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);
637d523365SDimitry Andric     bool maybeInternalize(GlobalValue &GV,
647d523365SDimitry Andric                           const std::set<const Comdat *> &ExternalComdats);
657d523365SDimitry Andric     void checkComdatVisibility(GlobalValue &GV,
667d523365SDimitry Andric                                std::set<const Comdat *> &ExternalComdats);
6791bc56edSDimitry Andric     bool runOnModule(Module &M) override;
68f22ef01cSRoman Divacky 
6991bc56edSDimitry Andric     void getAnalysisUsage(AnalysisUsage &AU) const override {
70f22ef01cSRoman Divacky       AU.setPreservesCFG();
7191bc56edSDimitry Andric       AU.addPreserved<CallGraphWrapperPass>();
72f22ef01cSRoman Divacky     }
73f22ef01cSRoman Divacky   };
74f22ef01cSRoman Divacky } // end anonymous namespace
75f22ef01cSRoman Divacky 
76f22ef01cSRoman Divacky char InternalizePass::ID = 0;
77e580952dSDimitry Andric INITIALIZE_PASS(InternalizePass, "internalize",
782754fe60SDimitry Andric                 "Internalize Global Symbols", false, false)
79f22ef01cSRoman Divacky 
8091bc56edSDimitry Andric InternalizePass::InternalizePass() : ModulePass(ID) {
812754fe60SDimitry Andric   initializeInternalizePassPass(*PassRegistry::getPassRegistry());
82f22ef01cSRoman Divacky   if (!APIFile.empty())           // If a filename is specified, use it.
83f22ef01cSRoman Divacky     LoadFile(APIFile.c_str());
84f22ef01cSRoman Divacky   ExternalNames.insert(APIList.begin(), APIList.end());
85f22ef01cSRoman Divacky }
86f22ef01cSRoman Divacky 
87f785676fSDimitry Andric InternalizePass::InternalizePass(ArrayRef<const char *> ExportList)
883861d79fSDimitry Andric     : ModulePass(ID) {
892754fe60SDimitry Andric   initializeInternalizePassPass(*PassRegistry::getPassRegistry());
90f785676fSDimitry Andric   for(ArrayRef<const char *>::const_iterator itr = ExportList.begin();
91f785676fSDimitry Andric         itr != ExportList.end(); itr++) {
92f22ef01cSRoman Divacky     ExternalNames.insert(*itr);
93f22ef01cSRoman Divacky   }
94f22ef01cSRoman Divacky }
95f22ef01cSRoman Divacky 
96f22ef01cSRoman Divacky void InternalizePass::LoadFile(const char *Filename) {
97f22ef01cSRoman Divacky   // Load the APIFile...
98f22ef01cSRoman Divacky   std::ifstream In(Filename);
99f22ef01cSRoman Divacky   if (!In.good()) {
100f22ef01cSRoman Divacky     errs() << "WARNING: Internalize couldn't load file '" << Filename
101f22ef01cSRoman Divacky          << "'! Continuing as if it's empty.\n";
102f22ef01cSRoman Divacky     return; // Just continue as if the file were empty
103f22ef01cSRoman Divacky   }
104f22ef01cSRoman Divacky   while (In) {
105f22ef01cSRoman Divacky     std::string Symbol;
106f22ef01cSRoman Divacky     In >> Symbol;
107f22ef01cSRoman Divacky     if (!Symbol.empty())
108f22ef01cSRoman Divacky       ExternalNames.insert(Symbol);
109f22ef01cSRoman Divacky   }
110f22ef01cSRoman Divacky }
111f22ef01cSRoman Divacky 
1127d523365SDimitry Andric static bool isExternallyVisible(const GlobalValue &GV,
113f785676fSDimitry Andric                                 const std::set<std::string> &ExternalNames) {
114f785676fSDimitry Andric   // Function must be defined here
115f785676fSDimitry Andric   if (GV.isDeclaration())
1167d523365SDimitry Andric     return true;
117139f7f9bSDimitry Andric 
118f785676fSDimitry Andric   // Available externally is really just a "declaration with a body".
119f785676fSDimitry Andric   if (GV.hasAvailableExternallyLinkage())
1207d523365SDimitry Andric     return true;
121f785676fSDimitry Andric 
12291bc56edSDimitry Andric   // Assume that dllexported symbols are referenced elsewhere
12391bc56edSDimitry Andric   if (GV.hasDLLExportStorageClass())
1247d523365SDimitry Andric     return true;
1257d523365SDimitry Andric 
1267d523365SDimitry Andric   // Marked to keep external?
1277d523365SDimitry Andric   if (!GV.hasLocalLinkage() && ExternalNames.count(GV.getName()))
1287d523365SDimitry Andric     return true;
1297d523365SDimitry Andric 
1307d523365SDimitry Andric   return false;
1317d523365SDimitry Andric }
1327d523365SDimitry Andric 
1337d523365SDimitry Andric // Internalize GV if it is possible to do so, i.e. it is not externally visible
1347d523365SDimitry Andric // and is not a member of an externally visible comdat.
1357d523365SDimitry Andric bool InternalizePass::maybeInternalize(
1367d523365SDimitry Andric     GlobalValue &GV, const std::set<const Comdat *> &ExternalComdats) {
1377d523365SDimitry Andric   if (Comdat *C = GV.getComdat()) {
1387d523365SDimitry Andric     if (ExternalComdats.count(C))
13991bc56edSDimitry Andric       return false;
14091bc56edSDimitry Andric 
1417d523365SDimitry Andric     // If a comdat is not externally visible we can drop it.
1427d523365SDimitry Andric     if (auto GO = dyn_cast<GlobalObject>(&GV))
1437d523365SDimitry Andric       GO->setComdat(nullptr);
1447d523365SDimitry Andric 
1457d523365SDimitry Andric     if (GV.hasLocalLinkage())
1467d523365SDimitry Andric       return false;
1477d523365SDimitry Andric   } else {
148f785676fSDimitry Andric     if (GV.hasLocalLinkage())
149f785676fSDimitry Andric       return false;
150f785676fSDimitry Andric 
1517d523365SDimitry Andric     if (isExternallyVisible(GV, ExternalNames))
152f785676fSDimitry Andric       return false;
1537d523365SDimitry Andric   }
154f785676fSDimitry Andric 
1557d523365SDimitry Andric   GV.setVisibility(GlobalValue::DefaultVisibility);
1567d523365SDimitry Andric   GV.setLinkage(GlobalValue::InternalLinkage);
157f785676fSDimitry Andric   return true;
158139f7f9bSDimitry Andric }
159139f7f9bSDimitry Andric 
1607d523365SDimitry Andric // If GV is part of a comdat and is externally visible, keep track of its
1617d523365SDimitry Andric // comdat so that we don't internalize any of its members.
1627d523365SDimitry Andric void InternalizePass::checkComdatVisibility(
1637d523365SDimitry Andric     GlobalValue &GV, std::set<const Comdat *> &ExternalComdats) {
1647d523365SDimitry Andric   Comdat *C = GV.getComdat();
1657d523365SDimitry Andric   if (!C)
1667d523365SDimitry Andric     return;
1677d523365SDimitry Andric 
1687d523365SDimitry Andric   if (isExternallyVisible(GV, ExternalNames))
1697d523365SDimitry Andric     ExternalComdats.insert(C);
1707d523365SDimitry Andric }
1717d523365SDimitry Andric 
172f22ef01cSRoman Divacky bool InternalizePass::runOnModule(Module &M) {
17391bc56edSDimitry Andric   CallGraphWrapperPass *CGPass = getAnalysisIfAvailable<CallGraphWrapperPass>();
17491bc56edSDimitry Andric   CallGraph *CG = CGPass ? &CGPass->getCallGraph() : nullptr;
17591bc56edSDimitry Andric   CallGraphNode *ExternalNode = CG ? CG->getExternalCallingNode() : nullptr;
176f22ef01cSRoman Divacky 
177f785676fSDimitry Andric   SmallPtrSet<GlobalValue *, 8> Used;
178f785676fSDimitry Andric   collectUsedGlobalVariables(M, Used, false);
179f785676fSDimitry Andric 
1807d523365SDimitry Andric   // Collect comdat visiblity information for the module.
1817d523365SDimitry Andric   std::set<const Comdat *> ExternalComdats;
1827d523365SDimitry Andric   if (!M.getComdatSymbolTable().empty()) {
1837d523365SDimitry Andric     for (Function &F : M)
1847d523365SDimitry Andric       checkComdatVisibility(F, ExternalComdats);
1857d523365SDimitry Andric     for (GlobalVariable &GV : M.globals())
1867d523365SDimitry Andric       checkComdatVisibility(GV, ExternalComdats);
1877d523365SDimitry Andric     for (GlobalAlias &GA : M.aliases())
1887d523365SDimitry Andric       checkComdatVisibility(GA, ExternalComdats);
1897d523365SDimitry Andric   }
1907d523365SDimitry Andric 
191f785676fSDimitry Andric   // We must assume that globals in llvm.used have a reference that not even
192f785676fSDimitry Andric   // the linker can see, so we don't internalize them.
193f785676fSDimitry Andric   // For llvm.compiler.used the situation is a bit fuzzy. The assembler and
194f785676fSDimitry Andric   // linker can drop those symbols. If this pass is running as part of LTO,
195f785676fSDimitry Andric   // one might think that it could just drop llvm.compiler.used. The problem
196f785676fSDimitry Andric   // is that even in LTO llvm doesn't see every reference. For example,
197f785676fSDimitry Andric   // we don't see references from function local inline assembly. To be
198f785676fSDimitry Andric   // conservative, we internalize symbols in llvm.compiler.used, but we
199f785676fSDimitry Andric   // keep llvm.compiler.used so that the symbol is not deleted by llvm.
20039d628a0SDimitry Andric   for (GlobalValue *V : Used) {
201f785676fSDimitry Andric     ExternalNames.insert(V->getName());
202f785676fSDimitry Andric   }
203dff0c46cSDimitry Andric 
204f22ef01cSRoman Divacky   // Mark all functions not in the api as internal.
2057d523365SDimitry Andric   for (Function &I : M) {
2067d523365SDimitry Andric     if (!maybeInternalize(I, ExternalComdats))
207f785676fSDimitry Andric       continue;
208f785676fSDimitry Andric 
209f785676fSDimitry Andric     if (ExternalNode)
210f22ef01cSRoman Divacky       // Remove a callgraph edge from the external node to this function.
2117d523365SDimitry Andric       ExternalNode->removeOneAbstractEdgeTo((*CG)[&I]);
212f785676fSDimitry Andric 
213f22ef01cSRoman Divacky     ++NumFunctions;
2147d523365SDimitry Andric     DEBUG(dbgs() << "Internalizing func " << I.getName() << "\n");
215f22ef01cSRoman Divacky   }
216f22ef01cSRoman Divacky 
217f22ef01cSRoman Divacky   // Never internalize the llvm.used symbol.  It is used to implement
218f22ef01cSRoman Divacky   // attribute((used)).
219f22ef01cSRoman Divacky   // FIXME: Shouldn't this just filter on llvm.metadata section??
220f22ef01cSRoman Divacky   ExternalNames.insert("llvm.used");
221f22ef01cSRoman Divacky   ExternalNames.insert("llvm.compiler.used");
222f22ef01cSRoman Divacky 
223f22ef01cSRoman Divacky   // Never internalize anchors used by the machine module info, else the info
224f22ef01cSRoman Divacky   // won't find them.  (see MachineModuleInfo.)
225f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global_ctors");
226f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global_dtors");
227f22ef01cSRoman Divacky   ExternalNames.insert("llvm.global.annotations");
228f22ef01cSRoman Divacky 
229dff0c46cSDimitry Andric   // Never internalize symbols code-gen inserts.
230f785676fSDimitry Andric   // FIXME: We should probably add this (and the __stack_chk_guard) via some
231f785676fSDimitry Andric   // type of call-back in CodeGen.
232f785676fSDimitry Andric   ExternalNames.insert("__stack_chk_fail");
233dff0c46cSDimitry Andric   ExternalNames.insert("__stack_chk_guard");
234dff0c46cSDimitry Andric 
235f22ef01cSRoman Divacky   // Mark all global variables with initializers that are not in the api as
236f22ef01cSRoman Divacky   // internal as well.
237f22ef01cSRoman Divacky   for (Module::global_iterator I = M.global_begin(), E = M.global_end();
238f785676fSDimitry Andric        I != E; ++I) {
2397d523365SDimitry Andric     if (!maybeInternalize(*I, ExternalComdats))
240f785676fSDimitry Andric       continue;
241f785676fSDimitry Andric 
242f22ef01cSRoman Divacky     ++NumGlobals;
243f22ef01cSRoman Divacky     DEBUG(dbgs() << "Internalized gvar " << I->getName() << "\n");
244f22ef01cSRoman Divacky   }
245f22ef01cSRoman Divacky 
246f22ef01cSRoman Divacky   // Mark all aliases that are not in the api as internal as well.
247f22ef01cSRoman Divacky   for (Module::alias_iterator I = M.alias_begin(), E = M.alias_end();
248f785676fSDimitry Andric        I != E; ++I) {
2497d523365SDimitry Andric     if (!maybeInternalize(*I, ExternalComdats))
250f785676fSDimitry Andric       continue;
251f785676fSDimitry Andric 
252f22ef01cSRoman Divacky     ++NumAliases;
253f22ef01cSRoman Divacky     DEBUG(dbgs() << "Internalized alias " << I->getName() << "\n");
254f22ef01cSRoman Divacky   }
255f22ef01cSRoman Divacky 
2567d523365SDimitry Andric   // We do not keep track of whether this pass changed the module because
2577d523365SDimitry Andric   // it adds unnecessary complexity:
2587d523365SDimitry Andric   // 1) This pass will generally be near the start of the pass pipeline, so
2597d523365SDimitry Andric   //    there will be no analyses to invalidate.
2607d523365SDimitry Andric   // 2) This pass will most likely end up changing the module and it isn't worth
2617d523365SDimitry Andric   //    worrying about optimizing the case where the module is unchanged.
2627d523365SDimitry Andric   return true;
263f22ef01cSRoman Divacky }
264f22ef01cSRoman Divacky 
26591bc56edSDimitry Andric ModulePass *llvm::createInternalizePass() { return new InternalizePass(); }
266f22ef01cSRoman Divacky 
267f785676fSDimitry Andric ModulePass *llvm::createInternalizePass(ArrayRef<const char *> ExportList) {
268f785676fSDimitry Andric   return new InternalizePass(ExportList);
269f22ef01cSRoman Divacky }
270