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