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