xref: /llvm-project/llvm/lib/Transforms/Utils/ModuleUtils.cpp (revision d4b6fcb32e29d0cd834a3c89205fef48fbfc1d2d) 
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/MDBuilder.h"
19 #include "llvm/IR/Module.h"
20 #include "llvm/Support/raw_ostream.h"
21 #include "llvm/Support/xxhash.h"
22 using namespace llvm;
23 
24 #define DEBUG_TYPE "moduleutils"
25 
26 static void appendToGlobalArray(StringRef ArrayName, Module &M, Function *F,
27                                 int Priority, Constant *Data) {
28   IRBuilder<> IRB(M.getContext());
29   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
30 
31   // Get the current set of static global constructors and add the new ctor
32   // to the list.
33   SmallVector<Constant *, 16> CurrentCtors;
34   StructType *EltTy = StructType::get(
35       IRB.getInt32Ty(), PointerType::get(FnTy, F->getAddressSpace()),
36       IRB.getInt8PtrTy());
37 
38   if (GlobalVariable *GVCtor = M.getNamedGlobal(ArrayName)) {
39     if (Constant *Init = GVCtor->getInitializer()) {
40       unsigned n = Init->getNumOperands();
41       CurrentCtors.reserve(n + 1);
42       for (unsigned i = 0; i != n; ++i)
43         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
44     }
45     GVCtor->eraseFromParent();
46   }
47 
48   // Build a 3 field global_ctor entry.  We don't take a comdat key.
49   Constant *CSVals[3];
50   CSVals[0] = IRB.getInt32(Priority);
51   CSVals[1] = F;
52   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
53                    : Constant::getNullValue(IRB.getInt8PtrTy());
54   Constant *RuntimeCtorInit =
55       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
56 
57   CurrentCtors.push_back(RuntimeCtorInit);
58 
59   // Create a new initializer.
60   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
61   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
62 
63   // Create the new global variable and replace all uses of
64   // the old global variable with the new one.
65   (void)new GlobalVariable(M, NewInit->getType(), false,
66                            GlobalValue::AppendingLinkage, NewInit, ArrayName);
67 }
68 
69 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
70   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
71 }
72 
73 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
74   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
75 }
76 
77 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
78   GlobalVariable *GV = M.getGlobalVariable(Name);
79   SmallPtrSet<Constant *, 16> InitAsSet;
80   SmallVector<Constant *, 16> Init;
81   if (GV) {
82     if (GV->hasInitializer()) {
83       auto *CA = cast<ConstantArray>(GV->getInitializer());
84       for (auto &Op : CA->operands()) {
85         Constant *C = cast_or_null<Constant>(Op);
86         if (InitAsSet.insert(C).second)
87           Init.push_back(C);
88       }
89     }
90     GV->eraseFromParent();
91   }
92 
93   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
94   for (auto *V : Values) {
95     Constant *C = ConstantExpr::getPointerBitCastOrAddrSpaceCast(V, Int8PtrTy);
96     if (InitAsSet.insert(C).second)
97       Init.push_back(C);
98   }
99 
100   if (Init.empty())
101     return;
102 
103   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
104   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
105                                 ConstantArray::get(ATy, Init), Name);
106   GV->setSection("llvm.metadata");
107 }
108 
109 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
110   appendToUsedList(M, "llvm.used", Values);
111 }
112 
113 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
114   appendToUsedList(M, "llvm.compiler.used", Values);
115 }
116 
117 static void setKCFIType(Module &M, Function &F, StringRef MangledType) {
118   if (!M.getModuleFlag("kcfi"))
119     return;
120   // Matches CodeGenModule::CreateKCFITypeId in Clang.
121   LLVMContext &Ctx = M.getContext();
122   MDBuilder MDB(Ctx);
123   F.setMetadata(
124       LLVMContext::MD_kcfi_type,
125       MDNode::get(Ctx, MDB.createConstant(ConstantInt::get(
126                            Type::getInt32Ty(Ctx),
127                            static_cast<uint32_t>(xxHash64(MangledType))))));
128 }
129 
130 FunctionCallee
131 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
132                                    ArrayRef<Type *> InitArgTypes) {
133   assert(!InitName.empty() && "Expected init function name");
134   return M.getOrInsertFunction(
135       InitName,
136       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
137       AttributeList());
138 }
139 
140 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
141   Function *Ctor = Function::createWithDefaultAttr(
142       FunctionType::get(Type::getVoidTy(M.getContext()), false),
143       GlobalValue::InternalLinkage, M.getDataLayout().getProgramAddressSpace(),
144       CtorName, &M);
145   Ctor->addFnAttr(Attribute::NoUnwind);
146   setKCFIType(M, *Ctor, "_ZTSFvvE"); // void (*)(void)
147   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
148   ReturnInst::Create(M.getContext(), CtorBB);
149   // Ensure Ctor cannot be discarded, even if in a comdat.
150   appendToUsed(M, {Ctor});
151   return Ctor;
152 }
153 
154 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
155     Module &M, StringRef CtorName, StringRef InitName,
156     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
157     StringRef VersionCheckName) {
158   assert(!InitName.empty() && "Expected init function name");
159   assert(InitArgs.size() == InitArgTypes.size() &&
160          "Sanitizer's init function expects different number of arguments");
161   FunctionCallee InitFunction =
162       declareSanitizerInitFunction(M, InitName, InitArgTypes);
163   Function *Ctor = createSanitizerCtor(M, CtorName);
164   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
165   IRB.CreateCall(InitFunction, InitArgs);
166   if (!VersionCheckName.empty()) {
167     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
168         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
169         AttributeList());
170     IRB.CreateCall(VersionCheckFunction, {});
171   }
172   return std::make_pair(Ctor, InitFunction);
173 }
174 
175 std::pair<Function *, FunctionCallee>
176 llvm::getOrCreateSanitizerCtorAndInitFunctions(
177     Module &M, StringRef CtorName, StringRef InitName,
178     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
179     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
180     StringRef VersionCheckName) {
181   assert(!CtorName.empty() && "Expected ctor function name");
182 
183   if (Function *Ctor = M.getFunction(CtorName))
184     // FIXME: Sink this logic into the module, similar to the handling of
185     // globals. This will make moving to a concurrent model much easier.
186     if (Ctor->arg_empty() ||
187         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
188       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
189 
190   Function *Ctor;
191   FunctionCallee InitFunction;
192   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
193       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
194   FunctionsCreatedCallback(Ctor, InitFunction);
195   return std::make_pair(Ctor, InitFunction);
196 }
197 
198 void llvm::filterDeadComdatFunctions(
199     SmallVectorImpl<Function *> &DeadComdatFunctions) {
200   SmallPtrSet<Function *, 32> MaybeDeadFunctions;
201   SmallPtrSet<Comdat *, 32> MaybeDeadComdats;
202   for (Function *F : DeadComdatFunctions) {
203     MaybeDeadFunctions.insert(F);
204     if (Comdat *C = F->getComdat())
205       MaybeDeadComdats.insert(C);
206   }
207 
208   // Find comdats for which all users are dead now.
209   SmallPtrSet<Comdat *, 32> DeadComdats;
210   for (Comdat *C : MaybeDeadComdats) {
211     auto IsUserDead = [&](GlobalObject *GO) {
212       auto *F = dyn_cast<Function>(GO);
213       return F && MaybeDeadFunctions.contains(F);
214     };
215     if (all_of(C->getUsers(), IsUserDead))
216       DeadComdats.insert(C);
217   }
218 
219   // Only keep functions which have no comdat or a dead comdat.
220   erase_if(DeadComdatFunctions, [&](Function *F) {
221     Comdat *C = F->getComdat();
222     return C && !DeadComdats.contains(C);
223   });
224 }
225 
226 std::string llvm::getUniqueModuleId(Module *M) {
227   MD5 Md5;
228   bool ExportsSymbols = false;
229   auto AddGlobal = [&](GlobalValue &GV) {
230     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
231         !GV.hasExternalLinkage() || GV.hasComdat())
232       return;
233     ExportsSymbols = true;
234     Md5.update(GV.getName());
235     Md5.update(ArrayRef<uint8_t>{0});
236   };
237 
238   for (auto &F : *M)
239     AddGlobal(F);
240   for (auto &GV : M->globals())
241     AddGlobal(GV);
242   for (auto &GA : M->aliases())
243     AddGlobal(GA);
244   for (auto &IF : M->ifuncs())
245     AddGlobal(IF);
246 
247   if (!ExportsSymbols)
248     return "";
249 
250   MD5::MD5Result R;
251   Md5.final(R);
252 
253   SmallString<32> Str;
254   MD5::stringifyResult(R, Str);
255   return ("." + Str).str();
256 }
257 
258 void VFABI::setVectorVariantNames(CallInst *CI,
259                                   ArrayRef<std::string> VariantMappings) {
260   if (VariantMappings.empty())
261     return;
262 
263   SmallString<256> Buffer;
264   llvm::raw_svector_ostream Out(Buffer);
265   for (const std::string &VariantMapping : VariantMappings)
266     Out << VariantMapping << ",";
267   // Get rid of the trailing ','.
268   assert(!Buffer.str().empty() && "Must have at least one char.");
269   Buffer.pop_back();
270 
271   Module *M = CI->getModule();
272 #ifndef NDEBUG
273   for (const std::string &VariantMapping : VariantMappings) {
274     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
275     std::optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
276     assert(VI && "Cannot add an invalid VFABI name.");
277     assert(M->getNamedValue(VI.value().VectorName) &&
278            "Cannot add variant to attribute: "
279            "vector function declaration is missing.");
280   }
281 #endif
282   CI->addFnAttr(
283       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
284 }
285 
286 void llvm::embedBufferInModule(Module &M, MemoryBufferRef Buf,
287                                StringRef SectionName, Align Alignment) {
288   // Embed the memory buffer into the module.
289   Constant *ModuleConstant = ConstantDataArray::get(
290       M.getContext(), makeArrayRef(Buf.getBufferStart(), Buf.getBufferSize()));
291   GlobalVariable *GV = new GlobalVariable(
292       M, ModuleConstant->getType(), true, GlobalValue::PrivateLinkage,
293       ModuleConstant, "llvm.embedded.object");
294   GV->setSection(SectionName);
295   GV->setAlignment(Alignment);
296 
297   LLVMContext &Ctx = M.getContext();
298   NamedMDNode *MD = M.getOrInsertNamedMetadata("llvm.embedded.objects");
299   Metadata *MDVals[] = {ConstantAsMetadata::get(GV),
300                         MDString::get(Ctx, SectionName)};
301 
302   MD->addOperand(llvm::MDNode::get(Ctx, MDVals));
303   GV->setMetadata(LLVMContext::MD_exclude, llvm::MDNode::get(Ctx, {}));
304 
305   appendToCompilerUsed(M, GV);
306 }
307