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/TargetLibraryInfo.h"
15 #include "llvm/Analysis/VectorUtils.h"
16 #include "llvm/IR/DerivedTypes.h"
17 #include "llvm/IR/Function.h"
18 #include "llvm/IR/IRBuilder.h"
19 #include "llvm/IR/Module.h"
20 #include "llvm/Support/raw_ostream.h"
21 using namespace llvm;
22 
23 #define DEBUG_TYPE "moduleutils"
24 
25 static void appendToGlobalArray(const char *Array, Module &M, Function *F,
26                                 int Priority, Constant *Data) {
27   IRBuilder<> IRB(M.getContext());
28   FunctionType *FnTy = FunctionType::get(IRB.getVoidTy(), false);
29 
30   // Get the current set of static global constructors and add the new ctor
31   // to the list.
32   SmallVector<Constant *, 16> CurrentCtors;
33   StructType *EltTy = StructType::get(
34       IRB.getInt32Ty(), PointerType::getUnqual(FnTy), IRB.getInt8PtrTy());
35   if (GlobalVariable *GVCtor = M.getNamedGlobal(Array)) {
36     if (Constant *Init = GVCtor->getInitializer()) {
37       unsigned n = Init->getNumOperands();
38       CurrentCtors.reserve(n + 1);
39       for (unsigned i = 0; i != n; ++i)
40         CurrentCtors.push_back(cast<Constant>(Init->getOperand(i)));
41     }
42     GVCtor->eraseFromParent();
43   }
44 
45   // Build a 3 field global_ctor entry.  We don't take a comdat key.
46   Constant *CSVals[3];
47   CSVals[0] = IRB.getInt32(Priority);
48   CSVals[1] = F;
49   CSVals[2] = Data ? ConstantExpr::getPointerCast(Data, IRB.getInt8PtrTy())
50                    : Constant::getNullValue(IRB.getInt8PtrTy());
51   Constant *RuntimeCtorInit =
52       ConstantStruct::get(EltTy, makeArrayRef(CSVals, EltTy->getNumElements()));
53 
54   CurrentCtors.push_back(RuntimeCtorInit);
55 
56   // Create a new initializer.
57   ArrayType *AT = ArrayType::get(EltTy, CurrentCtors.size());
58   Constant *NewInit = ConstantArray::get(AT, CurrentCtors);
59 
60   // Create the new global variable and replace all uses of
61   // the old global variable with the new one.
62   (void)new GlobalVariable(M, NewInit->getType(), false,
63                            GlobalValue::AppendingLinkage, NewInit, Array);
64 }
65 
66 void llvm::appendToGlobalCtors(Module &M, Function *F, int Priority, Constant *Data) {
67   appendToGlobalArray("llvm.global_ctors", M, F, Priority, Data);
68 }
69 
70 void llvm::appendToGlobalDtors(Module &M, Function *F, int Priority, Constant *Data) {
71   appendToGlobalArray("llvm.global_dtors", M, F, Priority, Data);
72 }
73 
74 static void appendToUsedList(Module &M, StringRef Name, ArrayRef<GlobalValue *> Values) {
75   GlobalVariable *GV = M.getGlobalVariable(Name);
76   SmallPtrSet<Constant *, 16> InitAsSet;
77   SmallVector<Constant *, 16> Init;
78   if (GV) {
79     if (GV->hasInitializer()) {
80       auto *CA = cast<ConstantArray>(GV->getInitializer());
81       for (auto &Op : CA->operands()) {
82         Constant *C = cast_or_null<Constant>(Op);
83         if (InitAsSet.insert(C).second)
84           Init.push_back(C);
85       }
86     }
87     GV->eraseFromParent();
88   }
89 
90   Type *Int8PtrTy = llvm::Type::getInt8PtrTy(M.getContext());
91   for (auto *V : Values) {
92     Constant *C = ConstantExpr::getBitCast(V, Int8PtrTy);
93     if (InitAsSet.insert(C).second)
94       Init.push_back(C);
95   }
96 
97   if (Init.empty())
98     return;
99 
100   ArrayType *ATy = ArrayType::get(Int8PtrTy, Init.size());
101   GV = new llvm::GlobalVariable(M, ATy, false, GlobalValue::AppendingLinkage,
102                                 ConstantArray::get(ATy, Init), Name);
103   GV->setSection("llvm.metadata");
104 }
105 
106 void llvm::appendToUsed(Module &M, ArrayRef<GlobalValue *> Values) {
107   appendToUsedList(M, "llvm.used", Values);
108 }
109 
110 void llvm::appendToCompilerUsed(Module &M, ArrayRef<GlobalValue *> Values) {
111   appendToUsedList(M, "llvm.compiler.used", Values);
112 }
113 
114 FunctionCallee
115 llvm::declareSanitizerInitFunction(Module &M, StringRef InitName,
116                                    ArrayRef<Type *> InitArgTypes) {
117   assert(!InitName.empty() && "Expected init function name");
118   return M.getOrInsertFunction(
119       InitName,
120       FunctionType::get(Type::getVoidTy(M.getContext()), InitArgTypes, false),
121       AttributeList());
122 }
123 
124 Function *llvm::createSanitizerCtor(Module &M, StringRef CtorName) {
125   Function *Ctor = Function::createWithDefaultAttr(
126       FunctionType::get(Type::getVoidTy(M.getContext()), false),
127       GlobalValue::InternalLinkage, 0, CtorName, &M);
128   Ctor->addAttribute(AttributeList::FunctionIndex, Attribute::NoUnwind);
129   BasicBlock *CtorBB = BasicBlock::Create(M.getContext(), "", Ctor);
130   ReturnInst::Create(M.getContext(), CtorBB);
131   return Ctor;
132 }
133 
134 std::pair<Function *, FunctionCallee> llvm::createSanitizerCtorAndInitFunctions(
135     Module &M, StringRef CtorName, StringRef InitName,
136     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
137     StringRef VersionCheckName) {
138   assert(!InitName.empty() && "Expected init function name");
139   assert(InitArgs.size() == InitArgTypes.size() &&
140          "Sanitizer's init function expects different number of arguments");
141   FunctionCallee InitFunction =
142       declareSanitizerInitFunction(M, InitName, InitArgTypes);
143   Function *Ctor = createSanitizerCtor(M, CtorName);
144   IRBuilder<> IRB(Ctor->getEntryBlock().getTerminator());
145   IRB.CreateCall(InitFunction, InitArgs);
146   if (!VersionCheckName.empty()) {
147     FunctionCallee VersionCheckFunction = M.getOrInsertFunction(
148         VersionCheckName, FunctionType::get(IRB.getVoidTy(), {}, false),
149         AttributeList());
150     IRB.CreateCall(VersionCheckFunction, {});
151   }
152   return std::make_pair(Ctor, InitFunction);
153 }
154 
155 std::pair<Function *, FunctionCallee>
156 llvm::getOrCreateSanitizerCtorAndInitFunctions(
157     Module &M, StringRef CtorName, StringRef InitName,
158     ArrayRef<Type *> InitArgTypes, ArrayRef<Value *> InitArgs,
159     function_ref<void(Function *, FunctionCallee)> FunctionsCreatedCallback,
160     StringRef VersionCheckName) {
161   assert(!CtorName.empty() && "Expected ctor function name");
162 
163   if (Function *Ctor = M.getFunction(CtorName))
164     // FIXME: Sink this logic into the module, similar to the handling of
165     // globals. This will make moving to a concurrent model much easier.
166     if (Ctor->arg_size() == 0 ||
167         Ctor->getReturnType() == Type::getVoidTy(M.getContext()))
168       return {Ctor, declareSanitizerInitFunction(M, InitName, InitArgTypes)};
169 
170   Function *Ctor;
171   FunctionCallee InitFunction;
172   std::tie(Ctor, InitFunction) = llvm::createSanitizerCtorAndInitFunctions(
173       M, CtorName, InitName, InitArgTypes, InitArgs, VersionCheckName);
174   FunctionsCreatedCallback(Ctor, InitFunction);
175   return std::make_pair(Ctor, InitFunction);
176 }
177 
178 Function *llvm::getOrCreateInitFunction(Module &M, StringRef Name) {
179   assert(!Name.empty() && "Expected init function name");
180   if (Function *F = M.getFunction(Name)) {
181     if (F->arg_size() != 0 ||
182         F->getReturnType() != Type::getVoidTy(M.getContext())) {
183       std::string Err;
184       raw_string_ostream Stream(Err);
185       Stream << "Sanitizer interface function defined with wrong type: " << *F;
186       report_fatal_error(Err);
187     }
188     return F;
189   }
190   Function *F =
191       cast<Function>(M.getOrInsertFunction(Name, AttributeList(),
192                                            Type::getVoidTy(M.getContext()))
193                          .getCallee());
194 
195   appendToGlobalCtors(M, F, 0);
196 
197   return F;
198 }
199 
200 void llvm::filterDeadComdatFunctions(
201     Module &M, SmallVectorImpl<Function *> &DeadComdatFunctions) {
202   // Build a map from the comdat to the number of entries in that comdat we
203   // think are dead. If this fully covers the comdat group, then the entire
204   // group is dead. If we find another entry in the comdat group though, we'll
205   // have to preserve the whole group.
206   SmallDenseMap<Comdat *, int, 16> ComdatEntriesCovered;
207   for (Function *F : DeadComdatFunctions) {
208     Comdat *C = F->getComdat();
209     assert(C && "Expected all input GVs to be in a comdat!");
210     ComdatEntriesCovered[C] += 1;
211   }
212 
213   auto CheckComdat = [&](Comdat &C) {
214     auto CI = ComdatEntriesCovered.find(&C);
215     if (CI == ComdatEntriesCovered.end())
216       return;
217 
218     // If this could have been covered by a dead entry, just subtract one to
219     // account for it.
220     if (CI->second > 0) {
221       CI->second -= 1;
222       return;
223     }
224 
225     // If we've already accounted for all the entries that were dead, the
226     // entire comdat is alive so remove it from the map.
227     ComdatEntriesCovered.erase(CI);
228   };
229 
230   auto CheckAllComdats = [&] {
231     for (Function &F : M.functions())
232       if (Comdat *C = F.getComdat()) {
233         CheckComdat(*C);
234         if (ComdatEntriesCovered.empty())
235           return;
236       }
237     for (GlobalVariable &GV : M.globals())
238       if (Comdat *C = GV.getComdat()) {
239         CheckComdat(*C);
240         if (ComdatEntriesCovered.empty())
241           return;
242       }
243     for (GlobalAlias &GA : M.aliases())
244       if (Comdat *C = GA.getComdat()) {
245         CheckComdat(*C);
246         if (ComdatEntriesCovered.empty())
247           return;
248       }
249   };
250   CheckAllComdats();
251 
252   if (ComdatEntriesCovered.empty()) {
253     DeadComdatFunctions.clear();
254     return;
255   }
256 
257   // Remove the entries that were not covering.
258   erase_if(DeadComdatFunctions, [&](GlobalValue *GV) {
259     return ComdatEntriesCovered.find(GV->getComdat()) ==
260            ComdatEntriesCovered.end();
261   });
262 }
263 
264 std::string llvm::getUniqueModuleId(Module *M) {
265   MD5 Md5;
266   bool ExportsSymbols = false;
267   auto AddGlobal = [&](GlobalValue &GV) {
268     if (GV.isDeclaration() || GV.getName().startswith("llvm.") ||
269         !GV.hasExternalLinkage() || GV.hasComdat())
270       return;
271     ExportsSymbols = true;
272     Md5.update(GV.getName());
273     Md5.update(ArrayRef<uint8_t>{0});
274   };
275 
276   for (auto &F : *M)
277     AddGlobal(F);
278   for (auto &GV : M->globals())
279     AddGlobal(GV);
280   for (auto &GA : M->aliases())
281     AddGlobal(GA);
282   for (auto &IF : M->ifuncs())
283     AddGlobal(IF);
284 
285   if (!ExportsSymbols)
286     return "";
287 
288   MD5::MD5Result R;
289   Md5.final(R);
290 
291   SmallString<32> Str;
292   MD5::stringifyResult(R, Str);
293   return ("." + Str).str();
294 }
295 
296 void VFABI::setVectorVariantNames(
297     CallInst *CI, const SmallVector<std::string, 8> &VariantMappings) {
298   if (VariantMappings.empty())
299     return;
300 
301   SmallString<256> Buffer;
302   llvm::raw_svector_ostream Out(Buffer);
303   for (const std::string &VariantMapping : VariantMappings)
304     Out << VariantMapping << ",";
305   // Get rid of the trailing ','.
306   assert(!Buffer.str().empty() && "Must have at least one char.");
307   Buffer.pop_back();
308 
309   Module *M = CI->getModule();
310 #ifndef NDEBUG
311   for (const std::string &VariantMapping : VariantMappings) {
312     LLVM_DEBUG(dbgs() << "VFABI: adding mapping '" << VariantMapping << "'\n");
313     Optional<VFInfo> VI = VFABI::tryDemangleForVFABI(VariantMapping, *M);
314     assert(VI.hasValue() && "Cannot add an invalid VFABI name.");
315     assert(M->getNamedValue(VI.getValue().VectorName) &&
316            "Cannot add variant to attribute: "
317            "vector function declaration is missing.");
318   }
319 #endif
320   CI->addAttribute(
321       AttributeList::FunctionIndex,
322       Attribute::get(M->getContext(), MappingsAttrName, Buffer.str()));
323 }
324