src/ld/ObjectFile.h

   1 /* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
   2  *
   3  * Copyright (c) 2005-2007 Apple Inc. All rights reserved.
   4  *
   5  * @APPLE_LICENSE_HEADER_START@
   6  *
   7  * This file contains Original Code and/or Modifications of Original Code
   8  * as defined in and that are subject to the Apple Public Source License
   9  * Version 2.0 (the 'License'). You may not use this file except in
  10  * compliance with the License. Please obtain a copy of the License at
  11  * http://www.opensource.apple.com/apsl/ and read it before using this
  12  * file.
  13  *
  14  * The Original Code and all software distributed under the License are
  15  * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  16  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  17  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  18  * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
  19  * Please see the License for the specific language governing rights and
  20  * limitations under the License.
  21  *
  22  * @APPLE_LICENSE_HEADER_END@
  23  */
  24
  25
  26 #ifndef __OBJECTFILE__
  27 #define __OBJECTFILE__
  28
  29 #include <stdint.h>
  30 #include <vector>
  31 #include <map>
  32 #include <set>
  33
  34
  35
  36 //
  37 // These classes represent the abstract Atoms and References that are the basis of the linker.
  38 // An Atom and a Reference correspond to a Node and Edge in graph theory.
  39 //
  40 // A Reader is a class which parses an object file and presents it as Atoms and References.
  41 // All linking operations are done on Atoms and References.  This makes the linker file
  42 // format independent.
  43 //
  44 // A Writer takes a vector of Atoms with all References resolved and produces an executable file.
  45 //
  46 //
  47
  48
  49
  50 namespace ObjectFile {
  51
  52
  53 struct LineInfo
  54 {
  55         uint32_t        atomOffset;
  56         const char* fileName;
  57         uint32_t        lineNumber;
  58 };
  59
  60
  61 class ReaderOptions
  62 {
  63 public:
  64                                                 ReaderOptions() : fFullyLoadArchives(false), fLoadAllObjcObjectsFromArchives(false), fFlatNamespace(false),
  65                                                                                 fLinkingMainExecutable(false),
  66                                                                                 fForFinalLinkedImage(false), fNoEHLabels(false), fForStatic(false), fForDyld(false), fMakeTentativeDefinitionsReal(false),
  67                                                                                 fWhyLoad(false), fRootSafe(false), fSetuidSafe(false),fDebugInfoStripping(kDebugInfoFull),
  68                                                                                 fImplicitlyLinkPublicDylibs(true),
  69                                                                                 fAddCompactUnwindEncoding(true),
  70                                                                                 fWarnCompactUnwind(false),
  71                                                                                 fRemoveDwarfUnwindIfCompactExists(false),
  72                                                                                 fMakeCompressedDyldInfo(false),
  73                                                                                 fAutoOrderInitializers(true),
  74                                                                                 fLogObjectFiles(false), fLogAllFiles(false),
  75                                                                                 fTraceDylibs(false), fTraceIndirectDylibs(false), fTraceArchives(false),
  76                                                                                 fTraceOutputFile(NULL), fMacVersionMin(kMinMacVersionUnset), fIPhoneVersionMin(kMinIPhoneVersionUnset) {}
  77         enum DebugInfoStripping { kDebugInfoNone, kDebugInfoMinimal, kDebugInfoFull };
  78         enum MacVersionMin { kMinMacVersionUnset, k10_1, k10_2, k10_3, k10_4, k10_5, k10_6 };
  79         enum IPhoneVersionMin { kMinIPhoneVersionUnset, k2_0, k2_1, k2_2, k3_0 };
  80
  81         struct AliasPair {
  82                 const char*                     realName;
  83                 const char*                     alias;
  84         };
  85
  86         bool                                    fFullyLoadArchives;
  87         bool                                    fLoadAllObjcObjectsFromArchives;
  88         bool                                    fFlatNamespace;
  89         bool                                    fLinkingMainExecutable;
  90         bool                                    fForFinalLinkedImage;
  91         bool                                    fNoEHLabels;
  92         bool                                    fForStatic;
  93         bool                                    fForDyld;
  94         bool                                    fMakeTentativeDefinitionsReal;
  95         bool                                    fWhyLoad;
  96         bool                                    fRootSafe;
  97         bool                                    fSetuidSafe;
  98         DebugInfoStripping              fDebugInfoStripping;
  99         bool                                    fImplicitlyLinkPublicDylibs;
 100         bool                                    fAddCompactUnwindEncoding;
 101         bool                                    fWarnCompactUnwind;
 102         bool                                    fRemoveDwarfUnwindIfCompactExists;
 103         bool                                    fMakeCompressedDyldInfo;
 104         bool                                    fAutoOrderInitializers;
 105         bool                                    fLogObjectFiles;
 106         bool                                    fLogAllFiles;
 107         bool                                    fTraceDylibs;
 108         bool                                    fTraceIndirectDylibs;
 109         bool                                    fTraceArchives;
 110         const char*                             fTraceOutputFile;
 111         MacVersionMin                   fMacVersionMin;
 112         IPhoneVersionMin                fIPhoneVersionMin;
 113         std::vector<AliasPair>  fAliases;
 114 };
 115
 116
 117 class Reader
 118 {
 119 public:
 120         enum DebugInfoKind { kDebugInfoNone=0, kDebugInfoStabs=1, kDebugInfoDwarf=2, kDebugInfoStabsUUID=3 };
 121         struct Stab
 122         {
 123                 class Atom*     atom;
 124                 uint8_t         type;
 125                 uint8_t         other;
 126                 uint16_t        desc;
 127                 uint32_t        value;
 128                 const char* string;
 129         };
 130         enum ObjcConstraint { kObjcNone,  kObjcRetainRelease,  kObjcRetainReleaseOrGC,  kObjcGC };
 131         enum CpuConstraint  { kCpuAny = 0 };
 132
 133         class DylibHander
 134         {
 135         public:
 136                 virtual                         ~DylibHander()  {}
 137                 virtual Reader*         findDylib(const char* installPath, const char* fromPath) = 0;
 138         };
 139
 140
 141         static Reader* createReader(const char* path, const ReaderOptions& options);
 142
 143         virtual const char*                                     getPath() = 0;
 144         virtual time_t                                          getModificationTime() = 0;
 145         virtual DebugInfoKind                           getDebugInfoKind() = 0;
 146         virtual std::vector<class Atom*>&       getAtoms() = 0;
 147         virtual std::vector<class Atom*>*       getJustInTimeAtomsFor(const char* name) = 0;
 148         virtual std::vector<Stab>*                      getStabs() = 0;
 149         virtual ObjcConstraint                          getObjCConstraint()                     { return kObjcNone; }
 150         virtual uint32_t                                        updateCpuConstraint(uint32_t current) { return current; }
 151         virtual bool                                            objcReplacementClasses()        { return false; }
 152
 153         // For relocatable object files only
 154         virtual bool                                            canScatterAtoms()                       { return true; }
 155         virtual bool                                            optimize(const std::vector<ObjectFile::Atom*>&, std::vector<ObjectFile::Atom*>&,
 156                                                                                                         std::vector<const char*>&, const std::set<ObjectFile::Atom*>&,
 157                                                                                                         std::vector<ObjectFile::Atom*>&,
 158                                                                                                         uint32_t, ObjectFile::Reader* writer,
 159                                                                                                         ObjectFile::Atom* entryPointAtom,
 160                                                                                                         const std::vector<const char*>& llvmOptions,
 161                                                                                                         bool allGlobalsAReDeadStripRoots, int okind,
 162                                                                                                         bool verbose, bool saveTemps, const char* outputFilePath,
 163                                                                                                         bool pie, bool allowTextRelocs) { return false; }
 164         virtual bool                                            hasLongBranchStubs()            { return false; }
 165
 166         // For Dynamic Libraries only
 167         virtual const char*                                     getInstallPath()                        { return NULL; }
 168         virtual uint32_t                                        getTimestamp()                          { return 0; }
 169         virtual uint32_t                                        getCurrentVersion()                     { return 0; }
 170         virtual uint32_t                                        getCompatibilityVersion()       { return 0; }
 171         virtual void                                            processIndirectLibraries(DylibHander* handler)  { }
 172         virtual void                                            setExplicitlyLinked()           { }
 173         virtual bool                                            explicitlyLinked()                      { return false; }
 174         virtual bool                                            implicitlyLinked()                      { return false; }
 175         virtual bool                                            providedExportAtom()            { return false; }
 176         virtual const char*                                     parentUmbrella()                        { return NULL; }
 177         virtual std::vector<const char*>*       getAllowableClients()           { return NULL; }
 178         virtual bool                                            hasWeakExternals()                      { return false; }
 179         virtual bool                                            deadStrippable()                        { return false; }
 180         virtual bool                                            isLazyLoadedDylib()                     { return false; }
 181
 182 protected:
 183                                                                                 Reader() {}
 184         virtual                                                         ~Reader() {}
 185 };
 186
 187 class Segment
 188 {
 189 public:
 190         virtual const char*                     getName() const  = 0;
 191         virtual bool                            isContentReadable() const = 0;
 192         virtual bool                            isContentWritable() const = 0;
 193         virtual bool                            isContentExecutable() const = 0;
 194
 195         uint64_t                                        getBaseAddress() const { return fBaseAddress; }
 196         void                                            setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 197         virtual bool                            hasFixedAddress() const { return false; }
 198
 199 protected:
 200                                                                 Segment() : fBaseAddress(0) {}
 201         virtual                                         ~Segment() {}
 202         uint64_t                                        fBaseAddress;
 203 };
 204
 205 class Reference;
 206
 207 class Section
 208 {
 209 public:
 210         unsigned int    getIndex() { return fIndex; }
 211         uint64_t                getBaseAddress() { return fBaseAddress; }
 212         void                    setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 213         void*                   fOther;
 214
 215 protected:
 216                                         Section() : fOther(NULL), fBaseAddress(0), fIndex(0)  {}
 217         uint64_t                fBaseAddress;
 218         unsigned int    fIndex;
 219 };
 220
 221
 222 struct Alignment
 223 {
 224                                 Alignment(int p2, int m=0) : powerOf2(p2), modulus(m) {}
 225         uint8_t         trailingZeros() const { return (modulus==0) ? powerOf2 : __builtin_ctz(modulus); }
 226         uint16_t        powerOf2;
 227         uint16_t        modulus;
 228 };
 229
 230 struct UnwindInfo
 231 {
 232         uint32_t        startOffset;
 233         uint32_t        unwindInfo;
 234
 235         typedef UnwindInfo* iterator;
 236
 237 };
 238
 239
 240 //
 241 // An atom is the fundamental unit of linking.  A C function or global variable is an atom.
 242 // An atom has content and some attributes. The content of a function atom is the instructions
 243 // that implement the function.  The content of a global variable atom is its initial bits.
 244 //
 245 // Name:
 246 // The name of an atom is the label name generated by the compiler.  A C compiler names foo()
 247 // as _foo.  A C++ compiler names foo() as __Z3foov.
 248 // The name refers to the first byte of the content.  An atom cannot have multiple entry points.
 249 // Such code is modeled as multiple atoms, each having a "follow on" reference to the next.
 250 // A "follow on" reference is a contraint to the linker to the atoms must be laid out contiguously.
 251 //
 252 // Scope:
 253 // An atom is in one of three scopes: translation-unit, linkage-unit, or global.  These correspond
 254 // to the C visibility of static, hidden, default.
 255 //
 256 // DefinitionKind:
 257 // An atom is one of five defintion kinds:
 258 //      regular                 Most atoms.
 259 //      weak                    C++ compiler makes some functions weak if there might be multiple copies
 260 //                                      that the linker needs to coalesce.
 261 //      tentative               A straggler from ancient C when the extern did not exist. "int foo;" is ambiguous.
 262 //                                      It could be a prototype or it could be a definition.
 263 //      external                This is a "proxy" atom produced by a dylib reader.  It has no content.  It exists
 264 //                                      so that all References can be resolved.
 265 //      external-weak   Same as external, but the definition in the dylib is weak.
 266 //
 267 // SymbolTableInclusion:
 268 // An atom may or may not be in the symbol table in an object file.
 269 //  in                          Most atoms for functions or global data
 270 //      not-in                  Anonymous atoms such literal c-strings, or other compiler generated data
 271 //      in-never-strip  Atom whose name the strip tool should never remove (e.g. REFERENCED_DYNAMICALLY in mach-o)
 272 //
 273 // Ordinal:
 274 // When a reader is created it is given a base ordinal number.  All atoms created by the reader
 275 // should return a contiguous range of ordinal values that start at the base ordinal.  The ordinal
 276 // values are used by the linker to sort the atom graph when producing the output file.
 277 //
 278 class Atom
 279 {
 280 public:
 281         enum Scope { scopeTranslationUnit, scopeLinkageUnit, scopeGlobal };
 282         enum DefinitionKind { kRegularDefinition, kWeakDefinition, kTentativeDefinition, kExternalDefinition, kExternalWeakDefinition, kAbsoluteSymbol };
 283         enum ContentType { kUnclassifiedType, kCStringType, kCFIType, kLSDAType };
 284         enum SymbolTableInclusion { kSymbolTableNotIn, kSymbolTableIn, kSymbolTableInAndNeverStrip, kSymbolTableInAsAbsolute };
 285
 286         virtual Reader*                                                 getFile() const = 0;
 287         virtual bool                                                    getTranslationUnitSource(const char** dir, const char** name) const = 0;
 288         virtual const char*                                             getName() const = 0;
 289         virtual const char*                                             getDisplayName() const = 0;
 290         virtual Scope                                                   getScope() const = 0;
 291         virtual DefinitionKind                                  getDefinitionKind() const = 0;
 292         virtual ContentType                                             getContentType() const { return kUnclassifiedType; }
 293         virtual SymbolTableInclusion                    getSymbolTableInclusion() const = 0;
 294         virtual bool                                                    dontDeadStrip() const = 0;
 295         virtual bool                                                    isZeroFill() const = 0;
 296         virtual bool                                                    isThumb() const = 0;
 297         virtual uint64_t                                                getSize() const = 0;
 298         virtual std::vector<ObjectFile::Reference*>&  getReferences() const = 0;
 299         virtual bool                                                    mustRemainInSection() const = 0;
 300         virtual const char*                                             getSectionName() const = 0;
 301         virtual Segment&                                                getSegment() const = 0;
 302         virtual Atom&                                                   getFollowOnAtom() const = 0;
 303         virtual uint32_t                                                getOrdinal() const = 0;
 304         virtual std::vector<LineInfo>*                  getLineInfo() const = 0;
 305         virtual Alignment                                               getAlignment() const = 0;
 306         virtual void                                                    copyRawContent(uint8_t buffer[]) const = 0;
 307         virtual void                                                    setScope(Scope) = 0;
 308         virtual UnwindInfo::iterator                    beginUnwind() { return NULL; }
 309         virtual UnwindInfo::iterator                    endUnwind() { return NULL; }
 310         virtual Reference*                                              getLSDA() { return NULL; }
 311         virtual Reference*                                              getFDE() { return NULL; }
 312         virtual Atom*                                                   getPersonalityPointer() { return NULL; }
 313
 314                         uint64_t                                                getSectionOffset() const        { return fSectionOffset; }
 315                         uint64_t                                                getAddress() const      { return fSection->getBaseAddress() + fSectionOffset; }
 316                         class Section*                                  getSection() const { return fSection; }
 317
 318         virtual void                                                    setSectionOffset(uint64_t offset) { fSectionOffset = offset; }
 319         virtual void                                                    setSection(class Section* sect) { fSection = sect; }
 320
 321 protected:
 322                                                                                         Atom() :  fSectionOffset(0), fSection(NULL) {}
 323                 virtual                                                         ~Atom() {}
 324
 325                 uint64_t                                                        fSectionOffset;
 326                 class Section*                                          fSection;
 327 };
 328
 329
 330 //
 331 // A Reference is a directed edge to another Atom.  When an instruction in
 332 // the content of an Atom refers to another Atom, that is represented by a
 333 // Reference.
 334 //
 335 // There are two kinds of references: direct and by-name.  With a direct Reference,
 336 // the target is bound by the Reader that created it.  For instance a reference to a
 337 // static would produce a direct reference.  A by-name reference requires the linker
 338 // to find the target Atom with the required name in order to be bound.
 339 //
 340 // For a link to succeed all References must be bound.
 341 //
 342 // A Reference has an optional "from" target.  This is used when the content to fix-up
 343 // is the difference of two Atom address.  For instance, if a pointer sized data Atom
 344 // is to contain A - B, then the Atom would have on Reference with a target of "A" and
 345 // a from-target of "B".
 346 //
 347 // A Reference also has a fix-up-offset.  This is the offset into the content of the
 348 // Atom holding the reference where the fix-up (relocation) will be applied.
 349 //
 350 //
 351 //
 352 class Reference
 353 {
 354 public:
 355         enum TargetBinding { kUnboundByName, kBoundDirectly, kBoundByName, kDontBind };
 356
 357         virtual TargetBinding   getTargetBinding() const = 0;
 358         virtual TargetBinding   getFromTargetBinding() const = 0;
 359         virtual uint8_t                 getKind() const = 0;
 360         virtual uint64_t                getFixUpOffset() const = 0;
 361         virtual const char*             getTargetName() const = 0;
 362         virtual Atom&                   getTarget() const = 0;
 363         virtual uint64_t                getTargetOffset() const = 0;
 364         virtual Atom&                   getFromTarget() const = 0;
 365         virtual const char*             getFromTargetName() const = 0;
 366         virtual uint64_t                getFromTargetOffset() const = 0;
 367
 368         virtual void                    setTarget(Atom&, uint64_t offset) = 0;
 369         virtual void                    setFromTarget(Atom&) = 0;
 370         virtual const char*             getDescription() const = 0;
 371         virtual bool                    isBranch() const  { return false; }
 372
 373 protected:
 374                                                         Reference() {}
 375         virtual                                 ~Reference() {}
 376 };
 377
 378
 379 };      // namespace ObjectFile
 380
 381
 382 #endif // __OBJECTFILE__