it as data.
436  if (getAssembler().getRelaxAll()) {
437    MCInst Relaxed;
438    getAssembler().getBackend().RelaxInstruction(Inst, Relaxed);
439    while (getAssembler().getBackend().MayNeedRelaxation(Relaxed))
440      getAssembler().getBackend().RelaxInstruction(Relaxed, Relaxed);
441    EmitInstToData(Relaxed);
442    return;
443  }
444
445  // Otherwise emit to a separate fragment.
446  EmitInstToFragment(Inst);
447}
448
449void MCMachOStreamer::Finish() {
450  // We have to set the fragment atom associations so we can relax properly for
451  // Mach-O.
452
453  // First, scan the symbol table to build a lookup table from fragments to
454  // defining symbols.
455  DenseMap<const MCFragment*, MCSymbolData*> DefiningSymbolMap;
456  for (MCAssembler::symbol_iterator it = getAssembler().symbol_begin(),
457         ie = getAssembler().symbol_end(); it != ie; ++it) {
458    if (getAssembler().isSymbolLinkerVisible(it->getSymbol()) &&
459        it->getFragment()) {
460      // An atom defining symbol should never be internal to a fragment.
461      assert(it->getOffset() == 0 && "Invalid offset in atom defining symbol!");
462      DefiningSymbolMap[it->getFragment()] = it;
463    }
464  }
465
466  // Set the fragment atom associations by tracking the last seen atom defining
467  // symbol.
468  for (MCAssembler::iterator it = getAssembler().begin(),
469         ie = getAssembler().end(); it != ie; ++it) {
470    MCSymbolData *CurrentAtom = 0;
471    for (MCSectionData::iterator it2 = it->begin(),
472           ie2 = it->end(); it2 != ie2; ++it2) {
473      if (MCSymbolData *SD = DefiningSymbolMap.lookup(it2))
474        CurrentAtom = SD;
475      it2->setAtom(CurrentAtom);
476    }
477  }
478
479  this->MCObjectStreamer::Finish();
480}
481
482MCStreamer *llvm::createMachOStreamer(MCContext &Context, TargetAsmBackend &TAB,
483                                      raw_ostream &OS, MCCodeEmitter *CE,
484                                      bool RelaxAll) {
485  MCMachOStreamer *S = new MCMachOStreamer(Context, TAB, OS, CE);
486  if (RelaxAll)
487    S->getAssembler().setRelaxAll(true);
488  return S;
489}
490