1 //===-- ModuleUtils.cpp - Functions to manipulate Modules -----------------===//
2 //
3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
4 // See https://llvm.org/LICENSE.txt for license information.
5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
6 //
7 //===----------------------------------------------------------------------===//
8 //
9 // This family of functions perform manipulations on Modules.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #include "llvm/Transforms/Utils/ModuleUtils.h"
14 #include "llvm/Analysis/TargetLibraryInfo.h"
15 #include "llvm/Analysis/VectorUtils.h"
16 #include "llvm/IR/DerivedTypes.h"
17 #include "llvm/IR/Function.h"
18 #include "llvm/IR/IRBuilder.h"
19 #include "llvm/IR/Module.h"
20 #include "llvm/Support/raw_ostream.h"
21 using namespace llvm;
22 
23 #define DEBUG_TYPE "moduleutils"
24 
25 static void appendToGlobalArray(const char *Array, Module &M, Function *F,
26                                 int Priority, Constant *Data) {
27   IRBuilder<> IRB(M.getContext());
28   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
29 
30   // Get the current set of static global constructors and add the new ctor
31   // to the list.
32   SmallVector<Constant *, 16> CurrentCtors;
33   StructType *EltTy = StructType::get(
34       IRB.getInt32Ty(), PointerType::getUnqual(FnTy), IRB.getInt8PtrTy());
35   if (GlobalVariable *GVCtor = M.getNamedGlobal(Array)) {
36     if (Constant *Init = GVCtor->getInitializer()) {
37       unsigned n = Init->getNumOperands();
38       CurrentCtors.reserve(n + 1);
39       for (unsigned i = 0; i != n; ++i)
40         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
41     }
42     GVCtor->eraseFromParent();
43   }
44 
45   // Build a 3 field global_ctor entry.  We don't take a comdat key.
46   Constant *CSVals[3];
47   CSVals[0] = IRB.getInt32(Priority);
48   CSVals[1] = F;
49   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
50                    : Constant::getNullValue(IRB.getInt8PtrTy());
51   Constant *RuntimeCtorInit =
52       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
53 
54   CurrentCtors.push_back(RuntimeCtorInit);
55 
56   // Create a new initializer.
57   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
58   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
59 
60   // Create the new global variable and replace all uses of
61   // the old global variable with the new one.
62   (void)new GlobalVariable(M, NewInit->getType(), false,
63                            GlobalValue::AppendingLinkage, NewInit, Array);
64 }
65 
66 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
67   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
68 }
69 
70 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
71   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
72 }
73 
74 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
75   GlobalVariable *GV = M.getGlobalVariable(Name);
76   SmallPtrSet<Constant *, 16> InitAsSet;
77   SmallVector<Constant *, 16> Init;
78   if (GV) {
79     auto *CA = cast<ConstantArray>(GV->getInitializer());
80     for (auto &Op : CA->operands()) {
81       Constant *C = cast_or_null<Constant>(Op);
82       if (InitAsSet.insert(C).second)
83         Init.push_back(C);
84     }
85     GV->eraseFromParent();
86   }
87 
88   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
89   for (auto *V : Values) {
90     Constant *C = ConstantExpr::getBitCast(V, Int8PtrTy);
91     if (InitAsSet.insert(C).second)
92       Init.push_back(C);
93   }
94 
95   if (Init.empty())
96     return;
97 
98   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
99   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
100                                 ConstantArray::get(ATy, Init), Name);
101   GV->setSection("llvm.metadata");
102 }
103 
104 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
105   appendToUsedList(M, "llvm.used", Values);
106 }
107 
108 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
109   appendToUsedList(M, "llvm.compiler.used", Values);
110 }
111 
112 FunctionCallee
113 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
114                                    ArrayRef<Type *> InitArgTypes) {
115   assert(!InitName.empty() && "Expected init function name");
116   return M.getOrInsertFunction(
117       InitName,
118       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
119       AttributeList());
120 }
121 
122 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
123     Module &M, StringRef CtorName, StringRef InitName,
124     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
125     StringRef VersionCheckName) {
126   assert(!InitName.empty() && "Expected init function name");
127   assert(InitArgs.size() == InitArgTypes.size() &&
128          "Sanitizer's init function expects different number of arguments");
129   FunctionCallee InitFunction =
130       declareSanitizerInitFunction(M, InitName, InitArgTypes);
131   Function *Ctor = Function::Create(
132       FunctionType::get(Type::getVoidTy(M.getContext()), false),
133       GlobalValue::InternalLinkage, CtorName, &M);
134   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
135   IRBuilder<> IRB(ReturnInst::Create(M.getContext(), CtorBB));
136   IRB.CreateCall(InitFunction, InitArgs);
137   if (!VersionCheckName.empty()) {
138     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
139         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
140         AttributeList());
141     IRB.CreateCall(VersionCheckFunction, {});
142   }
143   return std::make_pair(Ctor, InitFunction);
144 }
145 
146 std::pair<Function *, FunctionCallee>
147 llvm::getOrCreateSanitizerCtorAndInitFunctions(
148     Module &M, StringRef CtorName, StringRef InitName,
149     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
150     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
151     StringRef VersionCheckName) {
152   assert(!CtorName.empty() && "Expected ctor function name");
153 
154   if (Function *Ctor = M.getFunction(CtorName))
155     // FIXME: Sink this logic into the module, similar to the handling of
156     // globals. This will make moving to a concurrent model much easier.
157     if (Ctor->arg_size() == 0 ||
158         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
159       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
160 
161   Function *Ctor;
162   FunctionCallee InitFunction;
163   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
164       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
165   FunctionsCreatedCallback(Ctor, InitFunction);
166   return std::make_pair(Ctor, InitFunction);
167 }
168 
169 Function *llvm::getOrCreateInitFunction(Module &M, StringRef Name) {
170   assert(!Name.empty() && "Expected init function name");
171   if (Function *F = M.getFunction(Name)) {
172     if (F->arg_size() != 0 ||
173         F->getReturnType() != Type::getVoidTy(M.getContext())) {
174       std::string Err;
175       raw_string_ostream Stream(Err);
176       Stream << "Sanitizer interface function defined with wrong type: " << *F;
177       report_fatal_error(Err);
178     }
179     return F;
180   }
181   Function *F =
182       cast<Function>(M.getOrInsertFunction(Name, AttributeList(),
183                                            Type::getVoidTy(M.getContext()))
184                          .getCallee());
185 
186   appendToGlobalCtors(M, F, 0);
187 
188   return F;
189 }
190 
191 void llvm::filterDeadComdatFunctions(
192     Module &M, SmallVectorImpl<Function *> &DeadComdatFunctions) {
193   // Build a map from the comdat to the number of entries in that comdat we
194   // think are dead. If this fully covers the comdat group, then the entire
195   // group is dead. If we find another entry in the comdat group though, we'll
196   // have to preserve the whole group.
197   SmallDenseMap<Comdat *, int, 16> ComdatEntriesCovered;
198   for (Function *F : DeadComdatFunctions) {
199     Comdat *C = F->getComdat();
200     assert(C && "Expected all input GVs to be in a comdat!");
201     ComdatEntriesCovered[C] += 1;
202   }
203 
204   auto CheckComdat = [&](Comdat &C) {
205     auto CI = ComdatEntriesCovered.find(&C);
206     if (CI == ComdatEntriesCovered.end())
207       return;
208 
209     // If this could have been covered by a dead entry, just subtract one to
210     // account for it.
211     if (CI->second > 0) {
212       CI->second -= 1;
213       return;
214     }
215 
216     // If we've already accounted for all the entries that were dead, the
217     // entire comdat is alive so remove it from the map.
218     ComdatEntriesCovered.erase(CI);
219   };
220 
221   auto CheckAllComdats = [&] {
222     for (Function &F : M.functions())
223       if (Comdat *C = F.getComdat()) {
224         CheckComdat(*C);
225         if (ComdatEntriesCovered.empty())
226           return;
227       }
228     for (GlobalVariable &GV : M.globals())
229       if (Comdat *C = GV.getComdat()) {
230         CheckComdat(*C);
231         if (ComdatEntriesCovered.empty())
232           return;
233       }
234     for (GlobalAlias &GA : M.aliases())
235       if (Comdat *C = GA.getComdat()) {
236         CheckComdat(*C);
237         if (ComdatEntriesCovered.empty())
238           return;
239       }
240   };
241   CheckAllComdats();
242 
243   if (ComdatEntriesCovered.empty()) {
244     DeadComdatFunctions.clear();
245     return;
246   }
247 
248   // Remove the entries that were not covering.
249   erase_if(DeadComdatFunctions, [&](GlobalValue *GV) {
250     return ComdatEntriesCovered.find(GV->getComdat()) ==
251            ComdatEntriesCovered.end();
252   });
253 }
254 
255 std::string llvm::getUniqueModuleId(Module *M) {
256   MD5 Md5;
257   bool ExportsSymbols = false;
258   auto AddGlobal = [&](GlobalValue &GV) {
259     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
260         !GV.hasExternalLinkage() || GV.hasComdat())
261       return;
262     ExportsSymbols = true;
263     Md5.update(GV.getName());
264     Md5.update(ArrayRef<uint8_t>{0});
265   };
266 
267   for (auto &F : *M)
268     AddGlobal(F);
269   for (auto &GV : M->globals())
270     AddGlobal(GV);
271   for (auto &GA : M->aliases())
272     AddGlobal(GA);
273   for (auto &IF : M->ifuncs())
274     AddGlobal(IF);
275 
276   if (!ExportsSymbols)
277     return "";
278 
279   MD5::MD5Result R;
280   Md5.final(R);
281 
282   SmallString<32> Str;
283   MD5::stringifyResult(R, Str);
284   return ("$" + Str).str();
285 }
286 
287 void VFABI::setVectorVariantNames(
288     CallInst *CI, const SmallVector<std::string, 8> &VariantMappings) {
289   if (VariantMappings.empty())
290     return;
291 
292   SmallString<256> Buffer;
293   llvm::raw_svector_ostream Out(Buffer);
294   for (const std::string &VariantMapping : VariantMappings)
295     Out << VariantMapping << ",";
296   // Get rid of the trailing ','.
297   assert(!Buffer.str().empty() && "Must have at least one char.");
298   Buffer.pop_back();
299 
300   Module *M = CI->getModule();
301 #ifndef NDEBUG
302   for (const std::string &VariantMapping : VariantMappings) {
303     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
304     Optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
305     assert(VI.hasValue() && "Cannot add an invalid VFABI name.");
306     assert(M->getNamedValue(VI.getValue().VectorName) &&
307            "Cannot add variant to attribute: "
308            "vector function declaration is missing.");
309   }
310 #endif
311   CI->addAttribute(
312       AttributeList::FunctionIndex,
313       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
314 }
315