Internalize.cpp revision 36eef827f1576cb240418a78e2323863ef6b4f8c 
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/Support/raw_ostream.h"
25#include "llvm/ADT/Statistic.h"
26#include <fstream>
27#include <set>
28using namespace llvm;
29
30STATISTIC(NumAliases  , "Number of aliases internalized");
31STATISTIC(NumFunctions, "Number of functions internalized");
32STATISTIC(NumGlobals  , "Number of global vars internalized");
33
34// APIFile - A file which contains a list of symbols that should not be marked
35// external.
36static cl::opt<std::string>
37APIFile("internalize-public-api-file", cl::value_desc("filename"),
38        cl::desc("A file containing list of symbol names to preserve"));
39
40// APIList - A list of symbols that should not be marked internal.
41static cl::list<std::string>
42APIList("internalize-public-api-list", cl::value_desc("list"),
43        cl::desc("A list of symbol names to preserve"),
44        cl::CommaSeparated);
45
46namespace {
47  class VISIBILITY_HIDDEN InternalizePass : public ModulePass {
48    std::set<std::string> ExternalNames;
49    /// If no api symbols were specified and a main function is defined,
50    /// assume the main function is the only API
51    bool AllButMain;
52  public:
53    static char ID; // Pass identification, replacement for typeid
54    explicit InternalizePass(bool AllButMain = true);
55    explicit InternalizePass(const std::vector <const char *>& exportList);
56    void LoadFile(const char *Filename);
57    virtual bool runOnModule(Module &M);
58
59    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
60      AU.setPreservesCFG();
61      AU.addPreserved<CallGraph>();
62    }
63  };
64} // end anonymous namespace
65
66char InternalizePass::ID = 0;
67static RegisterPass<InternalizePass>
68X("internalize", "Internalize Global Symbols");
69
70InternalizePass::InternalizePass(bool AllButMain)
71  : ModulePass(&ID), AllButMain(AllButMain){
72  if (!APIFile.empty())           // If a filename is specified, use it.
73    LoadFile(APIFile.c_str());
74  if (!APIList.empty())           // If a list is specified, use it as well.
75    ExternalNames.insert(APIList.begin(), APIList.end());
76}
77
78InternalizePass::InternalizePass(const std::vector<const char *>&exportList)
79  : ModulePass(&ID), AllButMain(false){
80  for(std::vector<const char *>::const_iterator itr = exportList.begin();
81        itr != exportList.end(); itr++) {
82    ExternalNames.insert(*itr);
83  }
84}
85
86void InternalizePass::LoadFile(const char *Filename) {
87  // Load the APIFile...
88  std::ifstream In(Filename);
89  if (!In.good()) {
90    errs() << "WARNING: Internalize couldn't load file '" << Filename
91         << "'! Continuing as if it's empty.\n";
92    return; // Just continue as if the file were empty
93  }
94  while (In) {
95    std::string Symbol;
96    In >> Symbol;
97    if (!Symbol.empty())
98      ExternalNames.insert(Symbol);
99  }
100}
101
102bool InternalizePass::runOnModule(Module &M) {
103  CallGraph *CG = getAnalysisIfAvailable<CallGraph>();
104  CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : 0;
105
106  if (ExternalNames.empty()) {
107    // Return if we're not in 'all but main' mode and have no external api
108    if (!AllButMain)
109      return false;
110    // If no list or file of symbols was specified, check to see if there is a
111    // "main" symbol defined in the module.  If so, use it, otherwise do not
112    // internalize the module, it must be a library or something.
113    //
114    Function *MainFunc = M.getFunction("main");
115    if (MainFunc == 0 || MainFunc->isDeclaration())
116      return false;  // No main found, must be a library...
117
118    // Preserve main, internalize all else.
119    ExternalNames.insert(MainFunc->getName());
120  }
121
122  bool Changed = false;
123
124  // Mark all functions not in the api as internal.
125  // FIXME: maybe use private linkage?
126  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
127    if (!I->isDeclaration() &&         // Function must be defined here
128        !I->hasLocalLinkage() &&  // Can't already have internal linkage
129        !ExternalNames.count(I->getName())) {// Not marked to keep external?
130      I->setLinkage(GlobalValue::InternalLinkage);
131      // Remove a callgraph edge from the external node to this function.
132      if (ExternalNode) ExternalNode->removeOneAbstractEdgeTo((*CG)[I]);
133      Changed = true;
134      ++NumFunctions;
135      DEBUG(errs() << "Internalizing func " << I->getName() << "\n");
136    }
137
138  // Never internalize the llvm.used symbol.  It is used to implement
139  // attribute((used)).
140  // FIXME: Shouldn't this just filter on llvm.metadata section??
141  ExternalNames.insert("llvm.used");
142  ExternalNames.insert("llvm.compiler.used");
143
144  // Never internalize anchors used by the machine module info, else the info
145  // won't find them.  (see MachineModuleInfo.)
146  ExternalNames.insert("llvm.dbg.compile_units");
147  ExternalNames.insert("llvm.dbg.global_variables");
148  ExternalNames.insert("llvm.dbg.subprograms");
149  ExternalNames.insert("llvm.global_ctors");
150  ExternalNames.insert("llvm.global_dtors");
151  ExternalNames.insert("llvm.noinline");
152  ExternalNames.insert("llvm.global.annotations");
153
154  // Mark all global variables with initializers that are not in the api as
155  // internal as well.
156  // FIXME: maybe use private linkage?
157  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
158       I != E; ++I)
159    if (!I->isDeclaration() && !I->hasLocalLinkage() &&
160        !ExternalNames.count(I->getName())) {
161      I->setLinkage(GlobalValue::InternalLinkage);
162      Changed = true;
163      ++NumGlobals;
164      DEBUG(errs() << "Internalized gvar " << I->getName() << "\n");
165    }
166
167  // Mark all aliases that are not in the api as internal as well.
168  for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
169       I != E; ++I)
170    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
171        !ExternalNames.count(I->getName())) {
172      I->setLinkage(GlobalValue::InternalLinkage);
173      Changed = true;
174      ++NumAliases;
175      DEBUG(errs() << "Internalized alias " << I->getName() << "\n");
176    }
177
178  return Changed;
179}
180
181ModulePass *llvm::createInternalizePass(bool AllButMain) {
182  return new InternalizePass(AllButMain);
183}
184
185ModulePass *llvm::createInternalizePass(const std::vector <const char *> &el) {
186  return new InternalizePass(el);
187}
188