Internalize.cpp revision 27a53009efcf1b0334dc17c3d54382798686ff59 
1//===-- Internalize.cpp - Mark functions internal -------------------------===//
2//
3//                     The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// 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/Analysis/CallGraph.h"
18#include "llvm/Transforms/IPO.h"
19#include "llvm/Pass.h"
20#include "llvm/Module.h"
21#include "llvm/Support/CommandLine.h"
22#include "llvm/Support/Compiler.h"
23#include "llvm/Support/Debug.h"
24#include "llvm/ADT/Statistic.h"
25#include <fstream>
26#include <set>
27using namespace llvm;
28
29STATISTIC(NumAliases  , "Number of aliases internalized");
30STATISTIC(NumFunctions, "Number of functions internalized");
31STATISTIC(NumGlobals  , "Number of global vars internalized");
32
33// APIFile - A file which contains a list of symbols that should not be marked
34// external.
35static cl::opt<std::string>
36APIFile("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.
40static cl::list<std::string>
41APIList("internalize-public-api-list", cl::value_desc("list"),
42        cl::desc("A list of symbol names to preserve"),
43        cl::CommaSeparated);
44
45namespace {
46  class VISIBILITY_HIDDEN InternalizePass : public ModulePass {
47    std::set<std::string> ExternalNames;
48    /// If no api symbols were specified and a main function is defined,
49    /// assume the main function is the only API
50    bool AllButMain;
51  public:
52    static char ID; // Pass identification, replacement for typeid
53    explicit InternalizePass(bool AllButMain = true);
54    explicit InternalizePass(const std::vector <const char *>& exportList);
55    void LoadFile(const char *Filename);
56    virtual bool runOnModule(Module &M);
57
58    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
59      AU.setPreservesCFG();
60      AU.addPreserved<CallGraph>();
61    }
62  };
63} // end anonymous namespace
64
65char InternalizePass::ID = 0;
66static RegisterPass<InternalizePass>
67X("internalize", "Internalize Global Symbols");
68
69InternalizePass::InternalizePass(bool AllButMain)
70  : ModulePass(&ID), AllButMain(AllButMain){
71  if (!APIFile.empty())           // If a filename is specified, use it.
72    LoadFile(APIFile.c_str());
73  if (!APIList.empty())           // If a list is specified, use it as well.
74    ExternalNames.insert(APIList.begin(), APIList.end());
75}
76
77InternalizePass::InternalizePass(const std::vector<const char *>&exportList)
78  : ModulePass(&ID), AllButMain(false){
79  for(std::vector<const char *>::const_iterator itr = exportList.begin();
80        itr != exportList.end(); itr++) {
81    ExternalNames.insert(*itr);
82  }
83}
84
85void InternalizePass::LoadFile(const char *Filename) {
86  // Load the APIFile...
87  std::ifstream In(Filename);
88  if (!In.good()) {
89    cerr << "WARNING: Internalize couldn't load file '" << Filename
90         << "'! Continuing as if it's empty.\n";
91    return; // Just continue as if the file were empty
92  }
93  while (In) {
94    std::string Symbol;
95    In >> Symbol;
96    if (!Symbol.empty())
97      ExternalNames.insert(Symbol);
98  }
99}
100
101bool InternalizePass::runOnModule(Module &M) {
102  CallGraph *CG = getAnalysisToUpdate<CallGraph>();
103  CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : 0;
104
105  if (ExternalNames.empty()) {
106    // Return if we're not in 'all but main' mode and have no external api
107    if (!AllButMain)
108      return false;
109    // If no list or file of symbols was specified, check to see if there is a
110    // "main" symbol defined in the module.  If so, use it, otherwise do not
111    // internalize the module, it must be a library or something.
112    //
113    Function *MainFunc = M.getFunction("main");
114    if (MainFunc == 0 || MainFunc->isDeclaration())
115      return false;  // No main found, must be a library...
116
117    // Preserve main, internalize all else.
118    ExternalNames.insert(MainFunc->getName());
119  }
120
121  bool Changed = false;
122
123  // Mark all functions not in the api as internal.
124  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
125    if (!I->isDeclaration() &&         // Function must be defined here
126        !I->hasInternalLinkage() &&  // Can't already have internal linkage
127        !ExternalNames.count(I->getName())) {// Not marked to keep external?
128      I->setLinkage(GlobalValue::InternalLinkage);
129      // Remove a callgraph edge from the external node to this function.
130      if (ExternalNode) ExternalNode->removeOneAbstractEdgeTo((*CG)[I]);
131      Changed = true;
132      ++NumFunctions;
133      DOUT << "Internalizing func " << I->getName() << "\n";
134    }
135
136  // Never internalize the llvm.used symbol.  It is used to implement
137  // attribute((used)).
138  ExternalNames.insert("llvm.used");
139
140  // Never internalize anchors used by the machine module info, else the info
141  // won't find them.  (see MachineModuleInfo.)
142  ExternalNames.insert("llvm.dbg.compile_units");
143  ExternalNames.insert("llvm.dbg.global_variables");
144  ExternalNames.insert("llvm.dbg.subprograms");
145  ExternalNames.insert("llvm.global_ctors");
146  ExternalNames.insert("llvm.global_dtors");
147  ExternalNames.insert("llvm.noinline");
148  ExternalNames.insert("llvm.global.annotations");
149
150  // Mark all global variables with initializers that are not in the api as
151  // internal as well.
152  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
153       I != E; ++I)
154    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
155        !ExternalNames.count(I->getName())) {
156      I->setLinkage(GlobalValue::InternalLinkage);
157      Changed = true;
158      ++NumGlobals;
159      DOUT << "Internalized gvar " << I->getName() << "\n";
160    }
161
162  // Mark all aliases that are not in the api as internal as well.
163  for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
164       I != E; ++I)
165    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
166        !ExternalNames.count(I->getName())) {
167      I->setLinkage(GlobalValue::InternalLinkage);
168      Changed = true;
169      ++NumAliases;
170      DOUT << "Internalized alias " << I->getName() << "\n";
171    }
172
173  return Changed;
174}
175
176ModulePass *llvm::createInternalizePass(bool AllButMain) {
177  return new InternalizePass(AllButMain);
178}
179
180ModulePass *llvm::createInternalizePass(const std::vector <const char *> &el) {
181  return new InternalizePass(el);
182}
183