xref: /llvm-project/llvm/lib/Transforms/Utils/ModuleUtils.cpp (revision 02ea3694a0a541f6328d120f02eb04977c1fc840) 
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(StringRef ArrayName, 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::get(FnTy, F->getAddressSpace()),
34       IRB.getInt8PtrTy());
35 
36   if (GlobalVariable *GVCtor = M.getNamedGlobal(ArrayName)) {
37     if (Constant *Init = GVCtor->getInitializer()) {
38       unsigned n = Init->getNumOperands();
39       CurrentCtors.reserve(n + 1);
40       for (unsigned i = 0; i != n; ++i)
41         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
42     }
43     GVCtor->eraseFromParent();
44   }
45 
46   // Build a 3 field global_ctor entry.  We don't take a comdat key.
47   Constant *CSVals[3];
48   CSVals[0] = IRB.getInt32(Priority);
49   CSVals[1] = F;
50   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
51                    : Constant::getNullValue(IRB.getInt8PtrTy());
52   Constant *RuntimeCtorInit =
53       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
54 
55   CurrentCtors.push_back(RuntimeCtorInit);
56 
57   // Create a new initializer.
58   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
59   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
60 
61   // Create the new global variable and replace all uses of
62   // the old global variable with the new one.
63   (void)new GlobalVariable(M, NewInit->getType(), false,
64                            GlobalValue::AppendingLinkage, NewInit, ArrayName);
65 }
66 
67 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
68   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
69 }
70 
71 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
72   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
73 }
74 
75 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
76   GlobalVariable *GV = M.getGlobalVariable(Name);
77   SmallPtrSet<Constant *, 16> InitAsSet;
78   SmallVector<Constant *, 16> Init;
79   if (GV) {
80     if (GV->hasInitializer()) {
81       auto *CA = cast<ConstantArray>(GV->getInitializer());
82       for (auto &Op : CA->operands()) {
83         Constant *C = cast_or_null<Constant>(Op);
84         if (InitAsSet.insert(C).second)
85           Init.push_back(C);
86       }
87     }
88     GV->eraseFromParent();
89   }
90 
91   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
92   for (auto *V : Values) {
93     Constant *C = ConstantExpr::getPointerBitCastOrAddrSpaceCast(V, Int8PtrTy);
94     if (InitAsSet.insert(C).second)
95       Init.push_back(C);
96   }
97 
98   if (Init.empty())
99     return;
100 
101   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
102   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
103                                 ConstantArray::get(ATy, Init), Name);
104   GV->setSection("llvm.metadata");
105 }
106 
107 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
108   appendToUsedList(M, "llvm.used", Values);
109 }
110 
111 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
112   appendToUsedList(M, "llvm.compiler.used", Values);
113 }
114 
115 FunctionCallee
116 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
117                                    ArrayRef<Type *> InitArgTypes) {
118   assert(!InitName.empty() && "Expected init function name");
119   return M.getOrInsertFunction(
120       InitName,
121       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
122       AttributeList());
123 }
124 
125 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
126   Function *Ctor = Function::createWithDefaultAttr(
127       FunctionType::get(Type::getVoidTy(M.getContext()), false),
128       GlobalValue::InternalLinkage, M.getDataLayout().getProgramAddressSpace(),
129       CtorName, &M);
130   Ctor->addFnAttr(Attribute::NoUnwind);
131   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
132   ReturnInst::Create(M.getContext(), CtorBB);
133   // Ensure Ctor cannot be discarded, even if in a comdat.
134   appendToUsed(M, {Ctor});
135   return Ctor;
136 }
137 
138 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
139     Module &M, StringRef CtorName, StringRef InitName,
140     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
141     StringRef VersionCheckName) {
142   assert(!InitName.empty() && "Expected init function name");
143   assert(InitArgs.size() == InitArgTypes.size() &&
144          "Sanitizer's init function expects different number of arguments");
145   FunctionCallee InitFunction =
146       declareSanitizerInitFunction(M, InitName, InitArgTypes);
147   Function *Ctor = createSanitizerCtor(M, CtorName);
148   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
149   IRB.CreateCall(InitFunction, InitArgs);
150   if (!VersionCheckName.empty()) {
151     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
152         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
153         AttributeList());
154     IRB.CreateCall(VersionCheckFunction, {});
155   }
156   return std::make_pair(Ctor, InitFunction);
157 }
158 
159 std::pair<Function *, FunctionCallee>
160 llvm::getOrCreateSanitizerCtorAndInitFunctions(
161     Module &M, StringRef CtorName, StringRef InitName,
162     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
163     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
164     StringRef VersionCheckName) {
165   assert(!CtorName.empty() && "Expected ctor function name");
166 
167   if (Function *Ctor = M.getFunction(CtorName))
168     // FIXME: Sink this logic into the module, similar to the handling of
169     // globals. This will make moving to a concurrent model much easier.
170     if (Ctor->arg_empty() ||
171         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
172       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
173 
174   Function *Ctor;
175   FunctionCallee InitFunction;
176   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
177       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
178   FunctionsCreatedCallback(Ctor, InitFunction);
179   return std::make_pair(Ctor, InitFunction);
180 }
181 
182 void llvm::filterDeadComdatFunctions(
183     SmallVectorImpl<Function *> &DeadComdatFunctions) {
184   SmallPtrSet<Function *, 32> MaybeDeadFunctions;
185   SmallPtrSet<Comdat *, 32> MaybeDeadComdats;
186   for (Function *F : DeadComdatFunctions) {
187     MaybeDeadFunctions.insert(F);
188     if (Comdat *C = F->getComdat())
189       MaybeDeadComdats.insert(C);
190   }
191 
192   // Find comdats for which all users are dead now.
193   SmallPtrSet<Comdat *, 32> DeadComdats;
194   for (Comdat *C : MaybeDeadComdats) {
195     auto IsUserDead = [&](GlobalObject *GO) {
196       auto *F = dyn_cast<Function>(GO);
197       return F && MaybeDeadFunctions.contains(F);
198     };
199     if (all_of(C->getUsers(), IsUserDead))
200       DeadComdats.insert(C);
201   }
202 
203   // Only keep functions which have no comdat or a dead comdat.
204   erase_if(DeadComdatFunctions, [&](Function *F) {
205     Comdat *C = F->getComdat();
206     return C && !DeadComdats.contains(C);
207   });
208 }
209 
210 std::string llvm::getUniqueModuleId(Module *M) {
211   MD5 Md5;
212   bool ExportsSymbols = false;
213   auto AddGlobal = [&](GlobalValue &GV) {
214     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
215         !GV.hasExternalLinkage() || GV.hasComdat())
216       return;
217     ExportsSymbols = true;
218     Md5.update(GV.getName());
219     Md5.update(ArrayRef<uint8_t>{0});
220   };
221 
222   for (auto &F : *M)
223     AddGlobal(F);
224   for (auto &GV : M->globals())
225     AddGlobal(GV);
226   for (auto &GA : M->aliases())
227     AddGlobal(GA);
228   for (auto &IF : M->ifuncs())
229     AddGlobal(IF);
230 
231   if (!ExportsSymbols)
232     return "";
233 
234   MD5::MD5Result R;
235   Md5.final(R);
236 
237   SmallString<32> Str;
238   MD5::stringifyResult(R, Str);
239   return ("." + Str).str();
240 }
241 
242 void VFABI::setVectorVariantNames(CallInst *CI,
243                                   ArrayRef<std::string> VariantMappings) {
244   if (VariantMappings.empty())
245     return;
246 
247   SmallString<256> Buffer;
248   llvm::raw_svector_ostream Out(Buffer);
249   for (const std::string &VariantMapping : VariantMappings)
250     Out << VariantMapping << ",";
251   // Get rid of the trailing ','.
252   assert(!Buffer.str().empty() && "Must have at least one char.");
253   Buffer.pop_back();
254 
255   Module *M = CI->getModule();
256 #ifndef NDEBUG
257   for (const std::string &VariantMapping : VariantMappings) {
258     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
259     Optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
260     assert(VI && "Cannot add an invalid VFABI name.");
261     assert(M->getNamedValue(VI.value().VectorName) &&
262            "Cannot add variant to attribute: "
263            "vector function declaration is missing.");
264   }
265 #endif
266   CI->addFnAttr(
267       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
268 }
269 
270 void llvm::embedBufferInModule(Module &M, MemoryBufferRef Buf,
271                                StringRef SectionName, Align Alignment) {
272   // Embed the memory buffer into the module.
273   Constant *ModuleConstant = ConstantDataArray::get(
274       M.getContext(), makeArrayRef(Buf.getBufferStart(), Buf.getBufferSize()));
275   GlobalVariable *GV = new GlobalVariable(
276       M, ModuleConstant->getType(), true, GlobalValue::PrivateLinkage,
277       ModuleConstant, "llvm.embedded.object");
278   GV->setSection(SectionName);
279   GV->setAlignment(Alignment);
280 
281   LLVMContext &Ctx = M.getContext();
282   NamedMDNode *MD = M.getOrInsertNamedMetadata("llvm.embedded.objects");
283   Metadata *MDVals[] = {ConstantAsMetadata::get(GV),
284                         MDString::get(Ctx, SectionName)};
285 
286   MD->addOperand(llvm::MDNode::get(Ctx, MDVals));
287   GV->setMetadata(LLVMContext::MD_exclude, llvm::MDNode::get(Ctx, {}));
288 
289   appendToCompilerUsed(M, GV);
290 }
291