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 and variables in the input module.
11// If the function or variable is not in the list of external names given to
12// the pass it is marked as internal.
13//
14//===----------------------------------------------------------------------===//
15
16#define DEBUG_TYPE "internalize"
17#include "llvm/Transforms/IPO.h"
18#include "llvm/ADT/Statistic.h"
19#include "llvm/Analysis/CallGraph.h"
20#include "llvm/IR/Module.h"
21#include "llvm/Pass.h"
22#include "llvm/Support/CommandLine.h"
23#include "llvm/Support/Debug.h"
24#include "llvm/Support/raw_ostream.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  public:
49    static char ID; // Pass identification, replacement for typeid
50    explicit InternalizePass();
51    explicit InternalizePass(ArrayRef<const char *> exportList);
52    void LoadFile(const char *Filename);
53    void ClearExportList();
54    void AddToExportList(const std::string &val);
55    virtual bool runOnModule(Module &M);
56
57    virtual void getAnalysisUsage(AnalysisUsage &AU) const {
58      AU.setPreservesCFG();
59      AU.addPreserved<CallGraph>();
60    }
61  };
62} // end anonymous namespace
63
64char InternalizePass::ID = 0;
65INITIALIZE_PASS(InternalizePass, "internalize",
66                "Internalize Global Symbols", false, false)
67
68InternalizePass::InternalizePass()
69  : ModulePass(ID) {
70  initializeInternalizePassPass(*PassRegistry::getPassRegistry());
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(ArrayRef<const char *> exportList)
78  : ModulePass(ID){
79  initializeInternalizePassPass(*PassRegistry::getPassRegistry());
80  for(ArrayRef<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
102void InternalizePass::ClearExportList() {
103  ExternalNames.clear();
104}
105
106void InternalizePass::AddToExportList(const std::string &val) {
107  ExternalNames.insert(val);
108}
109
110bool InternalizePass::runOnModule(Module &M) {
111  CallGraph *CG = getAnalysisIfAvailable<CallGraph>();
112  CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : 0;
113  bool Changed = false;
114
115  // Never internalize functions which code-gen might insert.
116  // FIXME: We should probably add this (and the __stack_chk_guard) via some
117  // type of call-back in CodeGen.
118  ExternalNames.insert("__stack_chk_fail");
119
120  // Mark all functions not in the api as internal.
121  // FIXME: maybe use private linkage?
122  for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I)
123    if (!I->isDeclaration() &&         // Function must be defined here
124        // Available externally is really just a "declaration with a body".
125        !I->hasAvailableExternallyLinkage() &&
126        !I->hasLocalLinkage() &&  // 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      DEBUG(dbgs() << "Internalizing func " << I->getName() << "\n");
134    }
135
136  // Never internalize the llvm.used symbol.  It is used to implement
137  // attribute((used)).
138  // FIXME: Shouldn't this just filter on llvm.metadata section??
139  ExternalNames.insert("llvm.used");
140  ExternalNames.insert("llvm.compiler.used");
141
142  // Never internalize anchors used by the machine module info, else the info
143  // won't find them.  (see MachineModuleInfo.)
144  ExternalNames.insert("llvm.global_ctors");
145  ExternalNames.insert("llvm.global_dtors");
146  ExternalNames.insert("llvm.global.annotations");
147
148  // Never internalize symbols code-gen inserts.
149  ExternalNames.insert("__stack_chk_guard");
150
151  // Mark all global variables with initializers that are not in the api as
152  // internal as well.
153  // FIXME: maybe use private linkage?
154  for (Module::global_iterator I = M.global_begin(), E = M.global_end();
155       I != E; ++I)
156    if (!I->isDeclaration() && !I->hasLocalLinkage() &&
157        // Available externally is really just a "declaration with a body".
158        !I->hasAvailableExternallyLinkage() &&
159        !ExternalNames.count(I->getName())) {
160      I->setLinkage(GlobalValue::InternalLinkage);
161      Changed = true;
162      ++NumGlobals;
163      DEBUG(dbgs() << "Internalized gvar " << I->getName() << "\n");
164    }
165
166  // Mark all aliases that are not in the api as internal as well.
167  for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
168       I != E; ++I)
169    if (!I->isDeclaration() && !I->hasInternalLinkage() &&
170        // Available externally is really just a "declaration with a body".
171        !I->hasAvailableExternallyLinkage() &&
172        !ExternalNames.count(I->getName())) {
173      I->setLinkage(GlobalValue::InternalLinkage);
174      Changed = true;
175      ++NumAliases;
176      DEBUG(dbgs() << "Internalized alias " << I->getName() << "\n");
177    }
178
179  return Changed;
180}
181
182ModulePass *llvm::createInternalizePass() {
183  return new InternalizePass();
184}
185
186ModulePass *llvm::createInternalizePass(ArrayRef<const char *> el) {
187  return new InternalizePass(el);
188}
189