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 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(TheTriple: TT, Error&: Err);
82 assert(T && T->hasMCAsmParser());
83
84 std::unique_ptr<MCRegisterInfo> MRI(T->createMCRegInfo(TT));
85 if (!MRI)
86 return;
87
88 MCTargetOptions MCOptions;
89 std::unique_ptr<MCAsmInfo> MAI(T->createMCAsmInfo(MRI: *MRI, TheTriple: TT, Options: MCOptions));
90 if (!MAI)
91 return;
92
93 std::unique_ptr<MCSubtargetInfo> STI(T->createMCSubtargetInfo(TheTriple: TT, CPU: "", Features: ""));
94 if (!STI)
95 return;
96
97 std::unique_ptr<MCInstrInfo> MCII(T->createMCInstrInfo());
98 if (!MCII)
99 return;
100
101 std::unique_ptr<MemoryBuffer> Buffer(
102 MemoryBuffer::getMemBuffer(InputData: InlineAsm, BufferName: "<inline asm>"));
103 SourceMgr SrcMgr;
104 SrcMgr.AddNewSourceBuffer(F: std::move(Buffer), IncludeLoc: SMLoc());
105
106 MCContext MCCtx(TT, MAI.get(), MRI.get(), STI.get(), &SrcMgr);
107 std::unique_ptr<MCObjectFileInfo> MOFI(
108 T->createMCObjectFileInfo(Ctx&: MCCtx, /*PIC=*/false));
109 MCCtx.setObjectFileInfo(MOFI.get());
110 RecordStreamer Streamer(MCCtx, M);
111 T->createNullTargetStreamer(S&: Streamer);
112
113 std::unique_ptr<MCAsmParser> Parser(
114 createMCAsmParser(SrcMgr, MCCtx, Streamer, *MAI));
115
116 std::unique_ptr<MCTargetAsmParser> TAP(
117 T->createMCAsmParser(STI: *STI, Parser&: *Parser, MII: *MCII, Options: MCOptions));
118 if (!TAP)
119 return;
120
121 MCCtx.setDiagnosticHandler([&](const SMDiagnostic &SMD, bool IsInlineAsm,
122 const SourceMgr &SrcMgr,
123 std::vector<const MDNode *> &LocInfos) {
124 M.getContext().diagnose(
125 DI: DiagnosticInfoSrcMgr(SMD, M.getName(), IsInlineAsm, /*LocCookie=*/0));
126 });
127
128 // Module-level inline asm is assumed to use At&t syntax (see
129 // AsmPrinter::doInitialization()).
130 Parser->setAssemblerDialect(InlineAsm::AD_ATT);
131
132 Parser->setSymbolScanningMode(true);
133
134 Parser->setTargetParser(*TAP);
135 if (Parser->Run(NoInitialTextSection: false))
136 return;
137
138 Init(Streamer);
139}
140
141void ModuleSymbolTable::CollectAsmSymbols(
142 const Module &M,
143 function_ref<void(StringRef, BasicSymbolRef::Flags)> AsmSymbol) {
144 initializeRecordStreamer(M, Init: [&](RecordStreamer &Streamer) {
145 Streamer.flushSymverDirectives();
146
147 for (auto &KV : Streamer) {
148 StringRef Key = KV.first();
149 RecordStreamer::State Value = KV.second;
150 // FIXME: For now we just assume that all asm symbols are executable.
151 uint32_t Res = BasicSymbolRef::SF_Executable;
152 switch (Value) {
153 case RecordStreamer::NeverSeen:
154 llvm_unreachable("NeverSeen should have been replaced earlier");
155 case RecordStreamer::DefinedGlobal:
156 Res |= BasicSymbolRef::SF_Global;
157 break;
158 case RecordStreamer::Defined:
159 break;
160 case RecordStreamer::Global:
161 case RecordStreamer::Used:
162 Res |= BasicSymbolRef::SF_Undefined;
163 Res |= BasicSymbolRef::SF_Global;
164 break;
165 case RecordStreamer::DefinedWeak:
166 Res |= BasicSymbolRef::SF_Weak;
167 Res |= BasicSymbolRef::SF_Global;
168 break;
169 case RecordStreamer::UndefinedWeak:
170 Res |= BasicSymbolRef::SF_Weak;
171 Res |= BasicSymbolRef::SF_Undefined;
172 }
173 AsmSymbol(Key, BasicSymbolRef::Flags(Res));
174 }
175 });
176
177 // In ELF, object code generated for x86-32 and some code models of x86-64 may
178 // reference the special symbol _GLOBAL_OFFSET_TABLE_ that is not used in the
179 // IR. Record it like inline asm symbols.
180 Triple TT(M.getTargetTriple());
181 if (!TT.isOSBinFormatELF() || !TT.isX86())
182 return;
183 auto CM = M.getCodeModel();
184 if (TT.getArch() == Triple::x86 || CM == CodeModel::Medium ||
185 CM == CodeModel::Large) {
186 AsmSymbol("_GLOBAL_OFFSET_TABLE_",
187 BasicSymbolRef::Flags(BasicSymbolRef::SF_Undefined |
188 BasicSymbolRef::SF_Global));
189 }
190}
191
192void ModuleSymbolTable::CollectAsmSymvers(
193 const Module &M, function_ref<void(StringRef, StringRef)> AsmSymver) {
194 initializeRecordStreamer(M, Init: [&](RecordStreamer &Streamer) {
195 for (auto &KV : Streamer.symverAliases())
196 for (auto &Alias : KV.second)
197 AsmSymver(KV.first->getName(), Alias);
198 });
199}
200
201void ModuleSymbolTable::printSymbolName(raw_ostream &OS, Symbol S) const {
202 if (isa<AsmSymbol *>(Val: S)) {
203 OS << cast<AsmSymbol *>(Val&: S)->first;
204 return;
205 }
206
207 auto *GV = cast<GlobalValue *>(Val&: S);
208 if (GV->hasDLLImportStorageClass())
209 OS << "__imp_";
210
211 Mang.getNameWithPrefix(OS, GV, CannotUsePrivateLabel: false);
212}
213
214uint32_t ModuleSymbolTable::getSymbolFlags(Symbol S) const {
215 if (isa<AsmSymbol *>(Val: S))
216 return cast<AsmSymbol *>(Val&: S)->second;
217
218 auto *GV = cast<GlobalValue *>(Val&: S);
219
220 uint32_t Res = BasicSymbolRef::SF_None;
221 if (GV->isDeclarationForLinker())
222 Res |= BasicSymbolRef::SF_Undefined;
223 else if (GV->hasHiddenVisibility() && !GV->hasLocalLinkage())
224 Res |= BasicSymbolRef::SF_Hidden;
225 if (const GlobalVariable *GVar = dyn_cast<GlobalVariable>(Val: GV)) {
226 if (GVar->isConstant())
227 Res |= BasicSymbolRef::SF_Const;
228 }
229 if (const GlobalObject *GO = GV->getAliaseeObject())
230 if (isa<Function>(Val: GO) || isa<GlobalIFunc>(Val: GO))
231 Res |= BasicSymbolRef::SF_Executable;
232 if (isa<GlobalAlias>(Val: GV))
233 Res |= BasicSymbolRef::SF_Indirect;
234 if (GV->hasPrivateLinkage())
235 Res |= BasicSymbolRef::SF_FormatSpecific;
236 if (!GV->hasLocalLinkage())
237 Res |= BasicSymbolRef::SF_Global;
238 if (GV->hasCommonLinkage())
239 Res |= BasicSymbolRef::SF_Common;
240 if (GV->hasLinkOnceLinkage() || GV->hasWeakLinkage() ||
241 GV->hasExternalWeakLinkage())
242 Res |= BasicSymbolRef::SF_Weak;
243
244 if (GV->getName().starts_with(Prefix: "llvm."))
245 Res |= BasicSymbolRef::SF_FormatSpecific;
246 else if (auto *Var = dyn_cast<GlobalVariable>(Val: GV)) {
247 if (Var->getSection() == "llvm.metadata")
248 Res |= BasicSymbolRef::SF_FormatSpecific;
249 }
250
251 return Res;
252}
253