xref: /llvm-project/llvm/lib/Transforms/Utils/ModuleUtils.cpp (revision a78cfd5909024f9daa9a5175e9d411b0fac0e9a1) 
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, ArrayRef(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 collectUsedGlobals(GlobalVariable *GV,
78                                SmallSetVector<Constant *, 16> &Init) {
79   if (!GV || !GV->hasInitializer())
80     return;
81 
82   auto *CA = cast<ConstantArray>(GV->getInitializer());
83   for (Use &Op : CA->operands())
84     Init.insert(cast<Constant>(Op));
85 }
86 
87 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
88   GlobalVariable *GV = M.getGlobalVariable(Name);
89 
90   SmallSetVector<Constant *, 16> Init;
91   collectUsedGlobals(GV, Init);
92   if (GV)
93     GV->eraseFromParent();
94 
95   Type *ArrayEltTy = llvm::Type::getInt8PtrTy(M.getContext());
96   for (auto *V : Values)
97     Init.insert(ConstantExpr::getPointerBitCastOrAddrSpaceCast(V, ArrayEltTy));
98 
99   if (Init.empty())
100     return;
101 
102   ArrayType *ATy = ArrayType::get(ArrayEltTy, Init.size());
103   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
104                                 ConstantArray::get(ATy, Init.getArrayRef()),
105                                 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 removeFromUsedList(Module &M, StringRef Name,
118                                function_ref<bool(Constant *)> ShouldRemove) {
119   GlobalVariable *GV = M.getNamedGlobal(Name);
120   if (!GV)
121     return;
122 
123   SmallSetVector<Constant *, 16> Init;
124   collectUsedGlobals(GV, Init);
125 
126   Type *ArrayEltTy = cast<ArrayType>(GV->getValueType())->getElementType();
127 
128   SmallVector<Constant *, 16> NewInit;
129   for (Constant *MaybeRemoved : Init) {
130     if (!ShouldRemove(MaybeRemoved->stripPointerCasts()))
131       NewInit.push_back(MaybeRemoved);
132   }
133 
134   if (!NewInit.empty()) {
135     ArrayType *ATy = ArrayType::get(ArrayEltTy, NewInit.size());
136     GlobalVariable *NewGV =
137         new GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
138                            ConstantArray::get(ATy, NewInit), "", GV,
139                            GV->getThreadLocalMode(), GV->getAddressSpace());
140     NewGV->setSection(GV->getSection());
141     NewGV->takeName(GV);
142   }
143 
144   GV->eraseFromParent();
145 }
146 
147 void llvm::removeFromUsedLists(Module &M,
148                                function_ref<bool(Constant *)> ShouldRemove) {
149   removeFromUsedList(M, "llvm.used", ShouldRemove);
150   removeFromUsedList(M, "llvm.compiler.used", ShouldRemove);
151 }
152 
153 void llvm::setKCFIType(Module &M, Function &F, StringRef MangledType) {
154   if (!M.getModuleFlag("kcfi"))
155     return;
156   // Matches CodeGenModule::CreateKCFITypeId in Clang.
157   LLVMContext &Ctx = M.getContext();
158   MDBuilder MDB(Ctx);
159   F.setMetadata(
160       LLVMContext::MD_kcfi_type,
161       MDNode::get(Ctx, MDB.createConstant(ConstantInt::get(
162                            Type::getInt32Ty(Ctx),
163                            static_cast<uint32_t>(xxHash64(MangledType))))));
164 }
165 
166 FunctionCallee
167 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
168                                    ArrayRef<Type *> InitArgTypes) {
169   assert(!InitName.empty() && "Expected init function name");
170   return M.getOrInsertFunction(
171       InitName,
172       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
173       AttributeList());
174 }
175 
176 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
177   Function *Ctor = Function::createWithDefaultAttr(
178       FunctionType::get(Type::getVoidTy(M.getContext()), false),
179       GlobalValue::InternalLinkage, M.getDataLayout().getProgramAddressSpace(),
180       CtorName, &M);
181   Ctor->addFnAttr(Attribute::NoUnwind);
182   setKCFIType(M, *Ctor, "_ZTSFvvE"); // void (*)(void)
183   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
184   ReturnInst::Create(M.getContext(), CtorBB);
185   // Ensure Ctor cannot be discarded, even if in a comdat.
186   appendToUsed(M, {Ctor});
187   return Ctor;
188 }
189 
190 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
191     Module &M, StringRef CtorName, StringRef InitName,
192     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
193     StringRef VersionCheckName) {
194   assert(!InitName.empty() && "Expected init function name");
195   assert(InitArgs.size() == InitArgTypes.size() &&
196          "Sanitizer's init function expects different number of arguments");
197   FunctionCallee InitFunction =
198       declareSanitizerInitFunction(M, InitName, InitArgTypes);
199   Function *Ctor = createSanitizerCtor(M, CtorName);
200   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
201   IRB.CreateCall(InitFunction, InitArgs);
202   if (!VersionCheckName.empty()) {
203     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
204         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
205         AttributeList());
206     IRB.CreateCall(VersionCheckFunction, {});
207   }
208   return std::make_pair(Ctor, InitFunction);
209 }
210 
211 std::pair<Function *, FunctionCallee>
212 llvm::getOrCreateSanitizerCtorAndInitFunctions(
213     Module &M, StringRef CtorName, StringRef InitName,
214     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
215     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
216     StringRef VersionCheckName) {
217   assert(!CtorName.empty() && "Expected ctor function name");
218 
219   if (Function *Ctor = M.getFunction(CtorName))
220     // FIXME: Sink this logic into the module, similar to the handling of
221     // globals. This will make moving to a concurrent model much easier.
222     if (Ctor->arg_empty() ||
223         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
224       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
225 
226   Function *Ctor;
227   FunctionCallee InitFunction;
228   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
229       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
230   FunctionsCreatedCallback(Ctor, InitFunction);
231   return std::make_pair(Ctor, InitFunction);
232 }
233 
234 void llvm::filterDeadComdatFunctions(
235     SmallVectorImpl<Function *> &DeadComdatFunctions) {
236   SmallPtrSet<Function *, 32> MaybeDeadFunctions;
237   SmallPtrSet<Comdat *, 32> MaybeDeadComdats;
238   for (Function *F : DeadComdatFunctions) {
239     MaybeDeadFunctions.insert(F);
240     if (Comdat *C = F->getComdat())
241       MaybeDeadComdats.insert(C);
242   }
243 
244   // Find comdats for which all users are dead now.
245   SmallPtrSet<Comdat *, 32> DeadComdats;
246   for (Comdat *C : MaybeDeadComdats) {
247     auto IsUserDead = [&](GlobalObject *GO) {
248       auto *F = dyn_cast<Function>(GO);
249       return F && MaybeDeadFunctions.contains(F);
250     };
251     if (all_of(C->getUsers(), IsUserDead))
252       DeadComdats.insert(C);
253   }
254 
255   // Only keep functions which have no comdat or a dead comdat.
256   erase_if(DeadComdatFunctions, [&](Function *F) {
257     Comdat *C = F->getComdat();
258     return C && !DeadComdats.contains(C);
259   });
260 }
261 
262 std::string llvm::getUniqueModuleId(Module *M) {
263   MD5 Md5;
264   bool ExportsSymbols = false;
265   auto AddGlobal = [&](GlobalValue &GV) {
266     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
267         !GV.hasExternalLinkage() || GV.hasComdat())
268       return;
269     ExportsSymbols = true;
270     Md5.update(GV.getName());
271     Md5.update(ArrayRef<uint8_t>{0});
272   };
273 
274   for (auto &F : *M)
275     AddGlobal(F);
276   for (auto &GV : M->globals())
277     AddGlobal(GV);
278   for (auto &GA : M->aliases())
279     AddGlobal(GA);
280   for (auto &IF : M->ifuncs())
281     AddGlobal(IF);
282 
283   if (!ExportsSymbols)
284     return "";
285 
286   MD5::MD5Result R;
287   Md5.final(R);
288 
289   SmallString<32> Str;
290   MD5::stringifyResult(R, Str);
291   return ("." + Str).str();
292 }
293 
294 void VFABI::setVectorVariantNames(CallInst *CI,
295                                   ArrayRef<std::string> VariantMappings) {
296   if (VariantMappings.empty())
297     return;
298 
299   SmallString<256> Buffer;
300   llvm::raw_svector_ostream Out(Buffer);
301   for (const std::string &VariantMapping : VariantMappings)
302     Out << VariantMapping << ",";
303   // Get rid of the trailing ','.
304   assert(!Buffer.str().empty() && "Must have at least one char.");
305   Buffer.pop_back();
306 
307   Module *M = CI->getModule();
308 #ifndef NDEBUG
309   for (const std::string &VariantMapping : VariantMappings) {
310     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
311     std::optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
312     assert(VI && "Cannot add an invalid VFABI name.");
313     assert(M->getNamedValue(VI->VectorName) &&
314            "Cannot add variant to attribute: "
315            "vector function declaration is missing.");
316   }
317 #endif
318   CI->addFnAttr(
319       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
320 }
321 
322 void llvm::embedBufferInModule(Module &M, MemoryBufferRef Buf,
323                                StringRef SectionName, Align Alignment) {
324   // Embed the memory buffer into the module.
325   Constant *ModuleConstant = ConstantDataArray::get(
326       M.getContext(), ArrayRef(Buf.getBufferStart(), Buf.getBufferSize()));
327   GlobalVariable *GV = new GlobalVariable(
328       M, ModuleConstant->getType(), true, GlobalValue::PrivateLinkage,
329       ModuleConstant, "llvm.embedded.object");
330   GV->setSection(SectionName);
331   GV->setAlignment(Alignment);
332 
333   LLVMContext &Ctx = M.getContext();
334   NamedMDNode *MD = M.getOrInsertNamedMetadata("llvm.embedded.objects");
335   Metadata *MDVals[] = {ConstantAsMetadata::get(GV),
336                         MDString::get(Ctx, SectionName)};
337 
338   MD->addOperand(llvm::MDNode::get(Ctx, MDVals));
339   GV->setMetadata(LLVMContext::MD_exclude, llvm::MDNode::get(Ctx, {}));
340 
341   appendToCompilerUsed(M, GV);
342 }
343