1 //===-- MCMachOStreamer.cpp - MachO Streamer ------------------------------===//
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 #include "llvm/MC/MCAsmBackend.h"
12 #include "llvm/MC/MCAssembler.h"
13 #include "llvm/MC/MCCodeEmitter.h"
14 #include "llvm/MC/MCContext.h"
15 #include "llvm/MC/MCDwarf.h"
16 #include "llvm/MC/MCExpr.h"
17 #include "llvm/MC/MCInst.h"
18 #include "llvm/MC/MCMachOSymbolFlags.h"
19 #include "llvm/MC/MCObjectFileInfo.h"
20 #include "llvm/MC/MCObjectStreamer.h"
21 #include "llvm/MC/MCSection.h"
22 #include "llvm/MC/MCSectionMachO.h"
23 #include "llvm/MC/MCSymbol.h"
24 #include "llvm/Support/Dwarf.h"
25 #include "llvm/Support/ErrorHandling.h"
26 #include "llvm/Support/raw_ostream.h"
27 
28 using namespace llvm;
29 
30 namespace {
31 
32 class MCMachOStreamer : public MCObjectStreamer {
33 private:
34   virtual void EmitInstToData(const MCInst &Inst, const MCSubtargetInfo &STI);
35 
36   void EmitDataRegion(DataRegionData::KindTy Kind);
37   void EmitDataRegionEnd();
38 public:
39   MCMachOStreamer(MCContext &Context, MCAsmBackend &MAB, raw_ostream &OS,
40                   MCCodeEmitter *Emitter)
41       : MCObjectStreamer(Context, MAB, OS, Emitter) {}
42 
43   /// @name MCStreamer Interface
44   /// @{
45 
46   virtual void EmitLabel(MCSymbol *Symbol);
47   virtual void EmitDebugLabel(MCSymbol *Symbol);
48   virtual void EmitEHSymAttributes(const MCSymbol *Symbol,
49                                    MCSymbol *EHSymbol);
50   virtual void EmitAssemblerFlag(MCAssemblerFlag Flag);
51   virtual void EmitLinkerOptions(ArrayRef<std::string> Options);
52   virtual void EmitDataRegion(MCDataRegionType Kind);
53   virtual void EmitThumbFunc(MCSymbol *Func);
54   virtual bool EmitSymbolAttribute(MCSymbol *Symbol, MCSymbolAttr Attribute);
55   virtual void EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue);
56   virtual void EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size,
57                                 unsigned ByteAlignment);
58   virtual void BeginCOFFSymbolDef(const MCSymbol *Symbol) {
59     llvm_unreachable("macho doesn't support this directive");
60   }
61   virtual void EmitCOFFSymbolStorageClass(int StorageClass) {
62     llvm_unreachable("macho doesn't support this directive");
63   }
64   virtual void EmitCOFFSymbolType(int Type) {
65     llvm_unreachable("macho doesn't support this directive");
66   }
67   virtual void EndCOFFSymbolDef() {
68     llvm_unreachable("macho doesn't support this directive");
69   }
70   virtual void EmitELFSize(MCSymbol *Symbol, const MCExpr *Value) {
71     llvm_unreachable("macho doesn't support this directive");
72   }
73   virtual void EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size,
74                                      unsigned ByteAlignment);
75   virtual void EmitZerofill(const MCSection *Section, MCSymbol *Symbol = 0,
76                             uint64_t Size = 0, unsigned ByteAlignment = 0);
77   virtual void EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol,
78                               uint64_t Size, unsigned ByteAlignment = 0);
79 
80   virtual void EmitFileDirective(StringRef Filename) {
81     // FIXME: Just ignore the .file; it isn't important enough to fail the
82     // entire assembly.
83 
84     // report_fatal_error("unsupported directive: '.file'");
85   }
86 
87   virtual void EmitIdent(StringRef IdentString) {
88     llvm_unreachable("macho doesn't support this directive");
89   }
90 
91   virtual void FinishImpl();
92 };
93 
94 } // end anonymous namespace.
95 
96 void MCMachOStreamer::EmitEHSymAttributes(const MCSymbol *Symbol,
97                                           MCSymbol *EHSymbol) {
98   MCSymbolData &SD =
99     getAssembler().getOrCreateSymbolData(*Symbol);
100   if (SD.isExternal())
101     EmitSymbolAttribute(EHSymbol, MCSA_Global);
102   if (SD.getFlags() & SF_WeakDefinition)
103     EmitSymbolAttribute(EHSymbol, MCSA_WeakDefinition);
104   if (SD.isPrivateExtern())
105     EmitSymbolAttribute(EHSymbol, MCSA_PrivateExtern);
106 }
107 
108 void MCMachOStreamer::EmitLabel(MCSymbol *Symbol) {
109   assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
110 
111   // isSymbolLinkerVisible uses the section.
112   AssignSection(Symbol, getCurrentSection().first);
113   // We have to create a new fragment if this is an atom defining symbol,
114   // fragments cannot span atoms.
115   if (getAssembler().isSymbolLinkerVisible(*Symbol))
116     insert(new MCDataFragment());
117 
118   MCObjectStreamer::EmitLabel(Symbol);
119 
120   MCSymbolData &SD = getAssembler().getSymbolData(*Symbol);
121   // This causes the reference type flag to be cleared. Darwin 'as' was "trying"
122   // to clear the weak reference and weak definition bits too, but the
123   // implementation was buggy. For now we just try to match 'as', for
124   // diffability.
125   //
126   // FIXME: Cleanup this code, these bits should be emitted based on semantic
127   // properties, not on the order of definition, etc.
128   SD.setFlags(SD.getFlags() & ~SF_ReferenceTypeMask);
129 }
130 
131 void MCMachOStreamer::EmitDebugLabel(MCSymbol *Symbol) {
132   EmitLabel(Symbol);
133 }
134 void MCMachOStreamer::EmitDataRegion(DataRegionData::KindTy Kind) {
135   if (!getAssembler().getBackend().hasDataInCodeSupport())
136     return;
137   // Create a temporary label to mark the start of the data region.
138   MCSymbol *Start = getContext().CreateTempSymbol();
139   EmitLabel(Start);
140   // Record the region for the object writer to use.
141   DataRegionData Data = { Kind, Start, NULL };
142   std::vector<DataRegionData> &Regions = getAssembler().getDataRegions();
143   Regions.push_back(Data);
144 }
145 
146 void MCMachOStreamer::EmitDataRegionEnd() {
147   if (!getAssembler().getBackend().hasDataInCodeSupport())
148     return;
149   std::vector<DataRegionData> &Regions = getAssembler().getDataRegions();
150   assert(Regions.size() && "Mismatched .end_data_region!");
151   DataRegionData &Data = Regions.back();
152   assert(Data.End == NULL && "Mismatched .end_data_region!");
153   // Create a temporary label to mark the end of the data region.
154   Data.End = getContext().CreateTempSymbol();
155   EmitLabel(Data.End);
156 }
157 
158 void MCMachOStreamer::EmitAssemblerFlag(MCAssemblerFlag Flag) {
159   // Let the target do whatever target specific stuff it needs to do.
160   getAssembler().getBackend().handleAssemblerFlag(Flag);
161   // Do any generic stuff we need to do.
162   switch (Flag) {
163   case MCAF_SyntaxUnified: return; // no-op here.
164   case MCAF_Code16: return; // Change parsing mode; no-op here.
165   case MCAF_Code32: return; // Change parsing mode; no-op here.
166   case MCAF_Code64: return; // Change parsing mode; no-op here.
167   case MCAF_SubsectionsViaSymbols:
168     getAssembler().setSubsectionsViaSymbols(true);
169     return;
170   }
171 }
172 
173 void MCMachOStreamer::EmitLinkerOptions(ArrayRef<std::string> Options) {
174   getAssembler().getLinkerOptions().push_back(Options);
175 }
176 
177 void MCMachOStreamer::EmitDataRegion(MCDataRegionType Kind) {
178   switch (Kind) {
179   case MCDR_DataRegion:
180     EmitDataRegion(DataRegionData::Data);
181     return;
182   case MCDR_DataRegionJT8:
183     EmitDataRegion(DataRegionData::JumpTable8);
184     return;
185   case MCDR_DataRegionJT16:
186     EmitDataRegion(DataRegionData::JumpTable16);
187     return;
188   case MCDR_DataRegionJT32:
189     EmitDataRegion(DataRegionData::JumpTable32);
190     return;
191   case MCDR_DataRegionEnd:
192     EmitDataRegionEnd();
193     return;
194   }
195 }
196 
197 void MCMachOStreamer::EmitThumbFunc(MCSymbol *Symbol) {
198   // Remember that the function is a thumb function. Fixup and relocation
199   // values will need adjusted.
200   getAssembler().setIsThumbFunc(Symbol);
201 
202   // Mark the thumb bit on the symbol.
203   MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
204   SD.setFlags(SD.getFlags() | SF_ThumbFunc);
205 }
206 
207 bool MCMachOStreamer::EmitSymbolAttribute(MCSymbol *Symbol,
208                                           MCSymbolAttr Attribute) {
209   // Indirect symbols are handled differently, to match how 'as' handles
210   // them. This makes writing matching .o files easier.
211   if (Attribute == MCSA_IndirectSymbol) {
212     // Note that we intentionally cannot use the symbol data here; this is
213     // important for matching the string table that 'as' generates.
214     IndirectSymbolData ISD;
215     ISD.Symbol = Symbol;
216     ISD.SectionData = getCurrentSectionData();
217     getAssembler().getIndirectSymbols().push_back(ISD);
218     return true;
219   }
220 
221   // Adding a symbol attribute always introduces the symbol, note that an
222   // important side effect of calling getOrCreateSymbolData here is to register
223   // the symbol with the assembler.
224   MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
225 
226   // The implementation of symbol attributes is designed to match 'as', but it
227   // leaves much to desired. It doesn't really make sense to arbitrarily add and
228   // remove flags, but 'as' allows this (in particular, see .desc).
229   //
230   // In the future it might be worth trying to make these operations more well
231   // defined.
232   switch (Attribute) {
233   case MCSA_Invalid:
234   case MCSA_ELF_TypeFunction:
235   case MCSA_ELF_TypeIndFunction:
236   case MCSA_ELF_TypeObject:
237   case MCSA_ELF_TypeTLS:
238   case MCSA_ELF_TypeCommon:
239   case MCSA_ELF_TypeNoType:
240   case MCSA_ELF_TypeGnuUniqueObject:
241   case MCSA_Hidden:
242   case MCSA_IndirectSymbol:
243   case MCSA_Internal:
244   case MCSA_Protected:
245   case MCSA_Weak:
246   case MCSA_Local:
247     return false;
248 
249   case MCSA_Global:
250     SD.setExternal(true);
251     // This effectively clears the undefined lazy bit, in Darwin 'as', although
252     // it isn't very consistent because it implements this as part of symbol
253     // lookup.
254     //
255     // FIXME: Cleanup this code, these bits should be emitted based on semantic
256     // properties, not on the order of definition, etc.
257     SD.setFlags(SD.getFlags() & ~SF_ReferenceTypeUndefinedLazy);
258     break;
259 
260   case MCSA_LazyReference:
261     // FIXME: This requires -dynamic.
262     SD.setFlags(SD.getFlags() | SF_NoDeadStrip);
263     if (Symbol->isUndefined())
264       SD.setFlags(SD.getFlags() | SF_ReferenceTypeUndefinedLazy);
265     break;
266 
267     // Since .reference sets the no dead strip bit, it is equivalent to
268     // .no_dead_strip in practice.
269   case MCSA_Reference:
270   case MCSA_NoDeadStrip:
271     SD.setFlags(SD.getFlags() | SF_NoDeadStrip);
272     break;
273 
274   case MCSA_SymbolResolver:
275     SD.setFlags(SD.getFlags() | SF_SymbolResolver);
276     break;
277 
278   case MCSA_PrivateExtern:
279     SD.setExternal(true);
280     SD.setPrivateExtern(true);
281     break;
282 
283   case MCSA_WeakReference:
284     // FIXME: This requires -dynamic.
285     if (Symbol->isUndefined())
286       SD.setFlags(SD.getFlags() | SF_WeakReference);
287     break;
288 
289   case MCSA_WeakDefinition:
290     // FIXME: 'as' enforces that this is defined and global. The manual claims
291     // it has to be in a coalesced section, but this isn't enforced.
292     SD.setFlags(SD.getFlags() | SF_WeakDefinition);
293     break;
294 
295   case MCSA_WeakDefAutoPrivate:
296     SD.setFlags(SD.getFlags() | SF_WeakDefinition | SF_WeakReference);
297     break;
298   }
299 
300   return true;
301 }
302 
303 void MCMachOStreamer::EmitSymbolDesc(MCSymbol *Symbol, unsigned DescValue) {
304   // Encode the 'desc' value into the lowest implementation defined bits.
305   assert(DescValue == (DescValue & SF_DescFlagsMask) &&
306          "Invalid .desc value!");
307   getAssembler().getOrCreateSymbolData(*Symbol).setFlags(
308     DescValue & SF_DescFlagsMask);
309 }
310 
311 void MCMachOStreamer::EmitCommonSymbol(MCSymbol *Symbol, uint64_t Size,
312                                        unsigned ByteAlignment) {
313   // FIXME: Darwin 'as' does appear to allow redef of a .comm by itself.
314   assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
315 
316   AssignSection(Symbol, NULL);
317 
318   MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
319   SD.setExternal(true);
320   SD.setCommon(Size, ByteAlignment);
321 }
322 
323 void MCMachOStreamer::EmitLocalCommonSymbol(MCSymbol *Symbol, uint64_t Size,
324                                             unsigned ByteAlignment) {
325   // '.lcomm' is equivalent to '.zerofill'.
326   return EmitZerofill(getContext().getObjectFileInfo()->getDataBSSSection(),
327                       Symbol, Size, ByteAlignment);
328 }
329 
330 void MCMachOStreamer::EmitZerofill(const MCSection *Section, MCSymbol *Symbol,
331                                    uint64_t Size, unsigned ByteAlignment) {
332   MCSectionData &SectData = getAssembler().getOrCreateSectionData(*Section);
333 
334   // The symbol may not be present, which only creates the section.
335   if (!Symbol)
336     return;
337 
338   // On darwin all virtual sections have zerofill type.
339   assert(Section->isVirtualSection() && "Section does not have zerofill type!");
340 
341   assert(Symbol->isUndefined() && "Cannot define a symbol twice!");
342 
343   MCSymbolData &SD = getAssembler().getOrCreateSymbolData(*Symbol);
344 
345   // Emit an align fragment if necessary.
346   if (ByteAlignment != 1)
347     new MCAlignFragment(ByteAlignment, 0, 0, ByteAlignment, &SectData);
348 
349   MCFragment *F = new MCFillFragment(0, 0, Size, &SectData);
350   SD.setFragment(F);
351 
352   AssignSection(Symbol, Section);
353 
354   // Update the maximum alignment on the zero fill section if necessary.
355   if (ByteAlignment > SectData.getAlignment())
356     SectData.setAlignment(ByteAlignment);
357 }
358 
359 // This should always be called with the thread local bss section.  Like the
360 // .zerofill directive this doesn't actually switch sections on us.
361 void MCMachOStreamer::EmitTBSSSymbol(const MCSection *Section, MCSymbol *Symbol,
362                                      uint64_t Size, unsigned ByteAlignment) {
363   EmitZerofill(Section, Symbol, Size, ByteAlignment);
364   return;
365 }
366 
367 void MCMachOStreamer::EmitInstToData(const MCInst &Inst,
368                                      const MCSubtargetInfo &STI) {
369   MCDataFragment *DF = getOrCreateDataFragment();
370 
371   SmallVector<MCFixup, 4> Fixups;
372   SmallString<256> Code;
373   raw_svector_ostream VecOS(Code);
374   getAssembler().getEmitter().EncodeInstruction(Inst, VecOS, Fixups, STI);
375   VecOS.flush();
376 
377   // Add the fixups and data.
378   for (unsigned i = 0, e = Fixups.size(); i != e; ++i) {
379     Fixups[i].setOffset(Fixups[i].getOffset() + DF->getContents().size());
380     DF->getFixups().push_back(Fixups[i]);
381   }
382   DF->getContents().append(Code.begin(), Code.end());
383 }
384 
385 void MCMachOStreamer::FinishImpl() {
386   EmitFrames(&getAssembler().getBackend(), true);
387 
388   // We have to set the fragment atom associations so we can relax properly for
389   // Mach-O.
390 
391   // First, scan the symbol table to build a lookup table from fragments to
392   // defining symbols.
393   DenseMap<const MCFragment*, MCSymbolData*> DefiningSymbolMap;
394   for (MCAssembler::symbol_iterator it = getAssembler().symbol_begin(),
395          ie = getAssembler().symbol_end(); it != ie; ++it) {
396     if (getAssembler().isSymbolLinkerVisible(it->getSymbol()) &&
397         it->getFragment()) {
398       // An atom defining symbol should never be internal to a fragment.
399       assert(it->getOffset() == 0 && "Invalid offset in atom defining symbol!");
400       DefiningSymbolMap[it->getFragment()] = it;
401     }
402   }
403 
404   // Set the fragment atom associations by tracking the last seen atom defining
405   // symbol.
406   for (MCAssembler::iterator it = getAssembler().begin(),
407          ie = getAssembler().end(); it != ie; ++it) {
408     MCSymbolData *CurrentAtom = 0;
409     for (MCSectionData::iterator it2 = it->begin(),
410            ie2 = it->end(); it2 != ie2; ++it2) {
411       if (MCSymbolData *SD = DefiningSymbolMap.lookup(it2))
412         CurrentAtom = SD;
413       it2->setAtom(CurrentAtom);
414     }
415   }
416 
417   this->MCObjectStreamer::FinishImpl();
418 }
419 
420 MCStreamer *llvm::createMachOStreamer(MCContext &Context, MCAsmBackend &MAB,
421                                       raw_ostream &OS, MCCodeEmitter *CE,
422                                       bool RelaxAll) {
423   MCMachOStreamer *S = new MCMachOStreamer(Context, MAB, OS, CE);
424   if (RelaxAll)
425     S->getAssembler().setRelaxAll(true);
426   return S;
427 }
428