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