1//===- IRSymtab.cpp - implementation of IR symbol tables ------------------===//
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#include "llvm/Object/IRSymtab.h"
10#include "llvm/ADT/ArrayRef.h"
11#include "llvm/ADT/DenseMap.h"
12#include "llvm/ADT/SmallPtrSet.h"
13#include "llvm/ADT/SmallString.h"
14#include "llvm/ADT/SmallVector.h"
15#include "llvm/ADT/StringRef.h"
16#include "llvm/Bitcode/BitcodeReader.h"
17#include "llvm/Config/llvm-config.h"
18#include "llvm/IR/Comdat.h"
19#include "llvm/IR/DataLayout.h"
20#include "llvm/IR/GlobalAlias.h"
21#include "llvm/IR/GlobalObject.h"
22#include "llvm/IR/Mangler.h"
23#include "llvm/IR/Metadata.h"
24#include "llvm/IR/Module.h"
25#include "llvm/IR/RuntimeLibcalls.h"
26#include "llvm/MC/StringTableBuilder.h"
27#include "llvm/Object/ModuleSymbolTable.h"
28#include "llvm/Object/SymbolicFile.h"
29#include "llvm/Support/Allocator.h"
30#include "llvm/Support/Casting.h"
31#include "llvm/Support/CommandLine.h"
32#include "llvm/Support/Error.h"
33#include "llvm/Support/StringSaver.h"
34#include "llvm/Support/VCSRevision.h"
35#include "llvm/Support/raw_ostream.h"
36#include "llvm/TargetParser/Triple.h"
37#include <cassert>
38#include <string>
39#include <utility>
40#include <vector>
41
42using namespace llvm;
43using namespace irsymtab;
44
45static cl::opt<bool> DisableBitcodeVersionUpgrade(
46 "disable-bitcode-version-upgrade", cl::Hidden,
47 cl::desc("Disable automatic bitcode upgrade for version mismatch"));
48
49static const char *PreservedSymbols[] = {
50 // There are global variables, so put it here instead of in
51 // RuntimeLibcalls.def.
52 // TODO: Are there similar such variables?
53 "__ssp_canary_word",
54 "__stack_chk_guard",
55};
56
57namespace {
58
59const char *getExpectedProducerName() {
60 static char DefaultName[] = LLVM_VERSION_STRING
61#ifdef LLVM_REVISION
62 " " LLVM_REVISION
63#endif
64 ;
65 // Allows for testing of the irsymtab writer and upgrade mechanism. This
66 // environment variable should not be set by users.
67 if (char *OverrideName = getenv(name: "LLVM_OVERRIDE_PRODUCER"))
68 return OverrideName;
69 return DefaultName;
70}
71
72const char *kExpectedProducerName = getExpectedProducerName();
73
74/// Stores the temporary state that is required to build an IR symbol table.
75struct Builder {
76 SmallVector<char, 0> &Symtab;
77 StringTableBuilder &StrtabBuilder;
78 StringSaver Saver;
79
80 // This ctor initializes a StringSaver using the passed in BumpPtrAllocator.
81 // The StringTableBuilder does not create a copy of any strings added to it,
82 // so this provides somewhere to store any strings that we create.
83 Builder(SmallVector<char, 0> &Symtab, StringTableBuilder &StrtabBuilder,
84 BumpPtrAllocator &Alloc)
85 : Symtab(Symtab), StrtabBuilder(StrtabBuilder), Saver(Alloc) {}
86
87 DenseMap<const Comdat *, int> ComdatMap;
88 Mangler Mang;
89 Triple TT;
90
91 std::vector<storage::Comdat> Comdats;
92 std::vector<storage::Module> Mods;
93 std::vector<storage::Symbol> Syms;
94 std::vector<storage::Uncommon> Uncommons;
95
96 std::string COFFLinkerOpts;
97 raw_string_ostream COFFLinkerOptsOS{COFFLinkerOpts};
98
99 std::vector<storage::Str> DependentLibraries;
100
101 void setStr(storage::Str &S, StringRef Value) {
102 S.Offset = StrtabBuilder.add(S: Value);
103 S.Size = Value.size();
104 }
105
106 template <typename T>
107 void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
108 R.Offset = Symtab.size();
109 R.Size = Objs.size();
110 Symtab.insert(I: Symtab.end(), From: reinterpret_cast<const char *>(Objs.data()),
111 To: reinterpret_cast<const char *>(Objs.data() + Objs.size()));
112 }
113
114 Expected<int> getComdatIndex(const Comdat *C, const Module *M);
115
116 Error addModule(Module *M);
117 Error addSymbol(const ModuleSymbolTable &Msymtab,
118 const SmallPtrSet<GlobalValue *, 4> &Used,
119 ModuleSymbolTable::Symbol Sym);
120
121 Error build(ArrayRef<Module *> Mods);
122};
123
124Error Builder::addModule(Module *M) {
125 if (M->getDataLayoutStr().empty())
126 return make_error<StringError>(Args: "input module has no datalayout",
127 Args: inconvertibleErrorCode());
128
129 // Symbols in the llvm.used list will get the FB_Used bit and will not be
130 // internalized. We do this for llvm.compiler.used as well:
131 //
132 // IR symbol table tracks module-level asm symbol references but not inline
133 // asm. A symbol only referenced by inline asm is not in the IR symbol table,
134 // so we may not know that the definition (in another translation unit) is
135 // referenced. That definition may have __attribute__((used)) (which lowers to
136 // llvm.compiler.used on ELF targets) to communicate to the compiler that it
137 // may be used by inline asm. The usage is perfectly fine, so we treat
138 // llvm.compiler.used conservatively as llvm.used to work around our own
139 // limitation.
140 SmallVector<GlobalValue *, 4> UsedV;
141 collectUsedGlobalVariables(M: *M, Vec&: UsedV, /*CompilerUsed=*/false);
142 collectUsedGlobalVariables(M: *M, Vec&: UsedV, /*CompilerUsed=*/true);
143 SmallPtrSet<GlobalValue *, 4> Used(UsedV.begin(), UsedV.end());
144
145 ModuleSymbolTable Msymtab;
146 Msymtab.addModule(M);
147
148 storage::Module Mod;
149 Mod.Begin = Syms.size();
150 Mod.End = Syms.size() + Msymtab.symbols().size();
151 Mod.UncBegin = Uncommons.size();
152 Mods.push_back(x: Mod);
153
154 if (TT.isOSBinFormatCOFF()) {
155 if (auto E = M->materializeMetadata())
156 return E;
157 if (NamedMDNode *LinkerOptions =
158 M->getNamedMetadata(Name: "llvm.linker.options")) {
159 for (MDNode *MDOptions : LinkerOptions->operands())
160 for (const MDOperand &MDOption : cast<MDNode>(Val: MDOptions)->operands())
161 COFFLinkerOptsOS << " " << cast<MDString>(Val: MDOption)->getString();
162 }
163 }
164
165 if (TT.isOSBinFormatELF()) {
166 if (auto E = M->materializeMetadata())
167 return E;
168 if (NamedMDNode *N = M->getNamedMetadata(Name: "llvm.dependent-libraries")) {
169 for (MDNode *MDOptions : N->operands()) {
170 const auto OperandStr =
171 cast<MDString>(Val: cast<MDNode>(Val: MDOptions)->getOperand(I: 0))->getString();
172 storage::Str Specifier;
173 setStr(S&: Specifier, Value: OperandStr);
174 DependentLibraries.emplace_back(args&: Specifier);
175 }
176 }
177 }
178
179 for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
180 if (Error Err = addSymbol(Msymtab, Used, Sym: Msym))
181 return Err;
182
183 return Error::success();
184}
185
186Expected<int> Builder::getComdatIndex(const Comdat *C, const Module *M) {
187 auto P = ComdatMap.insert(KV: std::make_pair(x&: C, y: Comdats.size()));
188 if (P.second) {
189 std::string Name;
190 if (TT.isOSBinFormatCOFF()) {
191 const GlobalValue *GV = M->getNamedValue(Name: C->getName());
192 if (!GV)
193 return make_error<StringError>(Args: "Could not find leader",
194 Args: inconvertibleErrorCode());
195 // Internal leaders do not affect symbol resolution, therefore they do not
196 // appear in the symbol table.
197 if (GV->hasLocalLinkage()) {
198 P.first->second = -1;
199 return -1;
200 }
201 llvm::raw_string_ostream OS(Name);
202 Mang.getNameWithPrefix(OS, GV, CannotUsePrivateLabel: false);
203 } else {
204 Name = std::string(C->getName());
205 }
206
207 storage::Comdat Comdat;
208 setStr(S&: Comdat.Name, Value: Saver.save(S: Name));
209 Comdat.SelectionKind = C->getSelectionKind();
210 Comdats.push_back(x: Comdat);
211 }
212
213 return P.first->second;
214}
215
216static DenseSet<StringRef> buildPreservedSymbolsSet(const Triple &TT) {
217 DenseSet<StringRef> PreservedSymbolSet(std::begin(arr&: PreservedSymbols),
218 std::end(arr&: PreservedSymbols));
219
220 RTLIB::RuntimeLibcallsInfo Libcalls(TT);
221 for (const char *Name : Libcalls.getLibcallNames()) {
222 if (Name)
223 PreservedSymbolSet.insert(V: Name);
224 }
225 return PreservedSymbolSet;
226}
227
228Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
229 const SmallPtrSet<GlobalValue *, 4> &Used,
230 ModuleSymbolTable::Symbol Msym) {
231 Syms.emplace_back();
232 storage::Symbol &Sym = Syms.back();
233 Sym = {};
234
235 storage::Uncommon *Unc = nullptr;
236 auto Uncommon = [&]() -> storage::Uncommon & {
237 if (Unc)
238 return *Unc;
239 Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
240 Uncommons.emplace_back();
241 Unc = &Uncommons.back();
242 *Unc = {};
243 setStr(S&: Unc->COFFWeakExternFallbackName, Value: "");
244 setStr(S&: Unc->SectionName, Value: "");
245 return *Unc;
246 };
247
248 SmallString<64> Name;
249 {
250 raw_svector_ostream OS(Name);
251 Msymtab.printSymbolName(OS, S: Msym);
252 }
253 setStr(S&: Sym.Name, Value: Saver.save(S: Name.str()));
254
255 auto Flags = Msymtab.getSymbolFlags(S: Msym);
256 if (Flags & object::BasicSymbolRef::SF_Undefined)
257 Sym.Flags |= 1 << storage::Symbol::FB_undefined;
258 if (Flags & object::BasicSymbolRef::SF_Weak)
259 Sym.Flags |= 1 << storage::Symbol::FB_weak;
260 if (Flags & object::BasicSymbolRef::SF_Common)
261 Sym.Flags |= 1 << storage::Symbol::FB_common;
262 if (Flags & object::BasicSymbolRef::SF_Indirect)
263 Sym.Flags |= 1 << storage::Symbol::FB_indirect;
264 if (Flags & object::BasicSymbolRef::SF_Global)
265 Sym.Flags |= 1 << storage::Symbol::FB_global;
266 if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
267 Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
268 if (Flags & object::BasicSymbolRef::SF_Executable)
269 Sym.Flags |= 1 << storage::Symbol::FB_executable;
270
271 Sym.ComdatIndex = -1;
272 auto *GV = dyn_cast_if_present<GlobalValue *>(Val&: Msym);
273 if (!GV) {
274 // Undefined module asm symbols act as GC roots and are implicitly used.
275 if (Flags & object::BasicSymbolRef::SF_Undefined)
276 Sym.Flags |= 1 << storage::Symbol::FB_used;
277 setStr(S&: Sym.IRName, Value: "");
278 return Error::success();
279 }
280
281 setStr(S&: Sym.IRName, Value: GV->getName());
282
283 static const DenseSet<StringRef> PreservedSymbolsSet =
284 buildPreservedSymbolsSet(
285 TT: llvm::Triple(GV->getParent()->getTargetTriple()));
286 bool IsPreservedSymbol = PreservedSymbolsSet.contains(V: GV->getName());
287
288 if (Used.count(Ptr: GV) || IsPreservedSymbol)
289 Sym.Flags |= 1 << storage::Symbol::FB_used;
290 if (GV->isThreadLocal())
291 Sym.Flags |= 1 << storage::Symbol::FB_tls;
292 if (GV->hasGlobalUnnamedAddr())
293 Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
294 if (GV->canBeOmittedFromSymbolTable())
295 Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
296 Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
297
298 if (Flags & object::BasicSymbolRef::SF_Common) {
299 auto *GVar = dyn_cast<GlobalVariable>(Val: GV);
300 if (!GVar)
301 return make_error<StringError>(Args: "Only variables can have common linkage!",
302 Args: inconvertibleErrorCode());
303 Uncommon().CommonSize =
304 GV->getDataLayout().getTypeAllocSize(Ty: GV->getValueType());
305 Uncommon().CommonAlign = GVar->getAlign() ? GVar->getAlign()->value() : 0;
306 }
307
308 const GlobalObject *GO = GV->getAliaseeObject();
309 if (!GO) {
310 if (isa<GlobalIFunc>(Val: GV))
311 GO = cast<GlobalIFunc>(Val: GV)->getResolverFunction();
312 if (!GO)
313 return make_error<StringError>(Args: "Unable to determine comdat of alias!",
314 Args: inconvertibleErrorCode());
315 }
316 if (const Comdat *C = GO->getComdat()) {
317 Expected<int> ComdatIndexOrErr = getComdatIndex(C, M: GV->getParent());
318 if (!ComdatIndexOrErr)
319 return ComdatIndexOrErr.takeError();
320 Sym.ComdatIndex = *ComdatIndexOrErr;
321 }
322
323 if (TT.isOSBinFormatCOFF()) {
324 emitLinkerFlagsForGlobalCOFF(OS&: COFFLinkerOptsOS, GV, TT, Mangler&: Mang);
325
326 if ((Flags & object::BasicSymbolRef::SF_Weak) &&
327 (Flags & object::BasicSymbolRef::SF_Indirect)) {
328 auto *Fallback = dyn_cast<GlobalValue>(
329 Val: cast<GlobalAlias>(Val: GV)->getAliasee()->stripPointerCasts());
330 if (!Fallback)
331 return make_error<StringError>(Args: "Invalid weak external",
332 Args: inconvertibleErrorCode());
333 std::string FallbackName;
334 raw_string_ostream OS(FallbackName);
335 Msymtab.printSymbolName(OS, S: Fallback);
336 OS.flush();
337 setStr(S&: Uncommon().COFFWeakExternFallbackName, Value: Saver.save(S: FallbackName));
338 }
339 }
340
341 if (!GO->getSection().empty())
342 setStr(S&: Uncommon().SectionName, Value: Saver.save(S: GO->getSection()));
343
344 return Error::success();
345}
346
347Error Builder::build(ArrayRef<Module *> IRMods) {
348 storage::Header Hdr;
349
350 assert(!IRMods.empty());
351 Hdr.Version = storage::Header::kCurrentVersion;
352 setStr(S&: Hdr.Producer, Value: kExpectedProducerName);
353 setStr(S&: Hdr.TargetTriple, Value: IRMods[0]->getTargetTriple());
354 setStr(S&: Hdr.SourceFileName, Value: IRMods[0]->getSourceFileName());
355 TT = Triple(IRMods[0]->getTargetTriple());
356
357 for (auto *M : IRMods)
358 if (Error Err = addModule(M))
359 return Err;
360
361 COFFLinkerOptsOS.flush();
362 setStr(S&: Hdr.COFFLinkerOpts, Value: Saver.save(S: COFFLinkerOpts));
363
364 // We are about to fill in the header's range fields, so reserve space for it
365 // and copy it in afterwards.
366 Symtab.resize(N: sizeof(storage::Header));
367 writeRange(R&: Hdr.Modules, Objs: Mods);
368 writeRange(R&: Hdr.Comdats, Objs: Comdats);
369 writeRange(R&: Hdr.Symbols, Objs: Syms);
370 writeRange(R&: Hdr.Uncommons, Objs: Uncommons);
371 writeRange(R&: Hdr.DependentLibraries, Objs: DependentLibraries);
372 *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
373 return Error::success();
374}
375
376} // end anonymous namespace
377
378Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
379 StringTableBuilder &StrtabBuilder,
380 BumpPtrAllocator &Alloc) {
381 return Builder(Symtab, StrtabBuilder, Alloc).build(IRMods: Mods);
382}
383
384// Upgrade a vector of bitcode modules created by an old version of LLVM by
385// creating an irsymtab for them in the current format.
386static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
387 FileContents FC;
388
389 LLVMContext Ctx;
390 std::vector<Module *> Mods;
391 std::vector<std::unique_ptr<Module>> OwnedMods;
392 for (auto BM : BMs) {
393 Expected<std::unique_ptr<Module>> MOrErr =
394 BM.getLazyModule(Context&: Ctx, /*ShouldLazyLoadMetadata*/ true,
395 /*IsImporting*/ false);
396 if (!MOrErr)
397 return MOrErr.takeError();
398
399 Mods.push_back(x: MOrErr->get());
400 OwnedMods.push_back(x: std::move(*MOrErr));
401 }
402
403 StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
404 BumpPtrAllocator Alloc;
405 if (Error E = build(Mods, Symtab&: FC.Symtab, StrtabBuilder, Alloc))
406 return std::move(E);
407
408 StrtabBuilder.finalizeInOrder();
409 FC.Strtab.resize(N: StrtabBuilder.getSize());
410 StrtabBuilder.write(Buf: (uint8_t *)FC.Strtab.data());
411
412 FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
413 {FC.Strtab.data(), FC.Strtab.size()}};
414 return std::move(FC);
415}
416
417Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
418 if (BFC.Mods.empty())
419 return make_error<StringError>(Args: "Bitcode file does not contain any modules",
420 Args: inconvertibleErrorCode());
421
422 if (!DisableBitcodeVersionUpgrade) {
423 if (BFC.StrtabForSymtab.empty() ||
424 BFC.Symtab.size() < sizeof(storage::Header))
425 return upgrade(BMs: BFC.Mods);
426
427 // We cannot use the regular reader to read the version and producer,
428 // because it will expect the header to be in the current format. The only
429 // thing we can rely on is that the version and producer will be present as
430 // the first struct elements.
431 auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
432 unsigned Version = Hdr->Version;
433 StringRef Producer = Hdr->Producer.get(Strtab: BFC.StrtabForSymtab);
434 if (Version != storage::Header::kCurrentVersion ||
435 Producer != kExpectedProducerName)
436 return upgrade(BMs: BFC.Mods);
437 }
438
439 FileContents FC;
440 FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
441 {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
442
443 // Finally, make sure that the number of modules in the symbol table matches
444 // the number of modules in the bitcode file. If they differ, it may mean that
445 // the bitcode file was created by binary concatenation, so we need to create
446 // a new symbol table from scratch.
447 if (FC.TheReader.getNumModules() != BFC.Mods.size())
448 return upgrade(BMs: std::move(BFC.Mods));
449
450 return std::move(FC);
451}
452