src/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
  33
  34
  35 //
  36 // These classes represent the abstract Atoms and References that are the basis of the linker.
  37 // An Atom and a Reference correspond to a Node and Edge in graph theory.
  38 //
  39 // A Reader is a class which parses an object file and presents it as Atoms and References.
  40 // All linking operations are done on Atoms and References.  This makes the linker file
  41 // format independent.
  42 //
  43 // A Writer takes a vector of Atoms with all References resolved and produces an executable file.
  44 //
  45 //
  46
  47
  48
  49 namespace ObjectFile {
  50
  51
  52 struct LineInfo
  53 {
  54         uint32_t        atomOffset;
  55         const char* fileName;
  56         uint32_t        lineNumber;
  57 };
  58
  59
  60 class ReaderOptions
  61 {
  62 public:
  63                                                 ReaderOptions() : fFullyLoadArchives(false), fLoadAllObjcObjectsFromArchives(false), fFlatNamespace(false),
  64                                                                                 fLinkingMainExecutable(false), fSlowx86Stubs(false),
  65                                                                                 fForFinalLinkedImage(false), fForStatic(false), fForDyld(false), fMakeTentativeDefinitionsReal(false),
  66                                                                                 fWhyLoad(false), fRootSafe(false), fSetuidSafe(false),fDebugInfoStripping(kDebugInfoFull),
  67                                                                                 fImplicitlyLinkPublicDylibs(true), fLogObjectFiles(false), fLogAllFiles(false),
  68                                                                                 fTraceDylibs(false), fTraceIndirectDylibs(false), fTraceArchives(false),
  69                                                                                 fTraceOutputFile(NULL), fVersionMin(kMinUnset) {}
  70         enum DebugInfoStripping { kDebugInfoNone, kDebugInfoMinimal, kDebugInfoFull };
  71         enum VersionMin { kMinUnset, k10_1, k10_2, k10_3, k10_4, k10_5, k10_6 };
  72
  73         struct AliasPair {
  74                 const char*                     realName;
  75                 const char*                     alias;
  76         };
  77
  78         bool                                    fFullyLoadArchives;
  79         bool                                    fLoadAllObjcObjectsFromArchives;
  80         bool                                    fFlatNamespace;
  81         bool                                    fLinkingMainExecutable;
  82         bool                                    fSlowx86Stubs;
  83         bool                                    fForFinalLinkedImage;
  84         bool                                    fForStatic;
  85         bool                                    fForDyld;
  86         bool                                    fMakeTentativeDefinitionsReal;
  87         bool                                    fWhyLoad;
  88         bool                                    fRootSafe;
  89         bool                                    fSetuidSafe;
  90         DebugInfoStripping              fDebugInfoStripping;
  91         bool                                    fImplicitlyLinkPublicDylibs;
  92         bool                                    fLogObjectFiles;
  93         bool                                    fLogAllFiles;
  94         bool                                    fTraceDylibs;
  95         bool                                    fTraceIndirectDylibs;
  96         bool                                    fTraceArchives;
  97         const char*                             fTraceOutputFile;
  98         VersionMin                              fVersionMin;
  99         std::vector<AliasPair>  fAliases;
 100 };
 101
 102
 103 class Reader
 104 {
 105 public:
 106         enum DebugInfoKind { kDebugInfoNone=0, kDebugInfoStabs=1, kDebugInfoDwarf=2, kDebugInfoStabsUUID=3 };
 107         struct Stab
 108         {
 109                 class Atom*     atom;
 110                 uint8_t         type;
 111                 uint8_t         other;
 112                 uint16_t        desc;
 113                 uint32_t        value;
 114                 const char* string;
 115         };
 116         enum ObjcConstraint { kObjcNone,  kObjcRetainRelease,  kObjcRetainReleaseOrGC,  kObjcGC };
 117         enum CpuConstraint  { kCpuAny = 0 };
 118
 119         class DylibHander
 120         {
 121         public:
 122                 virtual                         ~DylibHander()  {}
 123                 virtual Reader*         findDylib(const char* installPath, const char* fromPath) = 0;
 124         };
 125
 126
 127         static Reader* createReader(const char* path, const ReaderOptions& options);
 128
 129         virtual const char*                                     getPath() = 0;
 130         virtual time_t                                          getModificationTime() = 0;
 131         virtual DebugInfoKind                           getDebugInfoKind() = 0;
 132         virtual std::vector<class Atom*>&       getAtoms() = 0;
 133         virtual std::vector<class Atom*>*       getJustInTimeAtomsFor(const char* name) = 0;
 134         virtual std::vector<Stab>*                      getStabs() = 0;
 135         virtual ObjcConstraint                          getObjCConstraint()                     { return kObjcNone; }
 136         virtual uint32_t                                        updateCpuConstraint(uint32_t current) { return current; }
 137         virtual bool                                            objcReplacementClasses()        { return false; }
 138
 139         // For relocatable object files only
 140         virtual bool                                            canScatterAtoms()                       { return true; }
 141         virtual void                                            optimize(std::vector<ObjectFile::Atom*>&, std::vector<ObjectFile::Atom*>&,
 142                                                                                                         std::vector<const char*>&,  uint32_t, ObjectFile::Reader* writer,
 143                                                                                                         bool allGlobalsAReDeadStripRoots, int okind,
 144                                                                                                         bool verbose, bool saveTemps, const char* outputFilePath,
 145                                                                                                         bool pie, bool allowTextRelocs) { }
 146         virtual bool                                            hasLongBranchStubs()            { return false; }
 147
 148         // For Dynamic Libraries only
 149         virtual const char*                                     getInstallPath()                        { return NULL; }
 150         virtual uint32_t                                        getTimestamp()                          { return 0; }
 151         virtual uint32_t                                        getCurrentVersion()                     { return 0; }
 152         virtual uint32_t                                        getCompatibilityVersion()       { return 0; }
 153         virtual void                                            processIndirectLibraries(DylibHander* handler)  { }
 154         virtual void                                            setExplicitlyLinked()           { }
 155         virtual bool                                            explicitlyLinked()                      { return false; }
 156         virtual bool                                            implicitlyLinked()                      { return false; }
 157         virtual bool                                            providedExportAtom()            { return false; }
 158         virtual const char*                                     parentUmbrella()                        { return NULL; }
 159         virtual std::vector<const char*>*       getAllowableClients()           { return NULL; }
 160         virtual bool                                            hasWeakExternals()                      { return false; }
 161         virtual bool                                            isLazyLoadedDylib()                     { return false; }
 162
 163 protected:
 164                                                                                 Reader() {}
 165         virtual                                                         ~Reader() {}
 166 };
 167
 168 class Segment
 169 {
 170 public:
 171         virtual const char*                     getName() const  = 0;
 172         virtual bool                            isContentReadable() const = 0;
 173         virtual bool                            isContentWritable() const = 0;
 174         virtual bool                            isContentExecutable() const = 0;
 175
 176         uint64_t                                        getBaseAddress() const { return fBaseAddress; }
 177         void                                            setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 178         virtual bool                            hasFixedAddress() const { return false; }
 179
 180 protected:
 181                                                                 Segment() : fBaseAddress(0) {}
 182         virtual                                         ~Segment() {}
 183         uint64_t                                        fBaseAddress;
 184 };
 185
 186 class Reference;
 187
 188 class Section
 189 {
 190 public:
 191         unsigned int    getIndex() { return fIndex; }
 192         uint64_t                getBaseAddress() { return fBaseAddress; }
 193         void                    setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 194         void*                   fOther;
 195
 196 protected:
 197                                         Section() : fOther(NULL), fBaseAddress(0), fIndex(0)  {}
 198         uint64_t                fBaseAddress;
 199         unsigned int    fIndex;
 200 };
 201
 202
 203 struct Alignment
 204 {
 205                                 Alignment(int p2, int m=0) : powerOf2(p2), modulus(m) {}
 206         uint8_t         trailingZeros() const { return (modulus==0) ? powerOf2 : __builtin_ctz(modulus); }
 207         uint16_t        powerOf2;
 208         uint16_t        modulus;
 209 };
 210
 211 //
 212 // An atom is the fundamental unit of linking.  A C function or global variable is an atom.
 213 // An atom has content and some attributes. The content of a function atom is the instructions
 214 // that implement the function.  The content of a global variable atom is its initial bits.
 215 //
 216 // Name:
 217 // The name of an atom is the label name generated by the compiler.  A C compiler names foo()
 218 // as _foo.  A C++ compiler names foo() as __Z3foov.
 219 // The name refers to the first byte of the content.  An atom cannot have multiple entry points.
 220 // Such code is modeled as multiple atoms, each having a "follow on" reference to the next.
 221 // A "follow on" reference is a contraint to the linker to the atoms must be laid out contiguously.
 222 //
 223 // Scope:
 224 // An atom is in one of three scopes: translation-unit, linkage-unit, or global.  These correspond
 225 // to the C visibility of static, hidden, default.
 226 //
 227 // DefinitionKind:
 228 // An atom is one of five defintion kinds:
 229 //      regular                 Most atoms.
 230 //      weak                    C++ compiler makes some functions weak if there might be multiple copies
 231 //                                      that the linker needs to coalesce.
 232 //      tentative               A straggler from ancient C when the extern did not exist. "int foo;" is ambiguous.
 233 //                                      It could be a prototype or it could be a definition.
 234 //      external                This is a "proxy" atom produced by a dylib reader.  It has no content.  It exists
 235 //                                      so that all References can be resolved.
 236 //      external-weak   Same as external, but the definition in the dylib is weak.
 237 //
 238 // SymbolTableInclusion:
 239 // An atom may or may not be in the symbol table in an object file.
 240 //  in                          Most atoms for functions or global data
 241 //      not-in                  Anonymous atoms such literal c-strings, or other compiler generated data
 242 //      in-never-strip  Atom whose name the strip tool should never remove (e.g. REFERENCED_DYNAMICALLY in mach-o)
 243 //
 244 // Ordinal:
 245 // When a reader is created it is given a base ordinal number.  All atoms created by the reader
 246 // should return a contiguous range of ordinal values that start at the base ordinal.  The ordinal
 247 // values are used by the linker to sort the atom graph when producing the output file.
 248 //
 249 class Atom
 250 {
 251 public:
 252         enum Scope { scopeTranslationUnit, scopeLinkageUnit, scopeGlobal };
 253         enum DefinitionKind { kRegularDefinition, kWeakDefinition, kTentativeDefinition, kExternalDefinition, kExternalWeakDefinition, kAbsoluteSymbol };
 254         enum SymbolTableInclusion { kSymbolTableNotIn, kSymbolTableIn, kSymbolTableInAndNeverStrip, kSymbolTableInAsAbsolute };
 255
 256         virtual Reader*                                                 getFile() const = 0;
 257         virtual bool                                                    getTranslationUnitSource(const char** dir, const char** name) const = 0;
 258         virtual const char*                                             getName() const = 0;
 259         virtual const char*                                             getDisplayName() const = 0;
 260         virtual Scope                                                   getScope() const = 0;
 261         virtual DefinitionKind                                  getDefinitionKind() const = 0;
 262         virtual SymbolTableInclusion                    getSymbolTableInclusion() const = 0;
 263         virtual bool                                                    dontDeadStrip() const = 0;
 264         virtual bool                                                    isZeroFill() const = 0;
 265         virtual bool                                                    isThumb() const = 0;
 266         virtual uint64_t                                                getSize() const = 0;
 267         virtual std::vector<ObjectFile::Reference*>&  getReferences() const = 0;
 268         virtual bool                                                    mustRemainInSection() const = 0;
 269         virtual const char*                                             getSectionName() const = 0;
 270         virtual Segment&                                                getSegment() const = 0;
 271         virtual Atom&                                                   getFollowOnAtom() const = 0;
 272         virtual uint32_t                                                getOrdinal() const = 0;
 273         virtual std::vector<LineInfo>*                  getLineInfo() const = 0;
 274         virtual Alignment                                               getAlignment() const = 0;
 275         virtual void                                                    copyRawContent(uint8_t buffer[]) const = 0;
 276         virtual void                                                    setScope(Scope) = 0;
 277
 278
 279                         uint64_t                                                getSectionOffset() const        { return fSectionOffset; }
 280                         uint64_t                                                getAddress() const      { return fSection->getBaseAddress() + fSectionOffset; }
 281                         class Section*                                  getSection() const { return fSection; }
 282
 283         virtual void                                                    setSectionOffset(uint64_t offset) { fSectionOffset = offset; }
 284         virtual void                                                    setSection(class Section* sect) { fSection = sect; }
 285
 286 protected:
 287                                                                                         Atom() :  fSectionOffset(0), fSection(NULL) {}
 288                 virtual                                                         ~Atom() {}
 289
 290                 uint64_t                                                        fSectionOffset;
 291                 class Section*                                          fSection;
 292 };
 293
 294
 295 //
 296 // A Reference is a directed edge to another Atom.  When an instruction in
 297 // the content of an Atom refers to another Atom, that is represented by a
 298 // Reference.
 299 //
 300 // There are two kinds of references: direct and by-name.  With a direct Reference,
 301 // the target is bound by the Reader that created it.  For instance a reference to a
 302 // static would produce a direct reference.  A by-name reference requires the linker
 303 // to find the target Atom with the required name in order to be bound.
 304 //
 305 // For a link to succeed all References must be bound.
 306 //
 307 // A Reference has an optional "from" target.  This is used when the content to fix-up
 308 // is the difference of two Atom address.  For instance, if a pointer sized data Atom
 309 // is to contain A - B, then the Atom would have on Reference with a target of "A" and
 310 // a from-target of "B".
 311 //
 312 // A Reference also has a fix-up-offset.  This is the offset into the content of the
 313 // Atom holding the reference where the fix-up (relocation) will be applied.
 314 //
 315 //
 316 //
 317 class Reference
 318 {
 319 public:
 320         enum TargetBinding { kUnboundByName, kBoundDirectly, kBoundByName, kDontBind };
 321
 322         virtual TargetBinding   getTargetBinding() const = 0;
 323         virtual TargetBinding   getFromTargetBinding() const = 0;
 324         virtual uint8_t                 getKind() const = 0;
 325         virtual uint64_t                getFixUpOffset() const = 0;
 326         virtual const char*             getTargetName() const = 0;
 327         virtual Atom&                   getTarget() const = 0;
 328         virtual uint64_t                getTargetOffset() const = 0;
 329         virtual Atom&                   getFromTarget() const = 0;
 330         virtual const char*             getFromTargetName() const = 0;
 331         virtual uint64_t                getFromTargetOffset() const = 0;
 332
 333         virtual void                    setTarget(Atom&, uint64_t offset) = 0;
 334         virtual void                    setFromTarget(Atom&) = 0;
 335         virtual const char*             getDescription() const = 0;
 336
 337 protected:
 338                                                         Reference() {}
 339         virtual                                 ~Reference() {}
 340 };
 341
 342
 343 };      // namespace ObjectFile
 344
 345
 346 #endif // __OBJECTFILE__