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 namespace ObjectFile {
  49
  50
  51 struct LineInfo
  52 {
  53         uint32_t        atomOffset;
  54         const char* fileName;
  55         uint32_t        lineNumber;
  56 };
  57
  58
  59 class ReaderOptions
  60 {
  61 public:
  62                                                 ReaderOptions() : fFullyLoadArchives(false), fLoadAllObjcObjectsFromArchives(false), fFlatNamespace(false), fLinkingMainExecutable(false),
  63                                                                                 fForFinalLinkedImage(false), fForStatic(false), fForDyld(false), fMakeTentativeDefinitionsReal(false),
  64                                                                                 fWhyLoad(false), fRootSafe(false), fSetuidSafe(false),fDebugInfoStripping(kDebugInfoFull),
  65                                                                                 fLogObjectFiles(false), fLogAllFiles(false),
  66                                                                                 fTraceDylibs(false), fTraceIndirectDylibs(false), fTraceArchives(false),
  67                                                                                 fTraceOutputFile(NULL), fVersionMin(kMinUnset) {}
  68         enum DebugInfoStripping { kDebugInfoNone, kDebugInfoMinimal, kDebugInfoFull };
  69         enum VersionMin { kMinUnset, k10_1, k10_2, k10_3, k10_4, k10_5 };
  70
  71         struct AliasPair {
  72                 const char*                     realName;
  73                 const char*                     alias;
  74         };
  75
  76         bool                                    fFullyLoadArchives;
  77         bool                                    fLoadAllObjcObjectsFromArchives;
  78         bool                                    fFlatNamespace;
  79         bool                                    fLinkingMainExecutable;
  80         bool                                    fForFinalLinkedImage;
  81         bool                                    fForStatic;
  82         bool                                    fForDyld;
  83         bool                                    fMakeTentativeDefinitionsReal;
  84         bool                                    fWhyLoad;
  85         bool                                    fRootSafe;
  86         bool                                    fSetuidSafe;
  87         DebugInfoStripping              fDebugInfoStripping;
  88         bool                                    fLogObjectFiles;
  89         bool                                    fLogAllFiles;
  90         bool                                    fTraceDylibs;
  91         bool                                    fTraceIndirectDylibs;
  92         bool                                    fTraceArchives;
  93         const char*                             fTraceOutputFile;
  94         VersionMin                              fVersionMin;
  95         std::vector<AliasPair>  fAliases;
  96 };
  97
  98
  99 class Reader
 100 {
 101 public:
 102         enum DebugInfoKind { kDebugInfoNone=0, kDebugInfoStabs=1, kDebugInfoDwarf=2, kDebugInfoStabsUUID=3 };
 103         struct Stab
 104         {
 105                 class Atom*     atom;
 106                 uint8_t         type;
 107                 uint8_t         other;
 108                 uint16_t        desc;
 109                 uint32_t        value;
 110                 const char* string;
 111         };
 112         enum ObjcConstraint { kObjcNone,  kObjcRetainRelease,  kObjcRetainReleaseOrGC,  kObjcGC };
 113         enum CpuConstraint  { kCpuAny,  kCpuG3,  kCpuG4,  kCpuG5  };
 114
 115         class DylibHander
 116         {
 117         public:
 118                 virtual                         ~DylibHander()  {}
 119                 virtual Reader*         findDylib(const char* installPath, const char* fromPath) = 0;
 120         };
 121
 122
 123         static Reader* createReader(const char* path, const ReaderOptions& options);
 124
 125         virtual const char*                                     getPath() = 0;
 126         virtual time_t                                          getModificationTime() = 0;
 127         virtual DebugInfoKind                           getDebugInfoKind() = 0;
 128         virtual std::vector<class Atom*>&       getAtoms() = 0;
 129         virtual std::vector<class Atom*>*       getJustInTimeAtomsFor(const char* name) = 0;
 130         virtual std::vector<Stab>*                      getStabs() = 0;
 131         virtual ObjcConstraint                          getObjCConstraint()                     { return kObjcNone; }
 132         virtual CpuConstraint                           getCpuConstraint()                      { return kCpuAny; }
 133         virtual bool                                            objcReplacementClasses()        { return false; }
 134
 135         // For relocatable object files only
 136         virtual bool                                            canScatterAtoms()                       { return true; }
 137         virtual void                                            optimize(std::vector<ObjectFile::Atom*>&, std::vector<ObjectFile::Atom*>&, uint32_t)            { }
 138
 139         // For Dynamic Libraries only
 140         virtual const char*                                     getInstallPath()                        { return NULL; }
 141         virtual uint32_t                                        getTimestamp()                          { return 0; }
 142         virtual uint32_t                                        getCurrentVersion()                     { return 0; }
 143         virtual uint32_t                                        getCompatibilityVersion()       { return 0; }
 144         virtual void                                            processIndirectLibraries(DylibHander* handler)  { }
 145         virtual void                                            setExplicitlyLinked()           { }
 146         virtual bool                                            explicitlyLinked()                      { return false; }
 147         virtual bool                                            implicitlyLinked()                      { return false; }
 148         virtual bool                                            providedExportAtom()            { return false; }
 149         virtual const char*                                     parentUmbrella()                        { return NULL; }
 150         virtual std::vector<const char*>*       getAllowableClients()           { return NULL; }
 151
 152
 153 protected:
 154                                                                                 Reader() {}
 155         virtual                                                         ~Reader() {}
 156 };
 157
 158 class Segment
 159 {
 160 public:
 161         virtual const char*                     getName() const  = 0;
 162         virtual bool                            isContentReadable() const = 0;
 163         virtual bool                            isContentWritable() const = 0;
 164         virtual bool                            isContentExecutable() const = 0;
 165
 166         uint64_t                                        getBaseAddress() const { return fBaseAddress; }
 167         void                                            setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 168         virtual bool                            hasFixedAddress() const { return false; }
 169
 170 protected:
 171                                                                 Segment() : fBaseAddress(0) {}
 172         virtual                                         ~Segment() {}
 173         uint64_t                                        fBaseAddress;
 174 };
 175
 176 class Reference;
 177
 178 class Section
 179 {
 180 public:
 181         unsigned int    getIndex() { return fIndex; }
 182         uint64_t                getBaseAddress() { return fBaseAddress; }
 183         void                    setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 184         void*                   fOther;
 185
 186 protected:
 187                                         Section() : fOther(NULL), fBaseAddress(0), fIndex(0)  {}
 188         uint64_t                fBaseAddress;
 189         unsigned int    fIndex;
 190 };
 191
 192
 193 struct Alignment
 194 {
 195                                 Alignment(int p2, int m=0) : powerOf2(p2), modulus(m) {}
 196         uint8_t         trailingZeros() const { return (modulus==0) ? powerOf2 : __builtin_ctz(modulus); }
 197         uint16_t        powerOf2;
 198         uint16_t        modulus;
 199 };
 200
 201 //
 202 // An atom is the fundamental unit of linking.  A C function or global variable is an atom.
 203 // An atom has content and some attributes. The content of a function atom is the instructions
 204 // that implement the function.  The content of a global variable atom is its initial bits.
 205 //
 206 // Name:
 207 // The name of an atom is the label name generated by the compiler.  A C compiler names foo()
 208 // as _foo.  A C++ compiler names foo() as __Z3foov.
 209 // The name refers to the first byte of the content.  An atom cannot have multiple entry points.
 210 // Such code is modeled as multiple atoms, each having a "follow on" reference to the next.
 211 // A "follow on" reference is a contraint to the linker to the atoms must be laid out contiguously.
 212 //
 213 // Scope:
 214 // An atom is in one of three scopes: translation-unit, linkage-unit, or global.  These correspond
 215 // to the C visibility of static, hidden, default.
 216 //
 217 // DefinitionKind:
 218 // An atom is one of five defintion kinds:
 219 //      regular                 Most atoms.
 220 //      weak                    C++ compiler makes some functions weak if there might be multiple copies
 221 //                                      that the linker needs to coalesce.
 222 //      tentative               A straggler from ancient C when the extern did not exist. "int foo;" is ambiguous.
 223 //                                      It could be a prototype or it could be a definition.
 224 //      external                This is a "proxy" atom produced by a dylib reader.  It has no content.  It exists
 225 //                                      so that all References can be resolved.
 226 //      external-weak   Same as external, but the definition in the dylib is weak.
 227 //
 228 // SymbolTableInclusion:
 229 // An atom may or may not be in the symbol table in an object file.
 230 //  in                          Most atoms for functions or global data
 231 //      not-in                  Anonymous atoms such literal c-strings, or other compiler generated data
 232 //      in-never-strip  Atom whose name the strip tool should never remove (e.g. REFERENCED_DYNAMICALLY in mach-o)
 233 //
 234 // Ordinal:
 235 // When a reader is created it is given a base ordinal number.  All atoms created by the reader
 236 // should return a contiguous range of ordinal values that start at the base ordinal.  The ordinal
 237 // values are used by the linker to sort the atom graph when producing the output file.
 238 //
 239 class Atom
 240 {
 241 public:
 242         enum Scope { scopeTranslationUnit, scopeLinkageUnit, scopeGlobal };
 243         enum DefinitionKind { kRegularDefinition, kWeakDefinition, kTentativeDefinition, kExternalDefinition, kExternalWeakDefinition, kAbsoluteSymbol };
 244         enum SymbolTableInclusion { kSymbolTableNotIn, kSymbolTableIn, kSymbolTableInAndNeverStrip, kSymbolTableInAsAbsolute };
 245
 246         virtual Reader*                                                 getFile() const = 0;
 247         virtual bool                                                    getTranslationUnitSource(const char** dir, const char** name) const = 0;
 248         virtual const char*                                             getName() const = 0;
 249         virtual const char*                                             getDisplayName() const = 0;
 250         virtual Scope                                                   getScope() const = 0;
 251         virtual DefinitionKind                                  getDefinitionKind() const = 0;
 252         virtual SymbolTableInclusion                    getSymbolTableInclusion() const = 0;
 253         virtual bool                                                    dontDeadStrip() const = 0;
 254         virtual bool                                                    isZeroFill() const = 0;
 255         virtual uint64_t                                                getSize() const = 0;
 256         virtual std::vector<ObjectFile::Reference*>&  getReferences() const = 0;
 257         virtual bool                                                    mustRemainInSection() const = 0;
 258         virtual const char*                                             getSectionName() const = 0;
 259         virtual Segment&                                                getSegment() const = 0;
 260         virtual Atom&                                                   getFollowOnAtom() const = 0;
 261         virtual uint32_t                                                getOrdinal() const = 0;
 262         virtual std::vector<LineInfo>*                  getLineInfo() const = 0;
 263         virtual Alignment                                               getAlignment() const = 0;
 264         virtual void                                                    copyRawContent(uint8_t buffer[]) const = 0;
 265         virtual void                                                    setScope(Scope) = 0;
 266
 267
 268                         uint64_t                                                getSectionOffset() const        { return fSectionOffset; }
 269                         uint64_t                                                getAddress() const      { return fSection->getBaseAddress() + fSectionOffset; }
 270                         class Section*                                  getSection() const { return fSection; }
 271
 272         virtual void                                                    setSectionOffset(uint64_t offset) { fSectionOffset = offset; }
 273         virtual void                                                    setSection(class Section* sect) { fSection = sect; }
 274
 275 protected:
 276                                                                                         Atom() :  fSectionOffset(0), fSection(NULL) {}
 277                 virtual                                                         ~Atom() {}
 278
 279                 uint64_t                                                        fSectionOffset;
 280                 class Section*                                          fSection;
 281 };
 282
 283
 284 //
 285 // A Reference is a directed edge to another Atom.  When an instruction in
 286 // the content of an Atom refers to another Atom, that is represented by a
 287 // Reference.
 288 //
 289 // There are two kinds of references: direct and by-name.  With a direct Reference,
 290 // the target is bound by the Reader that created it.  For instance a reference to a
 291 // static would produce a direct reference.  A by-name reference requires the linker
 292 // to find the target Atom with the required name in order to be bound.
 293 //
 294 // For a link to succeed all References must be bound.
 295 //
 296 // A Reference has an optional "from" target.  This is used when the content to fix-up
 297 // is the difference of two Atom address.  For instance, if a pointer sized data Atom
 298 // is to contain A - B, then the Atom would have on Reference with a target of "A" and
 299 // a from-target of "B".
 300 //
 301 // A Reference also has a fix-up-offset.  This is the offset into the content of the
 302 // Atom holding the reference where the fix-up (relocation) will be applied.
 303 //
 304 //
 305 //
 306 class Reference
 307 {
 308 public:
 309         enum TargetBinding { kUnboundByName, kBoundDirectly, kBoundByName, kDontBind };
 310
 311         virtual TargetBinding   getTargetBinding() const = 0;
 312         virtual TargetBinding   getFromTargetBinding() const = 0;
 313         virtual uint8_t                 getKind() const = 0;
 314         virtual uint64_t                getFixUpOffset() const = 0;
 315         virtual const char*             getTargetName() const = 0;
 316         virtual Atom&                   getTarget() const = 0;
 317         virtual uint64_t                getTargetOffset() const = 0;
 318         virtual Atom&                   getFromTarget() const = 0;
 319         virtual const char*             getFromTargetName() const = 0;
 320         virtual uint64_t                getFromTargetOffset() const = 0;
 321
 322         virtual void                    setTarget(Atom&, uint64_t offset) = 0;
 323         virtual void                    setFromTarget(Atom&) = 0;
 324         virtual const char*             getDescription() const = 0;
 325
 326 protected:
 327                                                         Reference() {}
 328         virtual                                 ~Reference() {}
 329 };
 330
 331
 332 };      // namespace ObjectFile
 333
 334
 335 #endif // __OBJECTFILE__