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 #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), fSlowx86Stubs(false),
  66                                                                                 fForFinalLinkedImage(false), fForStatic(false), fForDyld(false), fMakeTentativeDefinitionsReal(false),
  67                                                                                 fWhyLoad(false), fRootSafe(false), fSetuidSafe(false),fDebugInfoStripping(kDebugInfoFull),
  68                                                                                 fImplicitlyLinkPublicDylibs(true), fLogObjectFiles(false), fLogAllFiles(false),
  69                                                                                 fTraceDylibs(false), fTraceIndirectDylibs(false), fTraceArchives(false),
  70                                                                                 fTraceOutputFile(NULL), fVersionMin(kMinUnset) {}
  71         enum DebugInfoStripping { kDebugInfoNone, kDebugInfoMinimal, kDebugInfoFull };
  72         enum VersionMin { kMinUnset, k10_1, k10_2, k10_3, k10_4, k10_5, k10_6 };
  73
  74         struct AliasPair {
  75                 const char*                     realName;
  76                 const char*                     alias;
  77         };
  78
  79         bool                                    fFullyLoadArchives;
  80         bool                                    fLoadAllObjcObjectsFromArchives;
  81         bool                                    fFlatNamespace;
  82         bool                                    fLinkingMainExecutable;
  83         bool                                    fSlowx86Stubs;
  84         bool                                    fForFinalLinkedImage;
  85         bool                                    fForStatic;
  86         bool                                    fForDyld;
  87         bool                                    fMakeTentativeDefinitionsReal;
  88         bool                                    fWhyLoad;
  89         bool                                    fRootSafe;
  90         bool                                    fSetuidSafe;
  91         DebugInfoStripping              fDebugInfoStripping;
  92         bool                                    fImplicitlyLinkPublicDylibs;
  93         bool                                    fLogObjectFiles;
  94         bool                                    fLogAllFiles;
  95         bool                                    fTraceDylibs;
  96         bool                                    fTraceIndirectDylibs;
  97         bool                                    fTraceArchives;
  98         const char*                             fTraceOutputFile;
  99         VersionMin                              fVersionMin;
 100         std::vector<AliasPair>  fAliases;
 101 };
 102
 103
 104 class Reader
 105 {
 106 public:
 107         enum DebugInfoKind { kDebugInfoNone=0, kDebugInfoStabs=1, kDebugInfoDwarf=2, kDebugInfoStabsUUID=3 };
 108         struct Stab
 109         {
 110                 class Atom*     atom;
 111                 uint8_t         type;
 112                 uint8_t         other;
 113                 uint16_t        desc;
 114                 uint32_t        value;
 115                 const char* string;
 116         };
 117         enum ObjcConstraint { kObjcNone,  kObjcRetainRelease,  kObjcRetainReleaseOrGC,  kObjcGC };
 118         enum CpuConstraint  { kCpuAny = 0 };
 119
 120         class DylibHander
 121         {
 122         public:
 123                 virtual                         ~DylibHander()  {}
 124                 virtual Reader*         findDylib(const char* installPath, const char* fromPath) = 0;
 125         };
 126
 127
 128         static Reader* createReader(const char* path, const ReaderOptions& options);
 129
 130         virtual const char*                                     getPath() = 0;
 131         virtual time_t                                          getModificationTime() = 0;
 132         virtual DebugInfoKind                           getDebugInfoKind() = 0;
 133         virtual std::vector<class Atom*>&       getAtoms() = 0;
 134         virtual std::vector<class Atom*>*       getJustInTimeAtomsFor(const char* name) = 0;
 135         virtual std::vector<Stab>*                      getStabs() = 0;
 136         virtual ObjcConstraint                          getObjCConstraint()                     { return kObjcNone; }
 137         virtual uint32_t                                        updateCpuConstraint(uint32_t current) { return current; }
 138         virtual bool                                            objcReplacementClasses()        { return false; }
 139
 140         // For relocatable object files only
 141         virtual bool                                            canScatterAtoms()                       { return true; }
 142         virtual void                                            optimize(std::vector<ObjectFile::Atom*>&, std::vector<ObjectFile::Atom*>&,
 143                                                                                                         std::vector<const char*>&, const std::set<ObjectFile::Atom*>&,
 144                                                                                                         uint32_t, ObjectFile::Reader* writer,
 145                                                                                                         const std::vector<const char*>& llvmOptions,
 146                                                                                                         bool allGlobalsAReDeadStripRoots, int okind,
 147                                                                                                         bool verbose, bool saveTemps, const char* outputFilePath,
 148                                                                                                         bool pie, bool allowTextRelocs) { }
 149         virtual bool                                            hasLongBranchStubs()            { return false; }
 150
 151         // For Dynamic Libraries only
 152         virtual const char*                                     getInstallPath()                        { return NULL; }
 153         virtual uint32_t                                        getTimestamp()                          { return 0; }
 154         virtual uint32_t                                        getCurrentVersion()                     { return 0; }
 155         virtual uint32_t                                        getCompatibilityVersion()       { return 0; }
 156         virtual void                                            processIndirectLibraries(DylibHander* handler)  { }
 157         virtual void                                            setExplicitlyLinked()           { }
 158         virtual bool                                            explicitlyLinked()                      { return false; }
 159         virtual bool                                            implicitlyLinked()                      { return false; }
 160         virtual bool                                            providedExportAtom()            { return false; }
 161         virtual const char*                                     parentUmbrella()                        { return NULL; }
 162         virtual std::vector<const char*>*       getAllowableClients()           { return NULL; }
 163         virtual bool                                            hasWeakExternals()                      { return false; }
 164         virtual bool                                            isLazyLoadedDylib()                     { return false; }
 165
 166 protected:
 167                                                                                 Reader() {}
 168         virtual                                                         ~Reader() {}
 169 };
 170
 171 class Segment
 172 {
 173 public:
 174         virtual const char*                     getName() const  = 0;
 175         virtual bool                            isContentReadable() const = 0;
 176         virtual bool                            isContentWritable() const = 0;
 177         virtual bool                            isContentExecutable() const = 0;
 178
 179         uint64_t                                        getBaseAddress() const { return fBaseAddress; }
 180         void                                            setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 181         virtual bool                            hasFixedAddress() const { return false; }
 182
 183 protected:
 184                                                                 Segment() : fBaseAddress(0) {}
 185         virtual                                         ~Segment() {}
 186         uint64_t                                        fBaseAddress;
 187 };
 188
 189 class Reference;
 190
 191 class Section
 192 {
 193 public:
 194         unsigned int    getIndex() { return fIndex; }
 195         uint64_t                getBaseAddress() { return fBaseAddress; }
 196         void                    setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 197         void*                   fOther;
 198
 199 protected:
 200                                         Section() : fOther(NULL), fBaseAddress(0), fIndex(0)  {}
 201         uint64_t                fBaseAddress;
 202         unsigned int    fIndex;
 203 };
 204
 205
 206 struct Alignment
 207 {
 208                                 Alignment(int p2, int m=0) : powerOf2(p2), modulus(m) {}
 209         uint8_t         trailingZeros() const { return (modulus==0) ? powerOf2 : __builtin_ctz(modulus); }
 210         uint16_t        powerOf2;
 211         uint16_t        modulus;
 212 };
 213
 214 //
 215 // An atom is the fundamental unit of linking.  A C function or global variable is an atom.
 216 // An atom has content and some attributes. The content of a function atom is the instructions
 217 // that implement the function.  The content of a global variable atom is its initial bits.
 218 //
 219 // Name:
 220 // The name of an atom is the label name generated by the compiler.  A C compiler names foo()
 221 // as _foo.  A C++ compiler names foo() as __Z3foov.
 222 // The name refers to the first byte of the content.  An atom cannot have multiple entry points.
 223 // Such code is modeled as multiple atoms, each having a "follow on" reference to the next.
 224 // A "follow on" reference is a contraint to the linker to the atoms must be laid out contiguously.
 225 //
 226 // Scope:
 227 // An atom is in one of three scopes: translation-unit, linkage-unit, or global.  These correspond
 228 // to the C visibility of static, hidden, default.
 229 //
 230 // DefinitionKind:
 231 // An atom is one of five defintion kinds:
 232 //      regular                 Most atoms.
 233 //      weak                    C++ compiler makes some functions weak if there might be multiple copies
 234 //                                      that the linker needs to coalesce.
 235 //      tentative               A straggler from ancient C when the extern did not exist. "int foo;" is ambiguous.
 236 //                                      It could be a prototype or it could be a definition.
 237 //      external                This is a "proxy" atom produced by a dylib reader.  It has no content.  It exists
 238 //                                      so that all References can be resolved.
 239 //      external-weak   Same as external, but the definition in the dylib is weak.
 240 //
 241 // SymbolTableInclusion:
 242 // An atom may or may not be in the symbol table in an object file.
 243 //  in                          Most atoms for functions or global data
 244 //      not-in                  Anonymous atoms such literal c-strings, or other compiler generated data
 245 //      in-never-strip  Atom whose name the strip tool should never remove (e.g. REFERENCED_DYNAMICALLY in mach-o)
 246 //
 247 // Ordinal:
 248 // When a reader is created it is given a base ordinal number.  All atoms created by the reader
 249 // should return a contiguous range of ordinal values that start at the base ordinal.  The ordinal
 250 // values are used by the linker to sort the atom graph when producing the output file.
 251 //
 252 class Atom
 253 {
 254 public:
 255         enum Scope { scopeTranslationUnit, scopeLinkageUnit, scopeGlobal };
 256         enum DefinitionKind { kRegularDefinition, kWeakDefinition, kTentativeDefinition, kExternalDefinition, kExternalWeakDefinition, kAbsoluteSymbol };
 257         enum SymbolTableInclusion { kSymbolTableNotIn, kSymbolTableIn, kSymbolTableInAndNeverStrip, kSymbolTableInAsAbsolute };
 258
 259         virtual Reader*                                                 getFile() const = 0;
 260         virtual bool                                                    getTranslationUnitSource(const char** dir, const char** name) const = 0;
 261         virtual const char*                                             getName() const = 0;
 262         virtual const char*                                             getDisplayName() const = 0;
 263         virtual Scope                                                   getScope() const = 0;
 264         virtual DefinitionKind                                  getDefinitionKind() const = 0;
 265         virtual SymbolTableInclusion                    getSymbolTableInclusion() const = 0;
 266         virtual bool                                                    dontDeadStrip() const = 0;
 267         virtual bool                                                    isZeroFill() const = 0;
 268         virtual bool                                                    isThumb() const = 0;
 269         virtual uint64_t                                                getSize() const = 0;
 270         virtual std::vector<ObjectFile::Reference*>&  getReferences() const = 0;
 271         virtual bool                                                    mustRemainInSection() const = 0;
 272         virtual const char*                                             getSectionName() const = 0;
 273         virtual Segment&                                                getSegment() const = 0;
 274         virtual Atom&                                                   getFollowOnAtom() const = 0;
 275         virtual uint32_t                                                getOrdinal() const = 0;
 276         virtual std::vector<LineInfo>*                  getLineInfo() const = 0;
 277         virtual Alignment                                               getAlignment() const = 0;
 278         virtual void                                                    copyRawContent(uint8_t buffer[]) const = 0;
 279         virtual void                                                    setScope(Scope) = 0;
 280
 281
 282                         uint64_t                                                getSectionOffset() const        { return fSectionOffset; }
 283                         uint64_t                                                getAddress() const      { return fSection->getBaseAddress() + fSectionOffset; }
 284                         class Section*                                  getSection() const { return fSection; }
 285
 286         virtual void                                                    setSectionOffset(uint64_t offset) { fSectionOffset = offset; }
 287         virtual void                                                    setSection(class Section* sect) { fSection = sect; }
 288
 289 protected:
 290                                                                                         Atom() :  fSectionOffset(0), fSection(NULL) {}
 291                 virtual                                                         ~Atom() {}
 292
 293                 uint64_t                                                        fSectionOffset;
 294                 class Section*                                          fSection;
 295 };
 296
 297
 298 //
 299 // A Reference is a directed edge to another Atom.  When an instruction in
 300 // the content of an Atom refers to another Atom, that is represented by a
 301 // Reference.
 302 //
 303 // There are two kinds of references: direct and by-name.  With a direct Reference,
 304 // the target is bound by the Reader that created it.  For instance a reference to a
 305 // static would produce a direct reference.  A by-name reference requires the linker
 306 // to find the target Atom with the required name in order to be bound.
 307 //
 308 // For a link to succeed all References must be bound.
 309 //
 310 // A Reference has an optional "from" target.  This is used when the content to fix-up
 311 // is the difference of two Atom address.  For instance, if a pointer sized data Atom
 312 // is to contain A - B, then the Atom would have on Reference with a target of "A" and
 313 // a from-target of "B".
 314 //
 315 // A Reference also has a fix-up-offset.  This is the offset into the content of the
 316 // Atom holding the reference where the fix-up (relocation) will be applied.
 317 //
 318 //
 319 //
 320 class Reference
 321 {
 322 public:
 323         enum TargetBinding { kUnboundByName, kBoundDirectly, kBoundByName, kDontBind };
 324
 325         virtual TargetBinding   getTargetBinding() const = 0;
 326         virtual TargetBinding   getFromTargetBinding() const = 0;
 327         virtual uint8_t                 getKind() const = 0;
 328         virtual uint64_t                getFixUpOffset() const = 0;
 329         virtual const char*             getTargetName() const = 0;
 330         virtual Atom&                   getTarget() const = 0;
 331         virtual uint64_t                getTargetOffset() const = 0;
 332         virtual Atom&                   getFromTarget() const = 0;
 333         virtual const char*             getFromTargetName() const = 0;
 334         virtual uint64_t                getFromTargetOffset() const = 0;
 335
 336         virtual void                    setTarget(Atom&, uint64_t offset) = 0;
 337         virtual void                    setFromTarget(Atom&) = 0;
 338         virtual const char*             getDescription() const = 0;
 339
 340 protected:
 341                                                         Reference() {}
 342         virtual                                 ~Reference() {}
 343 };
 344
 345
 346 };      // namespace ObjectFile
 347
 348
 349 #endif // __OBJECTFILE__