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
addModule(Module * M)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
initializeRecordStreamer(const Module & M,function_ref<void (RecordStreamer &)> Init)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 MOFI->setSDKVersion(M.getSDKVersion());
113 MCCtx.setObjectFileInfo(MOFI.get());
114 RecordStreamer Streamer(MCCtx, M);
115 T->createNullTargetStreamer(Streamer);
116
117 std::unique_ptr<MCAsmParser> Parser(
118 createMCAsmParser(SrcMgr, MCCtx, Streamer, *MAI));
119
120 std::unique_ptr<MCTargetAsmParser> TAP(
121 T->createMCAsmParser(*STI, *Parser, *MCII, MCOptions));
122 if (!TAP)
123 return;
124
125 MCCtx.setDiagnosticHandler([&](const SMDiagnostic &SMD, bool IsInlineAsm,
126 const SourceMgr &SrcMgr,
127 std::vector<const MDNode *> &LocInfos) {
128 M.getContext().diagnose(
129 DiagnosticInfoSrcMgr(SMD, M.getName(), IsInlineAsm, /*LocCookie=*/0));
130 });
131
132 // Module-level inline asm is assumed to use At&t syntax (see
133 // AsmPrinter::doInitialization()).
134 Parser->setAssemblerDialect(InlineAsm::AD_ATT);
135
136 Parser->setTargetParser(*TAP);
137 if (Parser->Run(false))
138 return;
139
140 Init(Streamer);
141 }
142
CollectAsmSymbols(const Module & M,function_ref<void (StringRef,BasicSymbolRef::Flags)> AsmSymbol)143 void ModuleSymbolTable::CollectAsmSymbols(
144 const Module &M,
145 function_ref<void(StringRef, BasicSymbolRef::Flags)> AsmSymbol) {
146 initializeRecordStreamer(M, [&](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
CollectAsmSymvers(const Module & M,function_ref<void (StringRef,StringRef)> AsmSymver)194 void ModuleSymbolTable::CollectAsmSymvers(
195 const Module &M, function_ref<void(StringRef, StringRef)> AsmSymver) {
196 initializeRecordStreamer(M, [&](RecordStreamer &Streamer) {
197 for (auto &KV : Streamer.symverAliases())
198 for (auto &Alias : KV.second)
199 AsmSymver(KV.first->getName(), Alias);
200 });
201 }
202
printSymbolName(raw_ostream & OS,Symbol S) const203 void ModuleSymbolTable::printSymbolName(raw_ostream &OS, Symbol S) const {
204 if (isa<AsmSymbol *>(S)) {
205 OS << cast<AsmSymbol *>(S)->first;
206 return;
207 }
208
209 auto *GV = cast<GlobalValue *>(S);
210 if (GV->hasDLLImportStorageClass())
211 OS << "__imp_";
212
213 Mang.getNameWithPrefix(OS, GV, false);
214 }
215
getSymbolFlags(Symbol S) const216 uint32_t ModuleSymbolTable::getSymbolFlags(Symbol S) const {
217 if (isa<AsmSymbol *>(S))
218 return cast<AsmSymbol *>(S)->second;
219
220 auto *GV = cast<GlobalValue *>(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>(GV)) {
228 if (GVar->isConstant())
229 Res |= BasicSymbolRef::SF_Const;
230 }
231 if (const GlobalObject *GO = GV->getAliaseeObject())
232 if (isa<Function>(GO) || isa<GlobalIFunc>(GO))
233 Res |= BasicSymbolRef::SF_Executable;
234 if (isa<GlobalAlias>(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("llvm."))
247 Res |= BasicSymbolRef::SF_FormatSpecific;
248 else if (auto *Var = dyn_cast<GlobalVariable>(GV)) {
249 if (Var->getSection() == "llvm.metadata")
250 Res |= BasicSymbolRef::SF_FormatSpecific;
251 }
252
253 return Res;
254 }
255