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
49using namespace llvm;
50using namespace object;
51
52void 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(x: &GV);
60
61 CollectAsmSymbols(M: *M, AsmSymbol: [this](StringRef Name, BasicSymbolRef::Flags Flags) {
62 SymTab.push_back(x: new (AsmSymbols.Allocate())
63 AsmSymbol(std::string(Name), Flags));
64 });
65}
66
67static void
68initializeRecordStreamer(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 if (!M.hasModuleInlineAsm())
76 return;
77
78 std::string Err;
79 const Triple TT(M.getTargetTriple());
80 const Target *T = TargetRegistry::lookupTarget(TheTriple: TT, Error&: Err);
81 assert(T && T->hasMCAsmParser());
82
83 std::unique_ptr<MCRegisterInfo> MRI(T->createMCRegInfo(TT));
84 if (!MRI)
85 return;
86
87 MCTargetOptions MCOptions;
88 std::unique_ptr<MCAsmInfo> MAI(T->createMCAsmInfo(MRI: *MRI, TheTriple: TT, Options: MCOptions));
89 if (!MAI)
90 return;
91
92 std::unique_ptr<MCInstrInfo> MCII(T->createMCInstrInfo());
93 if (!MCII)
94 return;
95
96 for (const Module::GlobalAsmFragment &Frag : M.getModuleInlineAsm()) {
97 std::unique_ptr<MCSubtargetInfo> STI(T->createMCSubtargetInfo(
98 TheTriple: TT, CPU: Frag.Props.TargetCPU, Features: Frag.Props.TargetFeatures));
99 if (!STI)
100 return;
101
102 std::unique_ptr<MemoryBuffer> Buffer(
103 MemoryBuffer::getMemBuffer(InputData: Frag.Asm, BufferName: "<inline asm>"));
104 SourceMgr SrcMgr;
105 SrcMgr.AddNewSourceBuffer(F: std::move(Buffer), IncludeLoc: SMLoc());
106
107 MCContext MCCtx(TT, *MAI, *MRI, *STI, &SrcMgr);
108 std::unique_ptr<MCObjectFileInfo> MOFI(
109 T->createMCObjectFileInfo(Ctx&: MCCtx, /*PIC=*/false));
110 MCCtx.setObjectFileInfo(MOFI.get());
111 RecordStreamer Streamer(MCCtx, M);
112 T->createNullTargetStreamer(S&: 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: *STI, Parser&: *Parser, MII: *MCII));
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 DI: 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->setSymbolScanningMode(true);
134
135 Parser->setTargetParser(*TAP);
136 if (Parser->Run(NoInitialTextSection: false))
137 return;
138
139 Init(Streamer);
140 }
141}
142
143void ModuleSymbolTable::CollectAsmSymbols(
144 const Module &M,
145 function_ref<void(StringRef, BasicSymbolRef::Flags)> AsmSymbol) {
146 initializeRecordStreamer(M, Init: [&](RecordStreamer &Streamer) {
147 Streamer.flushSymverDirectives();
148
149 for (auto &KV : Streamer) {
150 StringRef Key = KV.first();
151 RecordStreamer::State Value = KV.second;
152 // FIXME: For now we just assume that all asm symbols are executable.
153 uint32_t Res = BasicSymbolRef::SF_Executable;
154 switch (Value) {
155 case RecordStreamer::NeverSeen:
156 llvm_unreachable("NeverSeen should have been replaced earlier");
157 case RecordStreamer::DefinedGlobal:
158 Res |= BasicSymbolRef::SF_Global;
159 break;
160 case RecordStreamer::Defined:
161 break;
162 case RecordStreamer::Global:
163 case RecordStreamer::Used:
164 Res |= BasicSymbolRef::SF_Undefined;
165 Res |= BasicSymbolRef::SF_Global;
166 break;
167 case RecordStreamer::DefinedWeak:
168 Res |= BasicSymbolRef::SF_Weak;
169 Res |= BasicSymbolRef::SF_Global;
170 break;
171 case RecordStreamer::UndefinedWeak:
172 Res |= BasicSymbolRef::SF_Weak;
173 Res |= BasicSymbolRef::SF_Undefined;
174 }
175 AsmSymbol(Key, BasicSymbolRef::Flags(Res));
176 }
177 });
178
179 // In ELF, object code generated for x86-32 and some code models of x86-64 may
180 // reference the special symbol _GLOBAL_OFFSET_TABLE_ that is not used in the
181 // IR. Record it like inline asm symbols.
182 Triple TT(M.getTargetTriple());
183 if (!TT.isOSBinFormatELF() || !TT.isX86())
184 return;
185 auto CM = M.getCodeModel();
186 if (TT.getArch() == Triple::x86 || CM == CodeModel::Medium ||
187 CM == CodeModel::Large) {
188 AsmSymbol("_GLOBAL_OFFSET_TABLE_",
189 BasicSymbolRef::Flags(BasicSymbolRef::SF_Undefined |
190 BasicSymbolRef::SF_Global));
191 }
192}
193
194void ModuleSymbolTable::CollectAsmSymvers(
195 const Module &M, function_ref<void(StringRef, StringRef)> AsmSymver) {
196 initializeRecordStreamer(M, Init: [&](RecordStreamer &Streamer) {
197 for (auto &KV : Streamer.symverAliases())
198 for (auto &Alias : KV.second)
199 AsmSymver(KV.first->getName(), Alias);
200 });
201}
202
203void ModuleSymbolTable::printSymbolName(raw_ostream &OS, Symbol S) const {
204 if (isa<AsmSymbol *>(Val: S)) {
205 OS << cast<AsmSymbol *>(Val&: S)->first;
206 return;
207 }
208
209 auto *GV = cast<GlobalValue *>(Val&: S);
210 if (GV->hasDLLImportStorageClass())
211 OS << "__imp_";
212
213 Mang.getNameWithPrefix(OS, GV, CannotUsePrivateLabel: false);
214}
215
216uint32_t ModuleSymbolTable::getSymbolFlags(Symbol S) const {
217 if (isa<AsmSymbol *>(Val: S))
218 return cast<AsmSymbol *>(Val&: S)->second;
219
220 auto *GV = cast<GlobalValue *>(Val&: S);
221
222 uint32_t Res = BasicSymbolRef::SF_None;
223 if (GV->isDeclarationForLinker())
224 Res |= BasicSymbolRef::SF_Undefined;
225 else if (GV->hasHiddenVisibility() && !GV->hasLocalLinkage())
226 Res |= BasicSymbolRef::SF_Hidden;
227 if (const GlobalVariable *GVar = dyn_cast<GlobalVariable>(Val: GV)) {
228 if (GVar->isConstant())
229 Res |= BasicSymbolRef::SF_Const;
230 }
231 if (const GlobalObject *GO = GV->getAliaseeObject())
232 if (isa<Function>(Val: GO) || isa<GlobalIFunc>(Val: GO))
233 Res |= BasicSymbolRef::SF_Executable;
234 if (isa<GlobalAlias>(Val: GV))
235 Res |= BasicSymbolRef::SF_Indirect;
236 if (GV->hasPrivateLinkage())
237 Res |= BasicSymbolRef::SF_FormatSpecific;
238 if (!GV->hasLocalLinkage())
239 Res |= BasicSymbolRef::SF_Global;
240 if (GV->hasCommonLinkage())
241 Res |= BasicSymbolRef::SF_Common;
242 if (GV->hasLinkOnceLinkage() || GV->hasWeakLinkage() ||
243 GV->hasExternalWeakLinkage())
244 Res |= BasicSymbolRef::SF_Weak;
245
246 if (GV->getName().starts_with(Prefix: "llvm."))
247 Res |= BasicSymbolRef::SF_FormatSpecific;
248 else if (auto *Var = dyn_cast<GlobalVariable>(Val: GV)) {
249 if (Var->getSection() == "llvm.metadata")
250 Res |= BasicSymbolRef::SF_FormatSpecific;
251 }
252
253 return Res;
254}
255