1f22ef01cSRoman Divacky //===-- Internalize.cpp - Mark functions internal -------------------------===//
2f22ef01cSRoman Divacky //
3f22ef01cSRoman Divacky //                     The LLVM Compiler Infrastructure
4f22ef01cSRoman Divacky //
5f22ef01cSRoman Divacky // This file is distributed under the University of Illinois Open Source
6f22ef01cSRoman Divacky // License. See LICENSE.TXT for details.
7f22ef01cSRoman Divacky //
8f22ef01cSRoman Divacky //===----------------------------------------------------------------------===//
9f22ef01cSRoman Divacky //
103861d79fSDimitry Andric // This pass loops over all of the functions and variables in the input module.
113861d79fSDimitry Andric // If the function or variable is not in the list of external names given to
123861d79fSDimitry Andric // the pass it is marked as internal.
13f22ef01cSRoman Divacky //
14f785676fSDimitry Andric // This transformation would not be legal in a regular compilation, but it gets
15f785676fSDimitry Andric // extra information from the linker about what is safe.
16f785676fSDimitry Andric //
17f785676fSDimitry Andric // For example: Internalizing a function with external linkage. Only if we are
18f785676fSDimitry Andric // told it is only used from within this module, it is safe to do it.
19f785676fSDimitry Andric //
20f22ef01cSRoman Divacky //===----------------------------------------------------------------------===//
21f22ef01cSRoman Divacky 
22f22ef01cSRoman Divacky #include "llvm/Transforms/IPO.h"
23f785676fSDimitry Andric #include "llvm/ADT/SmallPtrSet.h"
24139f7f9bSDimitry Andric #include "llvm/ADT/Statistic.h"
25139f7f9bSDimitry Andric #include "llvm/Analysis/CallGraph.h"
26139f7f9bSDimitry Andric #include "llvm/IR/Module.h"
27f22ef01cSRoman Divacky #include "llvm/Pass.h"
28f22ef01cSRoman Divacky #include "llvm/Support/CommandLine.h"
29f22ef01cSRoman Divacky #include "llvm/Support/Debug.h"
30f22ef01cSRoman Divacky #include "llvm/Support/raw_ostream.h"
31f785676fSDimitry Andric #include "llvm/Transforms/Utils/GlobalStatus.h"
32f785676fSDimitry Andric #include "llvm/Transforms/Utils/ModuleUtils.h"
33f22ef01cSRoman Divacky #include <fstream>
34f22ef01cSRoman Divacky #include <set>
35f22ef01cSRoman Divacky using namespace llvm;
36f22ef01cSRoman Divacky 
3791bc56edSDimitry Andric #define DEBUG_TYPE "internalize"
3891bc56edSDimitry Andric 
39f22ef01cSRoman Divacky STATISTIC(NumAliases  , "Number of aliases internalized");
40f22ef01cSRoman Divacky STATISTIC(NumFunctions, "Number of functions internalized");
41f22ef01cSRoman Divacky STATISTIC(NumGlobals  , "Number of global vars internalized");
42f22ef01cSRoman Divacky 
43f22ef01cSRoman Divacky // APIFile - A file which contains a list of symbols that should not be marked
44f22ef01cSRoman Divacky // external.
45f22ef01cSRoman Divacky static cl::opt<std::string>
46f22ef01cSRoman Divacky APIFile("internalize-public-api-file", cl::value_desc("filename"),
47f22ef01cSRoman Divacky         cl::desc("A file containing list of symbol names to preserve"));
48f22ef01cSRoman Divacky 
49f22ef01cSRoman Divacky // APIList - A list of symbols that should not be marked internal.
50f22ef01cSRoman Divacky static cl::list<std::string>
51f22ef01cSRoman Divacky APIList("internalize-public-api-list", cl::value_desc("list"),
52f22ef01cSRoman Divacky         cl::desc("A list of symbol names to preserve"),
53f22ef01cSRoman Divacky         cl::CommaSeparated);
54f22ef01cSRoman Divacky 
55f22ef01cSRoman Divacky namespace {
56f22ef01cSRoman Divacky   class InternalizePass : public ModulePass {
57f22ef01cSRoman Divacky     std::set<std::string> ExternalNames;
58f22ef01cSRoman Divacky   public:
59f22ef01cSRoman Divacky     static char ID; // Pass identification, replacement for typeid
603861d79fSDimitry Andric     explicit InternalizePass();
61f785676fSDimitry Andric     explicit InternalizePass(ArrayRef<const char *> ExportList);
62f22ef01cSRoman Divacky     void LoadFile(const char *Filename);
6391bc56edSDimitry Andric     bool runOnModule(Module &M) override;
64f22ef01cSRoman Divacky 
6591bc56edSDimitry Andric     void getAnalysisUsage(AnalysisUsage &AU) const override {
66f22ef01cSRoman Divacky       AU.setPreservesCFG();
6791bc56edSDimitry Andric       AU.addPreserved<CallGraphWrapperPass>();
68f22ef01cSRoman Divacky     }
69f22ef01cSRoman Divacky   };
70f22ef01cSRoman Divacky } // end anonymous namespace
71f22ef01cSRoman Divacky 
72f22ef01cSRoman Divacky char InternalizePass::ID = 0;
73e580952dSDimitry Andric INITIALIZE_PASS(InternalizePass, "internalize",
742754fe60SDimitry Andric                 "Internalize Global Symbols", false, false)
75f22ef01cSRoman Divacky 
7691bc56edSDimitry Andric InternalizePass::InternalizePass() : ModulePass(ID) {
772754fe60SDimitry Andric   initializeInternalizePassPass(*PassRegistry::getPassRegistry());
78f22ef01cSRoman Divacky   if (!APIFile.empty())           // If a filename is specified, use it.
79f22ef01cSRoman Divacky     LoadFile(APIFile.c_str());
80f22ef01cSRoman Divacky   ExternalNames.insert(APIList.begin(), APIList.end());
81f22ef01cSRoman Divacky }
82f22ef01cSRoman Divacky 
83f785676fSDimitry Andric InternalizePass::InternalizePass(ArrayRef<const char *> ExportList)
843861d79fSDimitry Andric     : ModulePass(ID) {
852754fe60SDimitry Andric   initializeInternalizePassPass(*PassRegistry::getPassRegistry());
86f785676fSDimitry Andric   for(ArrayRef<const char *>::const_iterator itr = ExportList.begin();
87f785676fSDimitry Andric         itr != ExportList.end(); itr++) {
88f22ef01cSRoman Divacky     ExternalNames.insert(*itr);
89f22ef01cSRoman Divacky   }
90f22ef01cSRoman Divacky }
91f22ef01cSRoman Divacky 
92f22ef01cSRoman Divacky void InternalizePass::LoadFile(const char *Filename) {
93f22ef01cSRoman Divacky   // Load the APIFile...
94f22ef01cSRoman Divacky   std::ifstream In(Filename);
95f22ef01cSRoman Divacky   if (!In.good()) {
96f22ef01cSRoman Divacky     errs() << "WARNING: Internalize couldn't load file '" << Filename
97f22ef01cSRoman Divacky          << "'! Continuing as if it's empty.\n";
98f22ef01cSRoman Divacky     return; // Just continue as if the file were empty
99f22ef01cSRoman Divacky   }
100f22ef01cSRoman Divacky   while (In) {
101f22ef01cSRoman Divacky     std::string Symbol;
102f22ef01cSRoman Divacky     In >> Symbol;
103f22ef01cSRoman Divacky     if (!Symbol.empty())
104f22ef01cSRoman Divacky       ExternalNames.insert(Symbol);
105f22ef01cSRoman Divacky   }
106f22ef01cSRoman Divacky }
107f22ef01cSRoman Divacky 
108f785676fSDimitry Andric static bool shouldInternalize(const GlobalValue &GV,
109f785676fSDimitry Andric                               const std::set<std::string> &ExternalNames) {
110f785676fSDimitry Andric   // Function must be defined here
111f785676fSDimitry Andric   if (GV.isDeclaration())
112f785676fSDimitry Andric     return false;
113139f7f9bSDimitry Andric 
114f785676fSDimitry Andric   // Available externally is really just a "declaration with a body".
115f785676fSDimitry Andric   if (GV.hasAvailableExternallyLinkage())
116f785676fSDimitry Andric     return false;
117f785676fSDimitry Andric 
11891bc56edSDimitry Andric   // Assume that dllexported symbols are referenced elsewhere
11991bc56edSDimitry Andric   if (GV.hasDLLExportStorageClass())
12091bc56edSDimitry Andric     return false;
12191bc56edSDimitry Andric 
122f785676fSDimitry Andric   // Already has internal linkage
123f785676fSDimitry Andric   if (GV.hasLocalLinkage())
124f785676fSDimitry Andric     return false;
125f785676fSDimitry Andric 
126f785676fSDimitry Andric   // Marked to keep external?
127f785676fSDimitry Andric   if (ExternalNames.count(GV.getName()))
128f785676fSDimitry Andric     return false;
129f785676fSDimitry Andric 
130f785676fSDimitry Andric   return true;
131139f7f9bSDimitry Andric }
132139f7f9bSDimitry Andric 
133f22ef01cSRoman Divacky bool InternalizePass::runOnModule(Module &M) {
13491bc56edSDimitry Andric   CallGraphWrapperPass *CGPass = getAnalysisIfAvailable<CallGraphWrapperPass>();
13591bc56edSDimitry Andric   CallGraph *CG = CGPass ? &CGPass->getCallGraph() : nullptr;
13691bc56edSDimitry Andric   CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : nullptr;
137f22ef01cSRoman Divacky   bool Changed = false;
138f22ef01cSRoman Divacky 
139f785676fSDimitry Andric   SmallPtrSet<GlobalValue *, 8> Used;
140f785676fSDimitry Andric   collectUsedGlobalVariables(M, Used, false);
141f785676fSDimitry Andric 
142f785676fSDimitry Andric   // We must assume that globals in llvm.used have a reference that not even
143f785676fSDimitry Andric   // the linker can see, so we don't internalize them.
144f785676fSDimitry Andric   // For llvm.compiler.used the situation is a bit fuzzy. The assembler and
145f785676fSDimitry Andric   // linker can drop those symbols. If this pass is running as part of LTO,
146f785676fSDimitry Andric   // one might think that it could just drop llvm.compiler.used. The problem
147f785676fSDimitry Andric   // is that even in LTO llvm doesn't see every reference. For example,
148f785676fSDimitry Andric   // we don't see references from function local inline assembly. To be
149f785676fSDimitry Andric   // conservative, we internalize symbols in llvm.compiler.used, but we
150f785676fSDimitry Andric   // keep llvm.compiler.used so that the symbol is not deleted by llvm.
151f785676fSDimitry Andric   for (SmallPtrSet<GlobalValue *, 8>::iterator I = Used.begin(), E = Used.end();
152f785676fSDimitry Andric        I != E; ++I) {
153f785676fSDimitry Andric     GlobalValue *V = *I;
154f785676fSDimitry Andric     ExternalNames.insert(V->getName());
155f785676fSDimitry Andric   }
156dff0c46cSDimitry Andric 
157f22ef01cSRoman Divacky   // Mark all functions not in the api as internal.
158f785676fSDimitry Andric   for (Module::iterator I = M.begin(), E = M.end(); I != E; ++I) {
159f785676fSDimitry Andric     if (!shouldInternalize(*I, ExternalNames))
160f785676fSDimitry Andric       continue;
161f785676fSDimitry Andric 
16291bc56edSDimitry Andric     I->setVisibility(GlobalValue::DefaultVisibility);
163f22ef01cSRoman Divacky     I->setLinkage(GlobalValue::InternalLinkage);
164f785676fSDimitry Andric 
165f785676fSDimitry Andric     if (ExternalNode)
166f22ef01cSRoman Divacky       // Remove a callgraph edge from the external node to this function.
167f785676fSDimitry Andric       ExternalNode->removeOneAbstractEdgeTo((*CG)[I]);
168f785676fSDimitry Andric 
169f22ef01cSRoman Divacky     Changed = true;
170f22ef01cSRoman Divacky     ++NumFunctions;
171f22ef01cSRoman Divacky     DEBUG(dbgs() << "Internalizing func " << I->getName() << "\n");
172f22ef01cSRoman Divacky   }
173f22ef01cSRoman Divacky 
174f22ef01cSRoman Divacky   // Never internalize the llvm.used symbol.  It is used to implement
175f22ef01cSRoman Divacky   // attribute((used)).
176f22ef01cSRoman Divacky   // FIXME: Shouldn't this just filter on llvm.metadata section??
177f22ef01cSRoman Divacky   ExternalNames.insert("llvm.used");
178f22ef01cSRoman Divacky   ExternalNames.insert("llvm.compiler.used");
179f22ef01cSRoman Divacky 
180f22ef01cSRoman Divacky   // Never internalize anchors used by the machine module info, else the info
181f22ef01cSRoman Divacky   // won't find them.  (see MachineModuleInfo.)
182f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global_ctors");
183f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global_dtors");
184f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global.annotations");
185f22ef01cSRoman Divacky 
186dff0c46cSDimitry Andric   // Never internalize symbols code-gen inserts.
187f785676fSDimitry Andric   // FIXME: We should probably add this (and the __stack_chk_guard) via some
188f785676fSDimitry Andric   // type of call-back in CodeGen.
189f785676fSDimitry Andric   ExternalNames.insert("__stack_chk_fail");
190dff0c46cSDimitry Andric   ExternalNames.insert("__stack_chk_guard");
191dff0c46cSDimitry Andric 
192f22ef01cSRoman Divacky   // Mark all global variables with initializers that are not in the api as
193f22ef01cSRoman Divacky   // internal as well.
194f22ef01cSRoman Divacky   for (Module::global_iterator I = M.global_begin(), E = M.global_end();
195f785676fSDimitry Andric        I != E; ++I) {
196f785676fSDimitry Andric     if (!shouldInternalize(*I, ExternalNames))
197f785676fSDimitry Andric       continue;
198f785676fSDimitry Andric 
19991bc56edSDimitry Andric     I->setVisibility(GlobalValue::DefaultVisibility);
200f22ef01cSRoman Divacky     I->setLinkage(GlobalValue::InternalLinkage);
201f22ef01cSRoman Divacky     Changed = true;
202f22ef01cSRoman Divacky     ++NumGlobals;
203f22ef01cSRoman Divacky     DEBUG(dbgs() << "Internalized gvar " << I->getName() << "\n");
204f22ef01cSRoman Divacky   }
205f22ef01cSRoman Divacky 
206f22ef01cSRoman Divacky   // Mark all aliases that are not in the api as internal as well.
207f22ef01cSRoman Divacky   for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
208f785676fSDimitry Andric        I != E; ++I) {
209f785676fSDimitry Andric     if (!shouldInternalize(*I, ExternalNames))
210f785676fSDimitry Andric       continue;
211f785676fSDimitry Andric 
21291bc56edSDimitry Andric     I->setVisibility(GlobalValue::DefaultVisibility);
213f22ef01cSRoman Divacky     I->setLinkage(GlobalValue::InternalLinkage);
214f22ef01cSRoman Divacky     Changed = true;
215f22ef01cSRoman Divacky     ++NumAliases;
216f22ef01cSRoman Divacky     DEBUG(dbgs() << "Internalized alias " << I->getName() << "\n");
217f22ef01cSRoman Divacky   }
218f22ef01cSRoman Divacky 
219f22ef01cSRoman Divacky   return Changed;
220f22ef01cSRoman Divacky }
221f22ef01cSRoman Divacky 
22291bc56edSDimitry Andric ModulePass *llvm::createInternalizePass() { return new InternalizePass(); }
223f22ef01cSRoman Divacky 
224f785676fSDimitry Andric ModulePass *llvm::createInternalizePass(ArrayRef<const char *> ExportList) {
225f785676fSDimitry Andric   return new InternalizePass(ExportList);
226f22ef01cSRoman Divacky }
227