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/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/Bitcode/BitcodeReader.h"
26 #include "llvm/MC/StringTableBuilder.h"
27 #include "llvm/Object/IRObjectFile.h"
28 #include "llvm/Object/ModuleSymbolTable.h"
29 #include "llvm/Object/SymbolicFile.h"
30 #include "llvm/Support/Allocator.h"
31 #include "llvm/Support/Casting.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 static const char *LibcallRoutineNames[] = {
45 #define HANDLE_LIBCALL(code, name) name,
46 #include "llvm/IR/RuntimeLibcalls.def"
47 #undef HANDLE_LIBCALL
48 };
49 
50 namespace {
51 
52 const char *getExpectedProducerName() {
53   static char DefaultName[] = LLVM_VERSION_STRING
54 #ifdef LLVM_REVISION
55       " " LLVM_REVISION
56 #endif
57       ;
58   // Allows for testing of the irsymtab writer and upgrade mechanism. This
59   // environment variable should not be set by users.
60   if (char *OverrideName = getenv("LLVM_OVERRIDE_PRODUCER"))
61     return OverrideName;
62   return DefaultName;
63 }
64 
65 const char *kExpectedProducerName = getExpectedProducerName();
66 
67 /// Stores the temporary state that is required to build an IR symbol table.
68 struct Builder {
69   SmallVector<char, 0> &Symtab;
70   StringTableBuilder &StrtabBuilder;
71   StringSaver Saver;
72 
73   // This ctor initializes a StringSaver using the passed in BumpPtrAllocator.
74   // The StringTableBuilder does not create a copy of any strings added to it,
75   // so this provides somewhere to store any strings that we create.
76   Builder(SmallVector<char, 0> &Symtab, StringTableBuilder &StrtabBuilder,
77           BumpPtrAllocator &Alloc)
78       : Symtab(Symtab), StrtabBuilder(StrtabBuilder), Saver(Alloc) {}
79 
80   DenseMap<const Comdat *, int> ComdatMap;
81   Mangler Mang;
82   Triple TT;
83 
84   std::vector<storage::Comdat> Comdats;
85   std::vector<storage::Module> Mods;
86   std::vector<storage::Symbol> Syms;
87   std::vector<storage::Uncommon> Uncommons;
88 
89   std::string COFFLinkerOpts;
90   raw_string_ostream COFFLinkerOptsOS{COFFLinkerOpts};
91 
92   std::vector<storage::Str> DependentLibraries;
93 
94   void setStr(storage::Str &S, StringRef Value) {
95     S.Offset = StrtabBuilder.add(Value);
96     S.Size = Value.size();
97   }
98 
99   template <typename T>
100   void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
101     R.Offset = Symtab.size();
102     R.Size = Objs.size();
103     Symtab.insert(Symtab.end(), reinterpret_cast<const char *>(Objs.data()),
104                   reinterpret_cast<const char *>(Objs.data() + Objs.size()));
105   }
106 
107   Expected<int> getComdatIndex(const Comdat *C, const Module *M);
108 
109   Error addModule(Module *M);
110   Error addSymbol(const ModuleSymbolTable &Msymtab,
111                   const SmallPtrSet<GlobalValue *, 8> &Used,
112                   ModuleSymbolTable::Symbol Sym);
113 
114   Error build(ArrayRef<Module *> Mods);
115 };
116 
117 Error Builder::addModule(Module *M) {
118   if (M->getDataLayoutStr().empty())
119     return make_error<StringError>("input module has no datalayout",
120                                    inconvertibleErrorCode());
121 
122   SmallPtrSet<GlobalValue *, 8> Used;
123   collectUsedGlobalVariables(*M, Used, /*CompilerUsed*/ false);
124 
125   ModuleSymbolTable Msymtab;
126   Msymtab.addModule(M);
127 
128   storage::Module Mod;
129   Mod.Begin = Syms.size();
130   Mod.End = Syms.size() + Msymtab.symbols().size();
131   Mod.UncBegin = Uncommons.size();
132   Mods.push_back(Mod);
133 
134   if (TT.isOSBinFormatCOFF()) {
135     if (auto E = M->materializeMetadata())
136       return E;
137     if (NamedMDNode *LinkerOptions =
138             M->getNamedMetadata("llvm.linker.options")) {
139       for (MDNode *MDOptions : LinkerOptions->operands())
140         for (const MDOperand &MDOption : cast<MDNode>(MDOptions)->operands())
141           COFFLinkerOptsOS << " " << cast<MDString>(MDOption)->getString();
142     }
143   }
144 
145   if (TT.isOSBinFormatELF()) {
146     if (auto E = M->materializeMetadata())
147       return E;
148     if (NamedMDNode *N = M->getNamedMetadata("llvm.dependent-libraries")) {
149       for (MDNode *MDOptions : N->operands()) {
150         const auto OperandStr =
151             cast<MDString>(cast<MDNode>(MDOptions)->getOperand(0))->getString();
152         storage::Str Specifier;
153         setStr(Specifier, OperandStr);
154         DependentLibraries.emplace_back(Specifier);
155       }
156     }
157   }
158 
159   for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
160     if (Error Err = addSymbol(Msymtab, Used, Msym))
161       return Err;
162 
163   return Error::success();
164 }
165 
166 Expected<int> Builder::getComdatIndex(const Comdat *C, const Module *M) {
167   auto P = ComdatMap.insert(std::make_pair(C, Comdats.size()));
168   if (P.second) {
169     std::string Name;
170     if (TT.isOSBinFormatCOFF()) {
171       const GlobalValue *GV = M->getNamedValue(C->getName());
172       if (!GV)
173         return make_error<StringError>("Could not find leader",
174                                        inconvertibleErrorCode());
175       // Internal leaders do not affect symbol resolution, therefore they do not
176       // appear in the symbol table.
177       if (GV->hasLocalLinkage()) {
178         P.first->second = -1;
179         return -1;
180       }
181       llvm::raw_string_ostream OS(Name);
182       Mang.getNameWithPrefix(OS, GV, false);
183     } else {
184       Name = std::string(C->getName());
185     }
186 
187     storage::Comdat Comdat;
188     setStr(Comdat.Name, Saver.save(Name));
189     Comdats.push_back(Comdat);
190   }
191 
192   return P.first->second;
193 }
194 
195 Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
196                          const SmallPtrSet<GlobalValue *, 8> &Used,
197                          ModuleSymbolTable::Symbol Msym) {
198   Syms.emplace_back();
199   storage::Symbol &Sym = Syms.back();
200   Sym = {};
201 
202   storage::Uncommon *Unc = nullptr;
203   auto Uncommon = [&]() -> storage::Uncommon & {
204     if (Unc)
205       return *Unc;
206     Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
207     Uncommons.emplace_back();
208     Unc = &Uncommons.back();
209     *Unc = {};
210     setStr(Unc->COFFWeakExternFallbackName, "");
211     setStr(Unc->SectionName, "");
212     return *Unc;
213   };
214 
215   SmallString<64> Name;
216   {
217     raw_svector_ostream OS(Name);
218     Msymtab.printSymbolName(OS, Msym);
219   }
220   setStr(Sym.Name, Saver.save(StringRef(Name)));
221 
222   auto Flags = Msymtab.getSymbolFlags(Msym);
223   if (Flags & object::BasicSymbolRef::SF_Undefined)
224     Sym.Flags |= 1 << storage::Symbol::FB_undefined;
225   if (Flags & object::BasicSymbolRef::SF_Weak)
226     Sym.Flags |= 1 << storage::Symbol::FB_weak;
227   if (Flags & object::BasicSymbolRef::SF_Common)
228     Sym.Flags |= 1 << storage::Symbol::FB_common;
229   if (Flags & object::BasicSymbolRef::SF_Indirect)
230     Sym.Flags |= 1 << storage::Symbol::FB_indirect;
231   if (Flags & object::BasicSymbolRef::SF_Global)
232     Sym.Flags |= 1 << storage::Symbol::FB_global;
233   if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
234     Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
235   if (Flags & object::BasicSymbolRef::SF_Executable)
236     Sym.Flags |= 1 << storage::Symbol::FB_executable;
237 
238   Sym.ComdatIndex = -1;
239   auto *GV = Msym.dyn_cast<GlobalValue *>();
240   if (!GV) {
241     // Undefined module asm symbols act as GC roots and are implicitly used.
242     if (Flags & object::BasicSymbolRef::SF_Undefined)
243       Sym.Flags |= 1 << storage::Symbol::FB_used;
244     setStr(Sym.IRName, "");
245     return Error::success();
246   }
247 
248   setStr(Sym.IRName, GV->getName());
249 
250   bool IsBuiltinFunc = false;
251 
252   for (const char *LibcallName : LibcallRoutineNames)
253     if (GV->getName() == LibcallName)
254       IsBuiltinFunc = true;
255 
256   if (Used.count(GV) || IsBuiltinFunc)
257     Sym.Flags |= 1 << storage::Symbol::FB_used;
258   if (GV->isThreadLocal())
259     Sym.Flags |= 1 << storage::Symbol::FB_tls;
260   if (GV->hasGlobalUnnamedAddr())
261     Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
262   if (GV->canBeOmittedFromSymbolTable())
263     Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
264   Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
265 
266   if (Flags & object::BasicSymbolRef::SF_Common) {
267     auto *GVar = dyn_cast<GlobalVariable>(GV);
268     if (!GVar)
269       return make_error<StringError>("Only variables can have common linkage!",
270                                      inconvertibleErrorCode());
271     Uncommon().CommonSize = GV->getParent()->getDataLayout().getTypeAllocSize(
272         GV->getType()->getElementType());
273     Uncommon().CommonAlign = GVar->getAlignment();
274   }
275 
276   const GlobalObject *Base = GV->getBaseObject();
277   if (!Base)
278     return make_error<StringError>("Unable to determine comdat of alias!",
279                                    inconvertibleErrorCode());
280   if (const Comdat *C = Base->getComdat()) {
281     Expected<int> ComdatIndexOrErr = getComdatIndex(C, GV->getParent());
282     if (!ComdatIndexOrErr)
283       return ComdatIndexOrErr.takeError();
284     Sym.ComdatIndex = *ComdatIndexOrErr;
285   }
286 
287   if (TT.isOSBinFormatCOFF()) {
288     emitLinkerFlagsForGlobalCOFF(COFFLinkerOptsOS, GV, TT, Mang);
289 
290     if ((Flags & object::BasicSymbolRef::SF_Weak) &&
291         (Flags & object::BasicSymbolRef::SF_Indirect)) {
292       auto *Fallback = dyn_cast<GlobalValue>(
293           cast<GlobalAlias>(GV)->getAliasee()->stripPointerCasts());
294       if (!Fallback)
295         return make_error<StringError>("Invalid weak external",
296                                        inconvertibleErrorCode());
297       std::string FallbackName;
298       raw_string_ostream OS(FallbackName);
299       Msymtab.printSymbolName(OS, Fallback);
300       OS.flush();
301       setStr(Uncommon().COFFWeakExternFallbackName, Saver.save(FallbackName));
302     }
303   }
304 
305   if (!Base->getSection().empty())
306     setStr(Uncommon().SectionName, Saver.save(Base->getSection()));
307 
308   return Error::success();
309 }
310 
311 Error Builder::build(ArrayRef<Module *> IRMods) {
312   storage::Header Hdr;
313 
314   assert(!IRMods.empty());
315   Hdr.Version = storage::Header::kCurrentVersion;
316   setStr(Hdr.Producer, kExpectedProducerName);
317   setStr(Hdr.TargetTriple, IRMods[0]->getTargetTriple());
318   setStr(Hdr.SourceFileName, IRMods[0]->getSourceFileName());
319   TT = Triple(IRMods[0]->getTargetTriple());
320 
321   for (auto *M : IRMods)
322     if (Error Err = addModule(M))
323       return Err;
324 
325   COFFLinkerOptsOS.flush();
326   setStr(Hdr.COFFLinkerOpts, Saver.save(COFFLinkerOpts));
327 
328   // We are about to fill in the header's range fields, so reserve space for it
329   // and copy it in afterwards.
330   Symtab.resize(sizeof(storage::Header));
331   writeRange(Hdr.Modules, Mods);
332   writeRange(Hdr.Comdats, Comdats);
333   writeRange(Hdr.Symbols, Syms);
334   writeRange(Hdr.Uncommons, Uncommons);
335   writeRange(Hdr.DependentLibraries, DependentLibraries);
336   *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
337   return Error::success();
338 }
339 
340 } // end anonymous namespace
341 
342 Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
343                       StringTableBuilder &StrtabBuilder,
344                       BumpPtrAllocator &Alloc) {
345   return Builder(Symtab, StrtabBuilder, Alloc).build(Mods);
346 }
347 
348 // Upgrade a vector of bitcode modules created by an old version of LLVM by
349 // creating an irsymtab for them in the current format.
350 static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
351   FileContents FC;
352 
353   LLVMContext Ctx;
354   std::vector<Module *> Mods;
355   std::vector<std::unique_ptr<Module>> OwnedMods;
356   for (auto BM : BMs) {
357     Expected<std::unique_ptr<Module>> MOrErr =
358         BM.getLazyModule(Ctx, /*ShouldLazyLoadMetadata*/ true,
359                          /*IsImporting*/ false);
360     if (!MOrErr)
361       return MOrErr.takeError();
362 
363     Mods.push_back(MOrErr->get());
364     OwnedMods.push_back(std::move(*MOrErr));
365   }
366 
367   StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
368   BumpPtrAllocator Alloc;
369   if (Error E = build(Mods, FC.Symtab, StrtabBuilder, Alloc))
370     return std::move(E);
371 
372   StrtabBuilder.finalizeInOrder();
373   FC.Strtab.resize(StrtabBuilder.getSize());
374   StrtabBuilder.write((uint8_t *)FC.Strtab.data());
375 
376   FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
377                   {FC.Strtab.data(), FC.Strtab.size()}};
378   return std::move(FC);
379 }
380 
381 Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
382   if (BFC.Mods.empty())
383     return make_error<StringError>("Bitcode file does not contain any modules",
384                                    inconvertibleErrorCode());
385 
386   if (BFC.StrtabForSymtab.empty() ||
387       BFC.Symtab.size() < sizeof(storage::Header))
388     return upgrade(BFC.Mods);
389 
390   // We cannot use the regular reader to read the version and producer, because
391   // it will expect the header to be in the current format. The only thing we
392   // can rely on is that the version and producer will be present as the first
393   // struct elements.
394   auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
395   unsigned Version = Hdr->Version;
396   StringRef Producer = Hdr->Producer.get(BFC.StrtabForSymtab);
397   if (Version != storage::Header::kCurrentVersion ||
398       Producer != kExpectedProducerName)
399     return upgrade(BFC.Mods);
400 
401   FileContents FC;
402   FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
403                   {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
404 
405   // Finally, make sure that the number of modules in the symbol table matches
406   // the number of modules in the bitcode file. If they differ, it may mean that
407   // the bitcode file was created by binary concatenation, so we need to create
408   // a new symbol table from scratch.
409   if (FC.TheReader.getNumModules() != BFC.Mods.size())
410     return upgrade(std::move(BFC.Mods));
411 
412   return std::move(FC);
413 }
414