1//===- MCAssembler.h - Object File Generation -------------------*- C++ -*-===//
2//
3// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4// See https://llvm.org/LICENSE.txt for license information.
5// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6//
7//===----------------------------------------------------------------------===//
8
9#ifndef LLVM_MC_MCASSEMBLER_H
10#define LLVM_MC_MCASSEMBLER_H
11
12#include "llvm/ADT/ArrayRef.h"
13#include "llvm/ADT/STLExtras.h"
14#include "llvm/ADT/SmallPtrSet.h"
15#include "llvm/ADT/StringRef.h"
16#include "llvm/ADT/iterator.h"
17#include "llvm/ADT/iterator_range.h"
18#include "llvm/BinaryFormat/MachO.h"
19#include "llvm/MC/MCDirectives.h"
20#include "llvm/MC/MCDwarf.h"
21#include "llvm/MC/MCFixup.h"
22#include "llvm/MC/MCFragment.h"
23#include "llvm/MC/MCLinkerOptimizationHint.h"
24#include "llvm/MC/MCSymbol.h"
25#include "llvm/Support/VersionTuple.h"
26#include <cassert>
27#include <cstddef>
28#include <cstdint>
29#include <string>
30#include <utility>
31#include <vector>
32
33namespace llvm {
34
35class MCAsmBackend;
36class MCAsmLayout;
37class MCContext;
38class MCCodeEmitter;
39class MCFragment;
40class MCObjectWriter;
41class MCSection;
42class MCValue;
43
44// FIXME: This really doesn't belong here. See comments below.
45struct IndirectSymbolData {
46  MCSymbol *Symbol;
47  MCSection *Section;
48};
49
50// FIXME: Ditto this. Purely so the Streamer and the ObjectWriter can talk
51// to one another.
52struct DataRegionData {
53  // This enum should be kept in sync w/ the mach-o definition in
54  // llvm/Object/MachOFormat.h.
55  enum KindTy { Data = 1, JumpTable8, JumpTable16, JumpTable32 } Kind;
56  MCSymbol *Start;
57  MCSymbol *End;
58};
59
60class MCAssembler {
61  friend class MCAsmLayout;
62
63public:
64  using SectionListType = std::vector<MCSection *>;
65  using SymbolDataListType = std::vector<const MCSymbol *>;
66
67  using const_iterator = pointee_iterator<SectionListType::const_iterator>;
68  using iterator = pointee_iterator<SectionListType::iterator>;
69
70  using const_symbol_iterator =
71      pointee_iterator<SymbolDataListType::const_iterator>;
72  using symbol_iterator = pointee_iterator<SymbolDataListType::iterator>;
73
74  using symbol_range = iterator_range<symbol_iterator>;
75  using const_symbol_range = iterator_range<const_symbol_iterator>;
76
77  using const_indirect_symbol_iterator =
78      std::vector<IndirectSymbolData>::const_iterator;
79  using indirect_symbol_iterator = std::vector<IndirectSymbolData>::iterator;
80
81  using const_data_region_iterator =
82      std::vector<DataRegionData>::const_iterator;
83  using data_region_iterator = std::vector<DataRegionData>::iterator;
84
85  /// MachO specific deployment target version info.
86  // A Major version of 0 indicates that no version information was supplied
87  // and so the corresponding load command should not be emitted.
88  using VersionInfoType = struct {
89    bool EmitBuildVersion;
90    union {
91      MCVersionMinType Type;          ///< Used when EmitBuildVersion==false.
92      MachO::PlatformType Platform;   ///< Used when EmitBuildVersion==true.
93    } TypeOrPlatform;
94    unsigned Major;
95    unsigned Minor;
96    unsigned Update;
97    /// An optional version of the SDK that was used to build the source.
98    VersionTuple SDKVersion;
99  };
100
101private:
102  MCContext &Context;
103
104  std::unique_ptr<MCAsmBackend> Backend;
105
106  std::unique_ptr<MCCodeEmitter> Emitter;
107
108  std::unique_ptr<MCObjectWriter> Writer;
109
110  SectionListType Sections;
111
112  SymbolDataListType Symbols;
113
114  std::vector<IndirectSymbolData> IndirectSymbols;
115
116  std::vector<DataRegionData> DataRegions;
117
118  /// The list of linker options to propagate into the object file.
119  std::vector<std::vector<std::string>> LinkerOptions;
120
121  /// List of declared file names
122  std::vector<std::string> FileNames;
123
124  MCDwarfLineTableParams LTParams;
125
126  /// The set of function symbols for which a .thumb_func directive has
127  /// been seen.
128  //
129  // FIXME: We really would like this in target specific code rather than
130  // here. Maybe when the relocation stuff moves to target specific,
131  // this can go with it? The streamer would need some target specific
132  // refactoring too.
133  mutable SmallPtrSet<const MCSymbol *, 32> ThumbFuncs;
134
135  /// The bundle alignment size currently set in the assembler.
136  ///
137  /// By default it's 0, which means bundling is disabled.
138  unsigned BundleAlignSize;
139
140  bool RelaxAll : 1;
141  bool SubsectionsViaSymbols : 1;
142  bool IncrementalLinkerCompatible : 1;
143
144  /// ELF specific e_header flags
145  // It would be good if there were an MCELFAssembler class to hold this.
146  // ELF header flags are used both by the integrated and standalone assemblers.
147  // Access to the flags is necessary in cases where assembler directives affect
148  // which flags to be set.
149  unsigned ELFHeaderEFlags;
150
151  /// Used to communicate Linker Optimization Hint information between
152  /// the Streamer and the .o writer
153  MCLOHContainer LOHContainer;
154
155  VersionInfoType VersionInfo;
156
157  /// Evaluate a fixup to a relocatable expression and the value which should be
158  /// placed into the fixup.
159  ///
160  /// \param Layout The layout to use for evaluation.
161  /// \param Fixup The fixup to evaluate.
162  /// \param DF The fragment the fixup is inside.
163  /// \param Target [out] On return, the relocatable expression the fixup
164  /// evaluates to.
165  /// \param Value [out] On return, the value of the fixup as currently laid
166  /// out.
167  /// \param WasForced [out] On return, the value in the fixup is set to the
168  /// correct value if WasForced is true, even if evaluateFixup returns false.
169  /// \return Whether the fixup value was fully resolved. This is true if the
170  /// \p Value result is fixed, otherwise the value may change due to
171  /// relocation.
172  bool evaluateFixup(const MCAsmLayout &Layout, const MCFixup &Fixup,
173                     const MCFragment *DF, MCValue &Target,
174                     uint64_t &Value, bool &WasForced) const;
175
176  /// Check whether a fixup can be satisfied, or whether it needs to be relaxed
177  /// (increased in size, in order to hold its value correctly).
178  bool fixupNeedsRelaxation(const MCFixup &Fixup, const MCRelaxableFragment *DF,
179                            const MCAsmLayout &Layout) const;
180
181  /// Check whether the given fragment needs relaxation.
182  bool fragmentNeedsRelaxation(const MCRelaxableFragment *IF,
183                               const MCAsmLayout &Layout) const;
184
185  /// Perform one layout iteration and return true if any offsets
186  /// were adjusted.
187  bool layoutOnce(MCAsmLayout &Layout);
188
189  /// Perform one layout iteration of the given section and return true
190  /// if any offsets were adjusted.
191  bool layoutSectionOnce(MCAsmLayout &Layout, MCSection &Sec);
192
193  bool relaxInstruction(MCAsmLayout &Layout, MCRelaxableFragment &IF);
194  bool relaxLEB(MCAsmLayout &Layout, MCLEBFragment &IF);
195  bool relaxBoundaryAlign(MCAsmLayout &Layout, MCBoundaryAlignFragment &BF);
196  bool relaxDwarfLineAddr(MCAsmLayout &Layout, MCDwarfLineAddrFragment &DF);
197  bool relaxDwarfCallFrameFragment(MCAsmLayout &Layout,
198                                   MCDwarfCallFrameFragment &DF);
199  bool relaxCVInlineLineTable(MCAsmLayout &Layout,
200                              MCCVInlineLineTableFragment &DF);
201  bool relaxCVDefRange(MCAsmLayout &Layout, MCCVDefRangeFragment &DF);
202
203  /// finishLayout - Finalize a layout, including fragment lowering.
204  void finishLayout(MCAsmLayout &Layout);
205
206  std::tuple<MCValue, uint64_t, bool>
207  handleFixup(const MCAsmLayout &Layout, MCFragment &F, const MCFixup &Fixup);
208
209public:
210  std::vector<std::pair<StringRef, const MCSymbol *>> Symvers;
211
212  /// Construct a new assembler instance.
213  //
214  // FIXME: How are we going to parameterize this? Two obvious options are stay
215  // concrete and require clients to pass in a target like object. The other
216  // option is to make this abstract, and have targets provide concrete
217  // implementations as we do with AsmParser.
218  MCAssembler(MCContext &Context, std::unique_ptr<MCAsmBackend> Backend,
219              std::unique_ptr<MCCodeEmitter> Emitter,
220              std::unique_ptr<MCObjectWriter> Writer);
221  MCAssembler(const MCAssembler &) = delete;
222  MCAssembler &operator=(const MCAssembler &) = delete;
223  ~MCAssembler();
224
225  /// Compute the effective fragment size assuming it is laid out at the given
226  /// \p SectionAddress and \p FragmentOffset.
227  uint64_t computeFragmentSize(const MCAsmLayout &Layout,
228                               const MCFragment &F) const;
229
230  /// Find the symbol which defines the atom containing the given symbol, or
231  /// null if there is no such symbol.
232  const MCSymbol *getAtom(const MCSymbol &S) const;
233
234  /// Check whether a particular symbol is visible to the linker and is required
235  /// in the symbol table, or whether it can be discarded by the assembler. This
236  /// also effects whether the assembler treats the label as potentially
237  /// defining a separate atom.
238  bool isSymbolLinkerVisible(const MCSymbol &SD) const;
239
240  /// Emit the section contents to \p OS.
241  void writeSectionData(raw_ostream &OS, const MCSection *Section,
242                        const MCAsmLayout &Layout) const;
243
244  /// Check whether a given symbol has been flagged with .thumb_func.
245  bool isThumbFunc(const MCSymbol *Func) const;
246
247  /// Flag a function symbol as the target of a .thumb_func directive.
248  void setIsThumbFunc(const MCSymbol *Func) { ThumbFuncs.insert(Func); }
249
250  /// ELF e_header flags
251  unsigned getELFHeaderEFlags() const { return ELFHeaderEFlags; }
252  void setELFHeaderEFlags(unsigned Flags) { ELFHeaderEFlags = Flags; }
253
254  /// MachO deployment target version information.
255  const VersionInfoType &getVersionInfo() const { return VersionInfo; }
256  void setVersionMin(MCVersionMinType Type, unsigned Major, unsigned Minor,
257                     unsigned Update,
258                     VersionTuple SDKVersion = VersionTuple()) {
259    VersionInfo.EmitBuildVersion = false;
260    VersionInfo.TypeOrPlatform.Type = Type;
261    VersionInfo.Major = Major;
262    VersionInfo.Minor = Minor;
263    VersionInfo.Update = Update;
264    VersionInfo.SDKVersion = SDKVersion;
265  }
266  void setBuildVersion(MachO::PlatformType Platform, unsigned Major,
267                       unsigned Minor, unsigned Update,
268                       VersionTuple SDKVersion = VersionTuple()) {
269    VersionInfo.EmitBuildVersion = true;
270    VersionInfo.TypeOrPlatform.Platform = Platform;
271    VersionInfo.Major = Major;
272    VersionInfo.Minor = Minor;
273    VersionInfo.Update = Update;
274    VersionInfo.SDKVersion = SDKVersion;
275  }
276
277  /// Reuse an assembler instance
278  ///
279  void reset();
280
281  MCContext &getContext() const { return Context; }
282
283  MCAsmBackend *getBackendPtr() const { return Backend.get(); }
284
285  MCCodeEmitter *getEmitterPtr() const { return Emitter.get(); }
286
287  MCObjectWriter *getWriterPtr() const { return Writer.get(); }
288
289  MCAsmBackend &getBackend() const { return *Backend; }
290
291  MCCodeEmitter &getEmitter() const { return *Emitter; }
292
293  MCObjectWriter &getWriter() const { return *Writer; }
294
295  MCDwarfLineTableParams getDWARFLinetableParams() const { return LTParams; }
296  void setDWARFLinetableParams(MCDwarfLineTableParams P) { LTParams = P; }
297
298  /// Finish - Do final processing and write the object to the output stream.
299  /// \p Writer is used for custom object writer (as the MCJIT does),
300  /// if not specified it is automatically created from backend.
301  void Finish();
302
303  // Layout all section and prepare them for emission.
304  void layout(MCAsmLayout &Layout);
305
306  // FIXME: This does not belong here.
307  bool getSubsectionsViaSymbols() const { return SubsectionsViaSymbols; }
308  void setSubsectionsViaSymbols(bool Value) { SubsectionsViaSymbols = Value; }
309
310  bool isIncrementalLinkerCompatible() const {
311    return IncrementalLinkerCompatible;
312  }
313  void setIncrementalLinkerCompatible(bool Value) {
314    IncrementalLinkerCompatible = Value;
315  }
316
317  bool getRelaxAll() const { return RelaxAll; }
318  void setRelaxAll(bool Value) { RelaxAll = Value; }
319
320  bool isBundlingEnabled() const { return BundleAlignSize != 0; }
321
322  unsigned getBundleAlignSize() const { return BundleAlignSize; }
323
324  void setBundleAlignSize(unsigned Size) {
325    assert((Size == 0 || !(Size & (Size - 1))) &&
326           "Expect a power-of-two bundle align size");
327    BundleAlignSize = Size;
328  }
329
330  /// \name Section List Access
331  /// @{
332
333  iterator begin() { return Sections.begin(); }
334  const_iterator begin() const { return Sections.begin(); }
335
336  iterator end() { return Sections.end(); }
337  const_iterator end() const { return Sections.end(); }
338
339  size_t size() const { return Sections.size(); }
340
341  /// @}
342  /// \name Symbol List Access
343  /// @{
344  symbol_iterator symbol_begin() { return Symbols.begin(); }
345  const_symbol_iterator symbol_begin() const { return Symbols.begin(); }
346
347  symbol_iterator symbol_end() { return Symbols.end(); }
348  const_symbol_iterator symbol_end() const { return Symbols.end(); }
349
350  symbol_range symbols() { return make_range(symbol_begin(), symbol_end()); }
351  const_symbol_range symbols() const {
352    return make_range(symbol_begin(), symbol_end());
353  }
354
355  size_t symbol_size() const { return Symbols.size(); }
356
357  /// @}
358  /// \name Indirect Symbol List Access
359  /// @{
360
361  // FIXME: This is a total hack, this should not be here. Once things are
362  // factored so that the streamer has direct access to the .o writer, it can
363  // disappear.
364  std::vector<IndirectSymbolData> &getIndirectSymbols() {
365    return IndirectSymbols;
366  }
367
368  indirect_symbol_iterator indirect_symbol_begin() {
369    return IndirectSymbols.begin();
370  }
371  const_indirect_symbol_iterator indirect_symbol_begin() const {
372    return IndirectSymbols.begin();
373  }
374
375  indirect_symbol_iterator indirect_symbol_end() {
376    return IndirectSymbols.end();
377  }
378  const_indirect_symbol_iterator indirect_symbol_end() const {
379    return IndirectSymbols.end();
380  }
381
382  size_t indirect_symbol_size() const { return IndirectSymbols.size(); }
383
384  /// @}
385  /// \name Linker Option List Access
386  /// @{
387
388  std::vector<std::vector<std::string>> &getLinkerOptions() {
389    return LinkerOptions;
390  }
391
392  /// @}
393  /// \name Data Region List Access
394  /// @{
395
396  // FIXME: This is a total hack, this should not be here. Once things are
397  // factored so that the streamer has direct access to the .o writer, it can
398  // disappear.
399  std::vector<DataRegionData> &getDataRegions() { return DataRegions; }
400
401  data_region_iterator data_region_begin() { return DataRegions.begin(); }
402  const_data_region_iterator data_region_begin() const {
403    return DataRegions.begin();
404  }
405
406  data_region_iterator data_region_end() { return DataRegions.end(); }
407  const_data_region_iterator data_region_end() const {
408    return DataRegions.end();
409  }
410
411  size_t data_region_size() const { return DataRegions.size(); }
412
413  /// @}
414  /// \name Data Region List Access
415  /// @{
416
417  // FIXME: This is a total hack, this should not be here. Once things are
418  // factored so that the streamer has direct access to the .o writer, it can
419  // disappear.
420  MCLOHContainer &getLOHContainer() { return LOHContainer; }
421  const MCLOHContainer &getLOHContainer() const {
422    return const_cast<MCAssembler *>(this)->getLOHContainer();
423  }
424
425  struct CGProfileEntry {
426    const MCSymbolRefExpr *From;
427    const MCSymbolRefExpr *To;
428    uint64_t Count;
429  };
430  std::vector<CGProfileEntry> CGProfile;
431  /// @}
432  /// \name Backend Data Access
433  /// @{
434
435  bool registerSection(MCSection &Section);
436
437  void registerSymbol(const MCSymbol &Symbol, bool *Created = nullptr);
438
439  ArrayRef<std::string> getFileNames() { return FileNames; }
440
441  void addFileName(StringRef FileName) {
442    if (!is_contained(FileNames, FileName))
443      FileNames.push_back(FileName);
444  }
445
446  /// Write the necessary bundle padding to \p OS.
447  /// Expects a fragment \p F containing instructions and its size \p FSize.
448  void writeFragmentPadding(raw_ostream &OS, const MCEncodedFragment &F,
449                            uint64_t FSize) const;
450
451  /// @}
452
453  void dump() const;
454};
455
456/// Compute the amount of padding required before the fragment \p F to
457/// obey bundling restrictions, where \p FOffset is the fragment's offset in
458/// its section and \p FSize is the fragment's size.
459uint64_t computeBundlePadding(const MCAssembler &Assembler,
460                              const MCEncodedFragment *F, uint64_t FOffset,
461                              uint64_t FSize);
462
463} // end namespace llvm
464
465#endif // LLVM_MC_MCASSEMBLER_H
466