src/ObjectFile.h

   1 /* -*- mode: C++; c-basic-offset: 4; tab-width: 4 -*-
   2  *
   3  * Copyright (c) 2005-2006 Apple Computer, 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), fLoadObjcClassesInArchives(false), fFlatNamespace(false),
  63                                                                                 fForFinalLinkedImage(false), fWhyLoad(false), fDebugInfoStripping(kDebugInfoFull),
  64                                                                                  fTraceDylibs(false), fTraceIndirectDylibs(false), fTraceArchives(false), fTraceOutputFile(NULL) {}
  65         enum DebugInfoStripping { kDebugInfoNone, kDebugInfoMinimal, kDebugInfoFull };
  66
  67         bool                            fFullyLoadArchives;
  68         bool                            fLoadObjcClassesInArchives;
  69         bool                            fFlatNamespace;
  70         bool                            fForFinalLinkedImage;
  71         bool                            fWhyLoad;
  72         DebugInfoStripping      fDebugInfoStripping;
  73         bool                            fTraceDylibs;
  74         bool                            fTraceIndirectDylibs;
  75         bool                            fTraceArchives;
  76         const char*                     fTraceOutputFile;
  77 };
  78
  79
  80 class Reader
  81 {
  82 public:
  83         enum DebugInfoKind { kDebugInfoNone=0, kDebugInfoStabs=1, kDebugInfoDwarf=2, kDebugInfoStabsUUID=3 };
  84         struct Stab
  85         {
  86                 class Atom*     atom;
  87                 uint8_t         type;
  88                 uint8_t         other;
  89                 uint16_t        desc;
  90                 uint32_t        value;
  91                 const char* string;
  92         };
  93
  94         static Reader* createReader(const char* path, const ReaderOptions& options);
  95
  96         virtual const char*                                     getPath() = 0;
  97         virtual time_t                                          getModificationTime() = 0;
  98         virtual DebugInfoKind                           getDebugInfoKind() = 0;
  99         virtual std::vector<class Atom*>&       getAtoms() = 0;
 100         virtual std::vector<class Atom*>*       getJustInTimeAtomsFor(const char* name) = 0;
 101         virtual std::vector<Stab>*                      getStabs() = 0;
 102                         unsigned int                            getSortOrder() const { return fSortOrder; }
 103                         void                                            setSortOrder(unsigned int order) { fSortOrder=order; }
 104
 105         // For Dynamic Libraries only
 106         virtual const char*                                     getInstallPath()                        { return NULL; }
 107         virtual uint32_t                                        getTimestamp()                          { return 0; }
 108         virtual uint32_t                                        getCurrentVersion()                     { return 0; }
 109         virtual uint32_t                                        getCompatibilityVersion()       { return 0; }
 110         virtual std::vector<const char*>*       getDependentLibraryPaths()      { return NULL; }
 111         virtual bool                                            reExports(Reader*)                      { return false; }
 112         virtual const char*                                     parentUmbrella()                        { return NULL; }
 113         virtual std::vector<const char*>*       getAllowableClients()           { return NULL; }
 114
 115 protected:
 116                                                                                 Reader() : fSortOrder(0) {}
 117         virtual                                                         ~Reader() {}
 118
 119         unsigned int                                            fSortOrder;
 120 };
 121
 122 class Segment
 123 {
 124 public:
 125         virtual const char*                     getName() const  = 0;
 126         virtual bool                            isContentReadable() const = 0;
 127         virtual bool                            isContentWritable() const = 0;
 128         virtual bool                            isContentExecutable() const = 0;
 129
 130         uint64_t                                        getBaseAddress() const { return fBaseAddress; }
 131         void                                            setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 132         virtual bool                            hasFixedAddress() const { return false; }
 133
 134 protected:
 135                                                                 Segment() : fBaseAddress(0) {}
 136         virtual                                         ~Segment() {}
 137         uint64_t                                        fBaseAddress;
 138 };
 139
 140 class Reference;
 141
 142 class Section
 143 {
 144 public:
 145         unsigned int    getIndex() { return fIndex; }
 146         uint64_t                getBaseAddress() { return fBaseAddress; }
 147         void                    setBaseAddress(uint64_t addr) { fBaseAddress = addr; }
 148         void*                   fOther;
 149
 150 protected:
 151                                         Section() : fOther(NULL), fBaseAddress(0), fIndex(0)  {}
 152         uint64_t                fBaseAddress;
 153         unsigned int    fIndex;
 154 };
 155
 156
 157
 158 //
 159 // An atom is the fundamental unit of linking.  A C function or global variable is an atom.
 160 // An atom has content and some attributes. The content of a function atom is the instructions
 161 // that implement the function.  The content of a global variable atom is its initial bits.
 162 //
 163 // Name:
 164 // The name of an atom is the label name generated by the compiler.  A C compiler names foo()
 165 // as _foo.  A C++ compiler names foo() as __Z3foov.
 166 // The name refers to the first byte of the content.  An atom cannot have multiple entry points.
 167 // Such code is modeled as multiple atoms, each having a "follow on" reference to the next.
 168 // A "follow on" reference is a contraint to the linker to the atoms must be laid out contiguously.
 169 //
 170 // Scope:
 171 // An atom is in one of three scopes: translation-unit, linkage-unit, or global.  These correspond
 172 // to the C visibility of static, hidden, default.
 173 //
 174 // DefinitionKind:
 175 // An atom is one of five defintion kinds:
 176 //      regular                 Most atoms.
 177 //      weak                    C++ compiler makes some functions weak if there might be multiple copies
 178 //                                      that the linker needs to coalesce.
 179 //      tentative               A straggler from ancient C when the extern did not exist. "int foo;" is ambiguous.
 180 //                                      It could be a prototype or it could be a definition.
 181 //      external                This is a "proxy" atom produced by a dylib reader.  It has no content.  It exists
 182 //                                      so that all References can be resolved.
 183 //      external-weak   Same as external, but the definition in the dylib is weak.
 184 //
 185 // SymbolTableInclusion:
 186 // An atom may or may not be in the symbol table in an object file.
 187 //  in                          Most atoms for functions or global data
 188 //      not-in                  Anonymous atoms such literal c-strings, or other compiler generated data
 189 //      in-never-strip  Atom whose name the strip tool should never remove (e.g. REFERENCED_DYNAMICALLY in mach-o)
 190 //
 191 class Atom
 192 {
 193 public:
 194         enum Scope { scopeTranslationUnit, scopeLinkageUnit, scopeGlobal };
 195         enum DefinitionKind { kRegularDefinition, kWeakDefinition, kTentativeDefinition, kExternalDefinition, kExternalWeakDefinition };
 196         enum SymbolTableInclusion { kSymbolTableNotIn, kSymbolTableIn, kSymbolTableInAndNeverStrip, kSymbolTableInAsAbsolute };
 197
 198         virtual Reader*                                                 getFile() const = 0;
 199         virtual bool                                                    getTranslationUnitSource(const char** dir, const char** name) const = 0;
 200         virtual const char*                                             getName() const = 0;
 201         virtual const char*                                             getDisplayName() const = 0;
 202         virtual Scope                                                   getScope() const = 0;
 203         virtual DefinitionKind                                  getDefinitionKind() const = 0;
 204         virtual SymbolTableInclusion                    getSymbolTableInclusion() const = 0;
 205         virtual bool                                                    dontDeadStrip() const = 0;
 206         virtual bool                                                    isZeroFill() const = 0;
 207         virtual uint64_t                                                getSize() const = 0;
 208         virtual std::vector<ObjectFile::Reference*>&  getReferences() const = 0;
 209         virtual bool                                                    mustRemainInSection() const = 0;
 210         virtual const char*                                             getSectionName() const = 0;
 211         virtual Segment&                                                getSegment() const = 0;
 212         virtual bool                                                    requiresFollowOnAtom() const = 0;
 213         virtual Atom&                                                   getFollowOnAtom() const = 0;
 214         virtual std::vector<LineInfo>*                  getLineInfo() const = 0;
 215         virtual uint8_t                                                 getAlignment() const = 0;
 216         virtual void                                                    copyRawContent(uint8_t buffer[]) const = 0;
 217         virtual void                                                    setScope(Scope) = 0;
 218
 219
 220                         uint64_t                                                getSectionOffset() const        { return fSectionOffset; }
 221                         uint64_t                                                getSegmentOffset() const        { return fSegmentOffset; }
 222                         uint64_t                                                getAddress() const      { return fSection->getBaseAddress() + fSectionOffset; }
 223                         unsigned int                                    getSortOrder() const { return fSortOrder; }
 224                         class Section*                                  getSection() const { return fSection; }
 225
 226                         void                                                    setSegmentOffset(uint64_t offset) { fSegmentOffset = offset; }
 227                         void                                                    setSectionOffset(uint64_t offset) { fSectionOffset = offset; }
 228                         void                                                    setSection(class Section* sect) { fSection = sect; }
 229                         unsigned int                                    setSortOrder(unsigned int order); // recursively sets follow-on atoms
 230
 231 protected:
 232                                                                                         Atom() : fSegmentOffset(0), fSectionOffset(0), fSortOrder(0), fSection(NULL) {}
 233                 virtual                                                         ~Atom() {}
 234
 235                 uint64_t                                                        fSegmentOffset;
 236                 uint64_t                                                        fSectionOffset;
 237                 unsigned int                                            fSortOrder;
 238                 class Section*                                          fSection;
 239 };
 240
 241
 242
 243 // recursively sets follow-on atoms
 244 inline unsigned int Atom::setSortOrder(unsigned int order)
 245 {
 246         if ( this->requiresFollowOnAtom() ) {
 247                 fSortOrder = order;
 248                 return this->getFollowOnAtom().setSortOrder(order+1);
 249         }
 250         else {
 251                 fSortOrder = order;
 252                 return (order + 1);
 253         }
 254 }
 255
 256
 257
 258 //
 259 // A Reference is a directed edge to another Atom.  When an instruction in
 260 // the content of an Atom refers to another Atom, that is represented by a
 261 // Reference.
 262 //
 263 // There are two kinds of references: direct and by-name.  With a direct Reference,
 264 // the target is bound by the Reader that created it.  For instance a reference to a
 265 // static would produce a direct reference.  A by-name reference requires the linker
 266 // to find the target Atom with the required name in order to be bound.
 267 //
 268 // For a link to succeed all References must be bound.
 269 //
 270 // A Reference has an optional "from" target.  This is used when the content to fix-up
 271 // is the difference of two Atom address.  For instance, if a pointer sized data Atom
 272 // is to contain A - B, then the Atom would have on Reference with a target of "A" and
 273 // a from-target of "B".
 274 //
 275 // A Reference also has a fix-up-offset.  This is the offset into the content of the
 276 // Atom holding the reference where the fix-up (relocation) will be applied.
 277 //
 278 //
 279 //
 280 class Reference
 281 {
 282 public:
 283
 284         virtual bool                    isTargetUnbound() const = 0;
 285         virtual bool                    isFromTargetUnbound() const = 0;
 286         virtual uint8_t                 getKind() const = 0;
 287         virtual uint64_t                getFixUpOffset() const = 0;
 288         virtual const char*             getTargetName() const = 0;
 289         virtual Atom&                   getTarget() const = 0;
 290         virtual uint64_t                getTargetOffset() const = 0;
 291         virtual bool                    hasFromTarget() const = 0;
 292         virtual Atom&                   getFromTarget() const = 0;
 293         virtual const char*             getFromTargetName() const = 0;
 294         virtual uint64_t                getFromTargetOffset() const = 0;
 295
 296         virtual void                    setTarget(Atom&, uint64_t offset) = 0;
 297         virtual void                    setFromTarget(Atom&) = 0;
 298         virtual const char*             getDescription() const = 0;
 299
 300 protected:
 301                                                         Reference() {}
 302         virtual                                 ~Reference() {}
 303 };
 304
 305
 306 };      // namespace ObjectFile
 307
 308
 309 #endif // __OBJECTFILE__