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/VectorUtils.h"
15 #include "llvm/IR/DerivedTypes.h"
16 #include "llvm/IR/Function.h"
17 #include "llvm/IR/IRBuilder.h"
18 #include "llvm/IR/Module.h"
19 #include "llvm/Support/raw_ostream.h"
20 using namespace llvm;
21 
22 #define DEBUG_TYPE "moduleutils"
23 
24 static void appendToGlobalArray(const char *Array, Module &M, Function *F,
25                                 int Priority, Constant *Data) {
26   IRBuilder<> IRB(M.getContext());
27   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
28 
29   // Get the current set of static global constructors and add the new ctor
30   // to the list.
31   SmallVector<Constant *, 16> CurrentCtors;
32   StructType *EltTy = StructType::get(
33       IRB.getInt32Ty(), PointerType::getUnqual(FnTy), IRB.getInt8PtrTy());
34   if (GlobalVariable *GVCtor = M.getNamedGlobal(Array)) {
35     if (Constant *Init = GVCtor->getInitializer()) {
36       unsigned n = Init->getNumOperands();
37       CurrentCtors.reserve(n + 1);
38       for (unsigned i = 0; i != n; ++i)
39         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
40     }
41     GVCtor->eraseFromParent();
42   }
43 
44   // Build a 3 field global_ctor entry.  We don't take a comdat key.
45   Constant *CSVals[3];
46   CSVals[0] = IRB.getInt32(Priority);
47   CSVals[1] = F;
48   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
49                    : Constant::getNullValue(IRB.getInt8PtrTy());
50   Constant *RuntimeCtorInit =
51       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
52 
53   CurrentCtors.push_back(RuntimeCtorInit);
54 
55   // Create a new initializer.
56   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
57   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
58 
59   // Create the new global variable and replace all uses of
60   // the old global variable with the new one.
61   (void)new GlobalVariable(M, NewInit->getType(), false,
62                            GlobalValue::AppendingLinkage, NewInit, Array);
63 }
64 
65 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
66   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
67 }
68 
69 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
70   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
71 }
72 
73 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
74   GlobalVariable *GV = M.getGlobalVariable(Name);
75   SmallPtrSet<Constant *, 16> InitAsSet;
76   SmallVector<Constant *, 16> Init;
77   if (GV) {
78     if (GV->hasInitializer()) {
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     }
86     GV->eraseFromParent();
87   }
88 
89   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
90   for (auto *V : Values) {
91     Constant *C = ConstantExpr::getPointerBitCastOrAddrSpaceCast(V, Int8PtrTy);
92     if (InitAsSet.insert(C).second)
93       Init.push_back(C);
94   }
95 
96   if (Init.empty())
97     return;
98 
99   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
100   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
101                                 ConstantArray::get(ATy, Init), Name);
102   GV->setSection("llvm.metadata");
103 }
104 
105 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
106   appendToUsedList(M, "llvm.used", Values);
107 }
108 
109 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
110   appendToUsedList(M, "llvm.compiler.used", Values);
111 }
112 
113 FunctionCallee
114 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
115                                    ArrayRef<Type *> InitArgTypes) {
116   assert(!InitName.empty() && "Expected init function name");
117   return M.getOrInsertFunction(
118       InitName,
119       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
120       AttributeList());
121 }
122 
123 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
124   Function *Ctor = Function::createWithDefaultAttr(
125       FunctionType::get(Type::getVoidTy(M.getContext()), false),
126       GlobalValue::InternalLinkage, 0, CtorName, &M);
127   Ctor->addFnAttr(Attribute::NoUnwind);
128   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
129   ReturnInst::Create(M.getContext(), CtorBB);
130   // Ensure Ctor cannot be discarded, even if in a comdat.
131   appendToUsed(M, {Ctor});
132   return Ctor;
133 }
134 
135 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
136     Module &M, StringRef CtorName, StringRef InitName,
137     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
138     StringRef VersionCheckName) {
139   assert(!InitName.empty() && "Expected init function name");
140   assert(InitArgs.size() == InitArgTypes.size() &&
141          "Sanitizer's init function expects different number of arguments");
142   FunctionCallee InitFunction =
143       declareSanitizerInitFunction(M, InitName, InitArgTypes);
144   Function *Ctor = createSanitizerCtor(M, CtorName);
145   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
146   IRB.CreateCall(InitFunction, InitArgs);
147   if (!VersionCheckName.empty()) {
148     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
149         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
150         AttributeList());
151     IRB.CreateCall(VersionCheckFunction, {});
152   }
153   return std::make_pair(Ctor, InitFunction);
154 }
155 
156 std::pair<Function *, FunctionCallee>
157 llvm::getOrCreateSanitizerCtorAndInitFunctions(
158     Module &M, StringRef CtorName, StringRef InitName,
159     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
160     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
161     StringRef VersionCheckName) {
162   assert(!CtorName.empty() && "Expected ctor function name");
163 
164   if (Function *Ctor = M.getFunction(CtorName))
165     // FIXME: Sink this logic into the module, similar to the handling of
166     // globals. This will make moving to a concurrent model much easier.
167     if (Ctor->arg_empty() ||
168         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
169       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
170 
171   Function *Ctor;
172   FunctionCallee InitFunction;
173   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
174       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
175   FunctionsCreatedCallback(Ctor, InitFunction);
176   return std::make_pair(Ctor, InitFunction);
177 }
178 
179 void llvm::filterDeadComdatFunctions(
180     SmallVectorImpl<Function *> &DeadComdatFunctions) {
181   SmallPtrSet<Function *, 32> MaybeDeadFunctions;
182   SmallPtrSet<Comdat *, 32> MaybeDeadComdats;
183   for (Function *F : DeadComdatFunctions) {
184     MaybeDeadFunctions.insert(F);
185     if (Comdat *C = F->getComdat())
186       MaybeDeadComdats.insert(C);
187   }
188 
189   // Find comdats for which all users are dead now.
190   SmallPtrSet<Comdat *, 32> DeadComdats;
191   for (Comdat *C : MaybeDeadComdats) {
192     auto IsUserDead = [&](GlobalObject *GO) {
193       auto *F = dyn_cast<Function>(GO);
194       return F && MaybeDeadFunctions.contains(F);
195     };
196     if (all_of(C->getUsers(), IsUserDead))
197       DeadComdats.insert(C);
198   }
199 
200   // Only keep functions which have no comdat or a dead comdat.
201   erase_if(DeadComdatFunctions, [&](Function *F) {
202     Comdat *C = F->getComdat();
203     return C && !DeadComdats.contains(C);
204   });
205 }
206 
207 std::string llvm::getUniqueModuleId(Module *M) {
208   MD5 Md5;
209   bool ExportsSymbols = false;
210   auto AddGlobal = [&](GlobalValue &GV) {
211     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
212         !GV.hasExternalLinkage() || GV.hasComdat())
213       return;
214     ExportsSymbols = true;
215     Md5.update(GV.getName());
216     Md5.update(ArrayRef<uint8_t>{0});
217   };
218 
219   for (auto &F : *M)
220     AddGlobal(F);
221   for (auto &GV : M->globals())
222     AddGlobal(GV);
223   for (auto &GA : M->aliases())
224     AddGlobal(GA);
225   for (auto &IF : M->ifuncs())
226     AddGlobal(IF);
227 
228   if (!ExportsSymbols)
229     return "";
230 
231   MD5::MD5Result R;
232   Md5.final(R);
233 
234   SmallString<32> Str;
235   MD5::stringifyResult(R, Str);
236   return ("." + Str).str();
237 }
238 
239 void VFABI::setVectorVariantNames(CallInst *CI,
240                                   ArrayRef<std::string> VariantMappings) {
241   if (VariantMappings.empty())
242     return;
243 
244   SmallString<256> Buffer;
245   llvm::raw_svector_ostream Out(Buffer);
246   for (const std::string &VariantMapping : VariantMappings)
247     Out << VariantMapping << ",";
248   // Get rid of the trailing ','.
249   assert(!Buffer.str().empty() && "Must have at least one char.");
250   Buffer.pop_back();
251 
252   Module *M = CI->getModule();
253 #ifndef NDEBUG
254   for (const std::string &VariantMapping : VariantMappings) {
255     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
256     Optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
257     assert(VI && "Cannot add an invalid VFABI name.");
258     assert(M->getNamedValue(VI.value().VectorName) &&
259            "Cannot add variant to attribute: "
260            "vector function declaration is missing.");
261   }
262 #endif
263   CI->addFnAttr(
264       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
265 }
266 
267 void llvm::embedBufferInModule(Module &M, MemoryBufferRef Buf,
268                                StringRef SectionName, Align Alignment) {
269   // Embed the memory buffer into the module.
270   Constant *ModuleConstant = ConstantDataArray::get(
271       M.getContext(), makeArrayRef(Buf.getBufferStart(), Buf.getBufferSize()));
272   GlobalVariable *GV = new GlobalVariable(
273       M, ModuleConstant->getType(), true, GlobalValue::PrivateLinkage,
274       ModuleConstant, "llvm.embedded.object");
275   GV->setSection(SectionName);
276   GV->setAlignment(Alignment);
277 
278   LLVMContext &Ctx = M.getContext();
279   NamedMDNode *MD = M.getOrInsertNamedMetadata("llvm.embedded.objects");
280   Metadata *MDVals[] = {ConstantAsMetadata::get(GV),
281                         MDString::get(Ctx, SectionName)};
282 
283   MD->addOperand(llvm::MDNode::get(Ctx, MDVals));
284   GV->setMetadata(LLVMContext::MD_exclude, llvm::MDNode::get(Ctx, {}));
285 
286   appendToCompilerUsed(M, GV);
287 }
288