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/Debug.h"
23#include "llvm/Support/raw_ostream.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 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;
66INITIALIZE_PASS(InternalizePass, "internalize",
67                "Internalize Global Symbols", false, false)
68
69InternalizePass::InternalizePass(bool AllButMain)
70  : ModulePass(ID), AllButMain(AllButMain){
71  initializeInternalizePassPass(*PassRegistry::getPassRegistry());
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  initializeInternalizePassPass(*PassRegistry::getPassRegistry());
81  for(std::vector<const char *>::const_iterator itr = exportList.begin();
82        itr != exportList.end(); itr++) {
83    ExternalNames.insert(*itr);
84  }
85}
86
87void InternalizePass::LoadFile(const char *Filename) {
88  // Load the APIFile...
89  std::ifstream In(Filename);
90  if (!In.good()) {
91    errs() << "WARNING: Internalize couldn't load file '" << Filename
92         << "'! Continuing as if it's empty.\n";
93    return; // Just continue as if the file were empty
94  }
95  while (In) {
96    std::string Symbol;
97    In >> Symbol;
98    if (!Symbol.empty())
99      ExternalNames.insert(Symbol);
100  }
101}
102
103bool InternalizePass::runOnModule(Module &M) {
104  CallGraph *CG = getAnalysisIfAvailable<CallGraph>();
105  CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : 0;
106
107  if (ExternalNames.empty()) {
108    // Return if we're not in 'all but main' mode and have no external api
109    if (!AllButMain)
110      return false;
111    // If no list or file of symbols was specified, check to see if there is a
112    // "main" symbol defined in the module.  If so, use it, otherwise do not
113    // internalize the module, it must be a library or something.
114    //
115    Function *MainFunc = M.getFunction("main");
116    if (MainFunc == 0 || MainFunc->isDeclaration())
117      return false;  // No main found, must be a library...
118
119    // Preserve main, internalize all else.
120    ExternalNames.insert(MainFunc->getName());
121  }
122
123  bool Changed = false;
124
125  // Mark all functions not in the api as internal.
126  // FIXME: maybe use private linkage?
127  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
128    if (!I->isDeclaration() &&         // Function must be defined here
129        // Available externally is really just a "declaration with a body".
130        !I->hasAvailableExternallyLinkage() &&
131        !I->hasLocalLinkage() &&  // Can't already have internal linkage
132        !ExternalNames.count(I->getName())) {// Not marked to keep external?
133      I->setLinkage(GlobalValue::InternalLinkage);
134      // Remove a callgraph edge from the external node to this function.
135      if (ExternalNode) ExternalNode->removeOneAbstractEdgeTo((*CG)[I]);
136      Changed = true;
137      ++NumFunctions;
138      DEBUG(dbgs() << "Internalizing func " << I->getName() << "\n");
139    }
140
141  // Never internalize the llvm.used symbol.  It is used to implement
142  // attribute((used)).
143  // FIXME: Shouldn't this just filter on llvm.metadata section??
144  ExternalNames.insert("llvm.used");
145  ExternalNames.insert("llvm.compiler.used");
146
147  // Never internalize anchors used by the machine module info, else the info
148  // won't find them.  (see MachineModuleInfo.)
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        // Available externally is really just a "declaration with a body".
161        !I->hasAvailableExternallyLinkage() &&
162        !ExternalNames.count(I->getName())) {
163      I->setLinkage(GlobalValue::InternalLinkage);
164      Changed = true;
165      ++NumGlobals;
166      DEBUG(dbgs() << "Internalized gvar " << I->getName() << "\n");
167    }
168
169  // Mark all aliases that are not in the api as internal as well.
170  for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
171       I != E; ++I)
172    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
173        // Available externally is really just a "declaration with a body".
174        !I->hasAvailableExternallyLinkage() &&
175        !ExternalNames.count(I->getName())) {
176      I->setLinkage(GlobalValue::InternalLinkage);
177      Changed = true;
178      ++NumAliases;
179      DEBUG(dbgs() << "Internalized alias " << I->getName() << "\n");
180    }
181
182  return Changed;
183}
184
185ModulePass *llvm::createInternalizePass(bool AllButMain) {
186  return new InternalizePass(AllButMain);
187}
188
189ModulePass *llvm::createInternalizePass(const std::vector <const char *> &el) {
190  return new InternalizePass(el);
191}
192