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/STLExtras.h" 18 #include "llvm/ADT/StringRef.h" 19 #include "llvm/IR/DiagnosticInfo.h" 20 #include "llvm/IR/Function.h" 21 #include "llvm/IR/GlobalAlias.h" 22 #include "llvm/IR/GlobalValue.h" 23 #include "llvm/IR/GlobalVariable.h" 24 #include "llvm/IR/InlineAsm.h" 25 #include "llvm/IR/Module.h" 26 #include "llvm/MC/MCAsmInfo.h" 27 #include "llvm/MC/MCContext.h" 28 #include "llvm/MC/MCInstrInfo.h" 29 #include "llvm/MC/MCObjectFileInfo.h" 30 #include "llvm/MC/MCParser/MCAsmParser.h" 31 #include "llvm/MC/MCParser/MCTargetAsmParser.h" 32 #include "llvm/MC/MCRegisterInfo.h" 33 #include "llvm/MC/MCSubtargetInfo.h" 34 #include "llvm/MC/MCSymbol.h" 35 #include "llvm/MC/MCTargetOptions.h" 36 #include "llvm/MC/TargetRegistry.h" 37 #include "llvm/Object/SymbolicFile.h" 38 #include "llvm/Support/Casting.h" 39 #include "llvm/Support/ErrorHandling.h" 40 #include "llvm/Support/MemoryBuffer.h" 41 #include "llvm/Support/SMLoc.h" 42 #include "llvm/Support/SourceMgr.h" 43 #include "llvm/Support/raw_ostream.h" 44 #include "llvm/TargetParser/Triple.h" 45 #include <algorithm> 46 #include <cassert> 47 #include <cstdint> 48 #include <memory> 49 #include <string> 50 51 using namespace llvm; 52 using namespace object; 53 54 void ModuleSymbolTable::addModule(Module *M) { 55 if (FirstMod) 56 assert(FirstMod->getTargetTriple() == M->getTargetTriple()); 57 else 58 FirstMod = M; 59 60 for (GlobalValue &GV : M->global_values()) 61 SymTab.push_back(&GV); 62 63 CollectAsmSymbols(*M, [this](StringRef Name, BasicSymbolRef::Flags Flags) { 64 SymTab.push_back(new (AsmSymbols.Allocate()) 65 AsmSymbol(std::string(Name), Flags)); 66 }); 67 } 68 69 static void 70 initializeRecordStreamer(const Module &M, 71 function_ref<void(RecordStreamer &)> Init) { 72 // This function may be called twice, once for ModuleSummaryIndexAnalysis and 73 // the other when writing the IR symbol table. If parsing inline assembly has 74 // caused errors in the first run, suppress the second run. 75 if (M.getContext().getDiagHandlerPtr()->HasErrors) 76 return; 77 StringRef InlineAsm = M.getModuleInlineAsm(); 78 if (InlineAsm.empty()) 79 return; 80 81 std::string Err; 82 const Triple TT(M.getTargetTriple()); 83 const Target *T = TargetRegistry::lookupTarget(TT.str(), Err); 84 assert(T && T->hasMCAsmParser()); 85 86 std::unique_ptr<MCRegisterInfo> MRI(T->createMCRegInfo(TT.str())); 87 if (!MRI) 88 return; 89 90 MCTargetOptions MCOptions; 91 std::unique_ptr<MCAsmInfo> MAI(T->createMCAsmInfo(*MRI, TT.str(), MCOptions)); 92 if (!MAI) 93 return; 94 95 std::unique_ptr<MCSubtargetInfo> STI( 96 T->createMCSubtargetInfo(TT.str(), "", "")); 97 if (!STI) 98 return; 99 100 std::unique_ptr<MCInstrInfo> MCII(T->createMCInstrInfo()); 101 if (!MCII) 102 return; 103 104 std::unique_ptr<MemoryBuffer> Buffer( 105 MemoryBuffer::getMemBuffer(InlineAsm, "<inline asm>")); 106 SourceMgr SrcMgr; 107 SrcMgr.AddNewSourceBuffer(std::move(Buffer), SMLoc()); 108 109 MCContext MCCtx(TT, MAI.get(), MRI.get(), STI.get(), &SrcMgr); 110 std::unique_ptr<MCObjectFileInfo> MOFI( 111 T->createMCObjectFileInfo(MCCtx, /*PIC=*/false)); 112 MCCtx.setObjectFileInfo(MOFI.get()); 113 RecordStreamer Streamer(MCCtx, M); 114 T->createNullTargetStreamer(Streamer); 115 116 std::unique_ptr<MCAsmParser> Parser( 117 createMCAsmParser(SrcMgr, MCCtx, Streamer, *MAI)); 118 119 std::unique_ptr<MCTargetAsmParser> TAP( 120 T->createMCAsmParser(*STI, *Parser, *MCII, MCOptions)); 121 if (!TAP) 122 return; 123 124 MCCtx.setDiagnosticHandler([&](const SMDiagnostic &SMD, bool IsInlineAsm, 125 const SourceMgr &SrcMgr, 126 std::vector<const MDNode *> &LocInfos) { 127 M.getContext().diagnose( 128 DiagnosticInfoSrcMgr(SMD, M.getName(), IsInlineAsm, /*LocCookie=*/0)); 129 }); 130 131 // Module-level inline asm is assumed to use At&t syntax (see 132 // AsmPrinter::doInitialization()). 133 Parser->setAssemblerDialect(InlineAsm::AD_ATT); 134 135 Parser->setTargetParser(*TAP); 136 if (Parser->Run(false)) 137 return; 138 139 Init(Streamer); 140 } 141 142 void ModuleSymbolTable::CollectAsmSymbols( 143 const Module &M, 144 function_ref<void(StringRef, BasicSymbolRef::Flags)> AsmSymbol) { 145 initializeRecordStreamer(M, [&](RecordStreamer &Streamer) { 146 Streamer.flushSymverDirectives(); 147 148 for (auto &KV : Streamer) { 149 StringRef Key = KV.first(); 150 RecordStreamer::State Value = KV.second; 151 // FIXME: For now we just assume that all asm symbols are executable. 152 uint32_t Res = BasicSymbolRef::SF_Executable; 153 switch (Value) { 154 case RecordStreamer::NeverSeen: 155 llvm_unreachable("NeverSeen should have been replaced earlier"); 156 case RecordStreamer::DefinedGlobal: 157 Res |= BasicSymbolRef::SF_Global; 158 break; 159 case RecordStreamer::Defined: 160 break; 161 case RecordStreamer::Global: 162 case RecordStreamer::Used: 163 Res |= BasicSymbolRef::SF_Undefined; 164 Res |= BasicSymbolRef::SF_Global; 165 break; 166 case RecordStreamer::DefinedWeak: 167 Res |= BasicSymbolRef::SF_Weak; 168 Res |= BasicSymbolRef::SF_Global; 169 break; 170 case RecordStreamer::UndefinedWeak: 171 Res |= BasicSymbolRef::SF_Weak; 172 Res |= BasicSymbolRef::SF_Undefined; 173 } 174 AsmSymbol(Key, BasicSymbolRef::Flags(Res)); 175 } 176 }); 177 178 // In ELF, object code generated for x86-32 and some code models of x86-64 may 179 // reference the special symbol _GLOBAL_OFFSET_TABLE_ that is not used in the 180 // IR. Record it like inline asm symbols. 181 Triple TT(M.getTargetTriple()); 182 if (!TT.isOSBinFormatELF() || !TT.isX86()) 183 return; 184 auto CM = M.getCodeModel(); 185 if (TT.getArch() == Triple::x86 || CM == CodeModel::Medium || 186 CM == CodeModel::Large) { 187 AsmSymbol("_GLOBAL_OFFSET_TABLE_", 188 BasicSymbolRef::Flags(BasicSymbolRef::SF_Undefined | 189 BasicSymbolRef::SF_Global)); 190 } 191 } 192 193 void ModuleSymbolTable::CollectAsmSymvers( 194 const Module &M, function_ref<void(StringRef, StringRef)> AsmSymver) { 195 initializeRecordStreamer(M, [&](RecordStreamer &Streamer) { 196 for (auto &KV : Streamer.symverAliases()) 197 for (auto &Alias : KV.second) 198 AsmSymver(KV.first->getName(), Alias); 199 }); 200 } 201 202 void ModuleSymbolTable::printSymbolName(raw_ostream &OS, Symbol S) const { 203 if (isa<AsmSymbol *>(S)) { 204 OS << cast<AsmSymbol *>(S)->first; 205 return; 206 } 207 208 auto *GV = cast<GlobalValue *>(S); 209 if (GV->hasDLLImportStorageClass()) 210 OS << "__imp_"; 211 212 Mang.getNameWithPrefix(OS, GV, false); 213 } 214 215 uint32_t ModuleSymbolTable::getSymbolFlags(Symbol S) const { 216 if (isa<AsmSymbol *>(S)) 217 return cast<AsmSymbol *>(S)->second; 218 219 auto *GV = cast<GlobalValue *>(S); 220 221 uint32_t Res = BasicSymbolRef::SF_None; 222 if (GV->isDeclarationForLinker()) 223 Res |= BasicSymbolRef::SF_Undefined; 224 else if (GV->hasHiddenVisibility() && !GV->hasLocalLinkage()) 225 Res |= BasicSymbolRef::SF_Hidden; 226 if (const GlobalVariable *GVar = dyn_cast<GlobalVariable>(GV)) { 227 if (GVar->isConstant()) 228 Res |= BasicSymbolRef::SF_Const; 229 } 230 if (const GlobalObject *GO = GV->getAliaseeObject()) 231 if (isa<Function>(GO) || isa<GlobalIFunc>(GO)) 232 Res |= BasicSymbolRef::SF_Executable; 233 if (isa<GlobalAlias>(GV)) 234 Res |= BasicSymbolRef::SF_Indirect; 235 if (GV->hasPrivateLinkage()) 236 Res |= BasicSymbolRef::SF_FormatSpecific; 237 if (!GV->hasLocalLinkage()) 238 Res |= BasicSymbolRef::SF_Global; 239 if (GV->hasCommonLinkage()) 240 Res |= BasicSymbolRef::SF_Common; 241 if (GV->hasLinkOnceLinkage() || GV->hasWeakLinkage() || 242 GV->hasExternalWeakLinkage()) 243 Res |= BasicSymbolRef::SF_Weak; 244 245 if (GV->getName().starts_with("llvm.")) 246 Res |= BasicSymbolRef::SF_FormatSpecific; 247 else if (auto *Var = dyn_cast<GlobalVariable>(GV)) { 248 if (Var->getSection() == "llvm.metadata") 249 Res |= BasicSymbolRef::SF_FormatSpecific; 250 } 251 252 return Res; 253 } 254