1 //===- ModuleSymbolTable.cpp - symbol table for in-memory IR --------------===// 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 class represents a symbol table built from in-memory IR. It provides 10 // access to GlobalValues and should only be used if such access is required 11 // (e.g. in the LTO implementation). 12 // 13 //===----------------------------------------------------------------------===// 14 15 #include "llvm/Object/ModuleSymbolTable.h" 16 #include "RecordStreamer.h" 17 #include "llvm/ADT/StringRef.h" 18 #include "llvm/IR/DiagnosticInfo.h" 19 #include "llvm/IR/Function.h" 20 #include "llvm/IR/GlobalAlias.h" 21 #include "llvm/IR/GlobalValue.h" 22 #include "llvm/IR/GlobalVariable.h" 23 #include "llvm/IR/InlineAsm.h" 24 #include "llvm/IR/Module.h" 25 #include "llvm/MC/MCAsmInfo.h" 26 #include "llvm/MC/MCContext.h" 27 #include "llvm/MC/MCInstrInfo.h" 28 #include "llvm/MC/MCObjectFileInfo.h" 29 #include "llvm/MC/MCParser/MCAsmParser.h" 30 #include "llvm/MC/MCParser/MCTargetAsmParser.h" 31 #include "llvm/MC/MCRegisterInfo.h" 32 #include "llvm/MC/MCSubtargetInfo.h" 33 #include "llvm/MC/MCSymbol.h" 34 #include "llvm/MC/MCTargetOptions.h" 35 #include "llvm/MC/TargetRegistry.h" 36 #include "llvm/Object/SymbolicFile.h" 37 #include "llvm/Support/Casting.h" 38 #include "llvm/Support/ErrorHandling.h" 39 #include "llvm/Support/MemoryBuffer.h" 40 #include "llvm/Support/SMLoc.h" 41 #include "llvm/Support/SourceMgr.h" 42 #include "llvm/Support/raw_ostream.h" 43 #include "llvm/TargetParser/Triple.h" 44 #include <cassert> 45 #include <cstdint> 46 #include <memory> 47 #include <string> 48 49 using namespace llvm; 50 using namespace object; 51 52 void ModuleSymbolTable::addModule(Module *M) { 53 if (FirstMod) 54 assert(FirstMod->getTargetTriple() == M->getTargetTriple()); 55 else 56 FirstMod = M; 57 58 for (GlobalValue &GV : M->global_values()) 59 SymTab.push_back(&GV); 60 61 CollectAsmSymbols(*M, [this](StringRef Name, BasicSymbolRef::Flags Flags) { 62 SymTab.push_back(new (AsmSymbols.Allocate()) 63 AsmSymbol(std::string(Name), Flags)); 64 }); 65 } 66 67 static void 68 initializeRecordStreamer(const Module &M, 69 function_ref<void(RecordStreamer &)> Init) { 70 // This function may be called twice, once for ModuleSummaryIndexAnalysis and 71 // the other when writing the IR symbol table. If parsing inline assembly has 72 // caused errors in the first run, suppress the second run. 73 if (M.getContext().getDiagHandlerPtr()->HasErrors) 74 return; 75 StringRef InlineAsm = M.getModuleInlineAsm(); 76 if (InlineAsm.empty()) 77 return; 78 79 std::string Err; 80 const Triple TT(M.getTargetTriple()); 81 const Target *T = TargetRegistry::lookupTarget(TT.str(), Err); 82 assert(T && T->hasMCAsmParser()); 83 84 std::unique_ptr<MCRegisterInfo> MRI(T->createMCRegInfo(TT.str())); 85 if (!MRI) 86 return; 87 88 MCTargetOptions MCOptions; 89 std::unique_ptr<MCAsmInfo> MAI(T->createMCAsmInfo(*MRI, TT.str(), MCOptions)); 90 if (!MAI) 91 return; 92 93 std::unique_ptr<MCSubtargetInfo> STI( 94 T->createMCSubtargetInfo(TT.str(), "", "")); 95 if (!STI) 96 return; 97 98 std::unique_ptr<MCInstrInfo> MCII(T->createMCInstrInfo()); 99 if (!MCII) 100 return; 101 102 std::unique_ptr<MemoryBuffer> Buffer( 103 MemoryBuffer::getMemBuffer(InlineAsm, "<inline asm>")); 104 SourceMgr SrcMgr; 105 SrcMgr.AddNewSourceBuffer(std::move(Buffer), SMLoc()); 106 107 MCContext MCCtx(TT, MAI.get(), MRI.get(), STI.get(), &SrcMgr); 108 std::unique_ptr<MCObjectFileInfo> MOFI( 109 T->createMCObjectFileInfo(MCCtx, /*PIC=*/false)); 110 MCCtx.setObjectFileInfo(MOFI.get()); 111 RecordStreamer Streamer(MCCtx, M); 112 T->createNullTargetStreamer(Streamer); 113 114 std::unique_ptr<MCAsmParser> Parser( 115 createMCAsmParser(SrcMgr, MCCtx, Streamer, *MAI)); 116 117 std::unique_ptr<MCTargetAsmParser> TAP( 118 T->createMCAsmParser(*STI, *Parser, *MCII, MCOptions)); 119 if (!TAP) 120 return; 121 122 MCCtx.setDiagnosticHandler([&](const SMDiagnostic &SMD, bool IsInlineAsm, 123 const SourceMgr &SrcMgr, 124 std::vector<const MDNode *> &LocInfos) { 125 M.getContext().diagnose( 126 DiagnosticInfoSrcMgr(SMD, M.getName(), IsInlineAsm, /*LocCookie=*/0)); 127 }); 128 129 // Module-level inline asm is assumed to use At&t syntax (see 130 // AsmPrinter::doInitialization()). 131 Parser->setAssemblerDialect(InlineAsm::AD_ATT); 132 133 Parser->setTargetParser(*TAP); 134 if (Parser->Run(false)) 135 return; 136 137 Init(Streamer); 138 } 139 140 void ModuleSymbolTable::CollectAsmSymbols( 141 const Module &M, 142 function_ref<void(StringRef, BasicSymbolRef::Flags)> AsmSymbol) { 143 initializeRecordStreamer(M, [&](RecordStreamer &Streamer) { 144 Streamer.flushSymverDirectives(); 145 146 for (auto &KV : Streamer) { 147 StringRef Key = KV.first(); 148 RecordStreamer::State Value = KV.second; 149 // FIXME: For now we just assume that all asm symbols are executable. 150 uint32_t Res = BasicSymbolRef::SF_Executable; 151 switch (Value) { 152 case RecordStreamer::NeverSeen: 153 llvm_unreachable("NeverSeen should have been replaced earlier"); 154 case RecordStreamer::DefinedGlobal: 155 Res |= BasicSymbolRef::SF_Global; 156 break; 157 case RecordStreamer::Defined: 158 break; 159 case RecordStreamer::Global: 160 case RecordStreamer::Used: 161 Res |= BasicSymbolRef::SF_Undefined; 162 Res |= BasicSymbolRef::SF_Global; 163 break; 164 case RecordStreamer::DefinedWeak: 165 Res |= BasicSymbolRef::SF_Weak; 166 Res |= BasicSymbolRef::SF_Global; 167 break; 168 case RecordStreamer::UndefinedWeak: 169 Res |= BasicSymbolRef::SF_Weak; 170 Res |= BasicSymbolRef::SF_Undefined; 171 } 172 AsmSymbol(Key, BasicSymbolRef::Flags(Res)); 173 } 174 }); 175 176 // In ELF, object code generated for x86-32 and some code models of x86-64 may 177 // reference the special symbol _GLOBAL_OFFSET_TABLE_ that is not used in the 178 // IR. Record it like inline asm symbols. 179 Triple TT(M.getTargetTriple()); 180 if (!TT.isOSBinFormatELF() || !TT.isX86()) 181 return; 182 auto CM = M.getCodeModel(); 183 if (TT.getArch() == Triple::x86 || CM == CodeModel::Medium || 184 CM == CodeModel::Large) { 185 AsmSymbol("_GLOBAL_OFFSET_TABLE_", 186 BasicSymbolRef::Flags(BasicSymbolRef::SF_Undefined | 187 BasicSymbolRef::SF_Global)); 188 } 189 } 190 191 void ModuleSymbolTable::CollectAsmSymvers( 192 const Module &M, function_ref<void(StringRef, StringRef)> AsmSymver) { 193 initializeRecordStreamer(M, [&](RecordStreamer &Streamer) { 194 for (auto &KV : Streamer.symverAliases()) 195 for (auto &Alias : KV.second) 196 AsmSymver(KV.first->getName(), Alias); 197 }); 198 } 199 200 void ModuleSymbolTable::printSymbolName(raw_ostream &OS, Symbol S) const { 201 if (isa<AsmSymbol *>(S)) { 202 OS << cast<AsmSymbol *>(S)->first; 203 return; 204 } 205 206 auto *GV = cast<GlobalValue *>(S); 207 if (GV->hasDLLImportStorageClass()) 208 OS << "__imp_"; 209 210 Mang.getNameWithPrefix(OS, GV, false); 211 } 212 213 uint32_t ModuleSymbolTable::getSymbolFlags(Symbol S) const { 214 if (isa<AsmSymbol *>(S)) 215 return cast<AsmSymbol *>(S)->second; 216 217 auto *GV = cast<GlobalValue *>(S); 218 219 uint32_t Res = BasicSymbolRef::SF_None; 220 if (GV->isDeclarationForLinker()) 221 Res |= BasicSymbolRef::SF_Undefined; 222 else if (GV->hasHiddenVisibility() && !GV->hasLocalLinkage()) 223 Res |= BasicSymbolRef::SF_Hidden; 224 if (const GlobalVariable *GVar = dyn_cast<GlobalVariable>(GV)) { 225 if (GVar->isConstant()) 226 Res |= BasicSymbolRef::SF_Const; 227 } 228 if (const GlobalObject *GO = GV->getAliaseeObject()) 229 if (isa<Function>(GO) || isa<GlobalIFunc>(GO)) 230 Res |= BasicSymbolRef::SF_Executable; 231 if (isa<GlobalAlias>(GV)) 232 Res |= BasicSymbolRef::SF_Indirect; 233 if (GV->hasPrivateLinkage()) 234 Res |= BasicSymbolRef::SF_FormatSpecific; 235 if (!GV->hasLocalLinkage()) 236 Res |= BasicSymbolRef::SF_Global; 237 if (GV->hasCommonLinkage()) 238 Res |= BasicSymbolRef::SF_Common; 239 if (GV->hasLinkOnceLinkage() || GV->hasWeakLinkage() || 240 GV->hasExternalWeakLinkage()) 241 Res |= BasicSymbolRef::SF_Weak; 242 243 if (GV->getName().starts_with("llvm.")) 244 Res |= BasicSymbolRef::SF_FormatSpecific; 245 else if (auto *Var = dyn_cast<GlobalVariable>(GV)) { 246 if (Var->getSection() == "llvm.metadata") 247 Res |= BasicSymbolRef::SF_FormatSpecific; 248 } 249 250 return Res; 251 } 252