Internalize.cpp revision 34c280e03131036065fb78613b754badb1c30514 
1//===-- Internalize.cpp - Mark functions internal -------------------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file was developed by the LLVM research group and is distributed under
6// the University of Illinois Open Source License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This pass loops over all of the functions in the input module, looking for a
11// main function.  If a main function is found, all other functions and all
12// global variables with initializers are marked as internal.
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "internalize"
17#include "llvm/Transforms/IPO.h"
18#include "llvm/Pass.h"
19#include "llvm/Module.h"
20#include "llvm/Support/CommandLine.h"
21#include "llvm/Support/Compiler.h"
22#include "llvm/Support/Debug.h"
23#include "llvm/ADT/Statistic.h"
24#include <fstream>
25#include <set>
26using namespace llvm;
27
28STATISTIC(NumFunctions, "Number of functions internalized");
29STATISTIC(NumGlobals  , "Number of global vars internalized");
30
31namespace {
32
33  // APIFile - A file which contains a list of symbols that should not be marked
34  // external.
35  cl::opt<std::string>
36  APIFile("internalize-public-api-file", cl::value_desc("filename"),
37          cl::desc("A file containing list of symbol names to preserve"));
38
39  // APIList - A list of symbols that should not be marked internal.
40  cl::list<std::string>
41  APIList("internalize-public-api-list", cl::value_desc("list"),
42          cl::desc("A list of symbol names to preserve"),
43          cl::CommaSeparated);
44
45  class VISIBILITY_HIDDEN InternalizePass : public ModulePass {
46    std::set<std::string> ExternalNames;
47    bool DontInternalize;
48  public:
49    static char ID; // Pass identification, replacement for typeid
50    explicit InternalizePass(bool InternalizeEverything = true);
51    explicit InternalizePass(const std::vector <const char *>& exportList);
52    void LoadFile(const char *Filename);
53    virtual bool runOnModule(Module &M);
54  };
55  char InternalizePass::ID = 0;
56  RegisterPass<InternalizePass> X("internalize", "Internalize Global Symbols");
57} // end anonymous namespace
58
59InternalizePass::InternalizePass(bool InternalizeEverything)
60  : ModulePass((intptr_t)&ID), DontInternalize(false){
61  if (!APIFile.empty())           // If a filename is specified, use it
62    LoadFile(APIFile.c_str());
63  else if (!APIList.empty())      // Else, if a list is specified, use it.
64    ExternalNames.insert(APIList.begin(), APIList.end());
65  else if (!InternalizeEverything)
66    // Finally, if we're allowed to, internalize all but main.
67    DontInternalize = true;
68}
69
70InternalizePass::InternalizePass(const std::vector<const char *>&exportList)
71  : ModulePass((intptr_t)&ID), DontInternalize(false){
72  for(std::vector<const char *>::const_iterator itr = exportList.begin();
73        itr != exportList.end(); itr++) {
74    ExternalNames.insert(*itr);
75  }
76}
77
78void InternalizePass::LoadFile(const char *Filename) {
79  // Load the APIFile...
80  std::ifstream In(Filename);
81  if (!In.good()) {
82    cerr << "WARNING: Internalize couldn't load file '" << Filename << "'!\n";
83    return;   // Do not internalize anything...
84  }
85  while (In) {
86    std::string Symbol;
87    In >> Symbol;
88    if (!Symbol.empty())
89      ExternalNames.insert(Symbol);
90  }
91}
92
93bool InternalizePass::runOnModule(Module &M) {
94  if (DontInternalize) return false;
95
96  // If no list or file of symbols was specified, check to see if there is a
97  // "main" symbol defined in the module.  If so, use it, otherwise do not
98  // internalize the module, it must be a library or something.
99  //
100  if (ExternalNames.empty()) {
101    Function *MainFunc = M.getFunction("main");
102    if (MainFunc == 0 || MainFunc->isDeclaration())
103      return false;  // No main found, must be a library...
104
105    // Preserve main, internalize all else.
106    ExternalNames.insert(MainFunc->getName());
107  }
108
109  bool Changed = false;
110
111  // Found a main function, mark all functions not named main as internal.
112  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
113    if (!I->isDeclaration() &&         // Function must be defined here
114        !I->hasInternalLinkage() &&  // Can't already have internal linkage
115        !ExternalNames.count(I->getName())) {// Not marked to keep external?
116      I->setLinkage(GlobalValue::InternalLinkage);
117      Changed = true;
118      ++NumFunctions;
119      DOUT << "Internalizing func " << I->getName() << "\n";
120    }
121
122  // Never internalize the llvm.used symbol.  It is used to implement
123  // attribute((used)).
124  ExternalNames.insert("llvm.used");
125
126  // Never internalize anchors used by the machine module info, else the info
127  // won't find them.  (see MachineModuleInfo.)
128  ExternalNames.insert("llvm.dbg.compile_units");
129  ExternalNames.insert("llvm.dbg.global_variables");
130  ExternalNames.insert("llvm.dbg.subprograms");
131  ExternalNames.insert("llvm.global_ctors");
132  ExternalNames.insert("llvm.global_dtors");
133
134  // Mark all global variables with initializers as internal as well.
135  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
136       I != E; ++I)
137    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
138        !ExternalNames.count(I->getName())) {
139      I->setLinkage(GlobalValue::InternalLinkage);
140      Changed = true;
141      ++NumGlobals;
142      DOUT << "Internalized gvar " << I->getName() << "\n";
143    }
144
145  return Changed;
146}
147
148ModulePass *llvm::createInternalizePass(bool InternalizeEverything) {
149  return new InternalizePass(InternalizeEverything);
150}
151
152ModulePass *llvm::createInternalizePass(const std::vector <const char *> &el) {
153  return new InternalizePass(el);
154}
155