blob: 668cf62d1e82b01d331f22e4e6767cd6c61f12d0 [file] [log] [blame]
Eugene Zelenko1df42fa2017-04-24 23:21:38 +00001//===- IRSymtab.cpp - implementation of IR symbol tables ------------------===//
Peter Collingbourne7b30f162017-03-31 04:47:07 +00002//
3// The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9
Chandler Carruth6bda14b2017-06-06 11:49:48 +000010#include "llvm/Object/IRSymtab.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000011#include "llvm/ADT/ArrayRef.h"
12#include "llvm/ADT/DenseMap.h"
13#include "llvm/ADT/SmallPtrSet.h"
14#include "llvm/ADT/SmallString.h"
15#include "llvm/ADT/SmallVector.h"
16#include "llvm/ADT/StringRef.h"
17#include "llvm/ADT/Triple.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000018#include "llvm/Analysis/ObjectUtils.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000019#include "llvm/IR/Comdat.h"
20#include "llvm/IR/DataLayout.h"
21#include "llvm/IR/GlobalAlias.h"
22#include "llvm/IR/GlobalObject.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000023#include "llvm/IR/Mangler.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000024#include "llvm/IR/Metadata.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000025#include "llvm/IR/Module.h"
Peter Collingbournec00c2b22017-06-08 01:26:14 +000026#include "llvm/Bitcode/BitcodeReader.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000027#include "llvm/MC/StringTableBuilder.h"
Peter Collingbournec00c2b22017-06-08 01:26:14 +000028#include "llvm/Object/IRObjectFile.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000029#include "llvm/Object/ModuleSymbolTable.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000030#include "llvm/Object/SymbolicFile.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000031#include "llvm/Support/Allocator.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000032#include "llvm/Support/Casting.h"
33#include "llvm/Support/Error.h"
Peter Collingbourne7b30f162017-03-31 04:47:07 +000034#include "llvm/Support/StringSaver.h"
Peter Collingbourne53ed8672017-06-27 23:49:58 +000035#include "llvm/Support/VCSRevision.h"
Chandler Carruth6bda14b2017-06-06 11:49:48 +000036#include "llvm/Support/raw_ostream.h"
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000037#include <cassert>
38#include <string>
39#include <utility>
40#include <vector>
Peter Collingbourne7b30f162017-03-31 04:47:07 +000041
42using namespace llvm;
43using namespace irsymtab;
44
45namespace {
46
Peter Collingbourne53ed8672017-06-27 23:49:58 +000047const char *getExpectedProducerName() {
48 static char DefaultName[] = LLVM_VERSION_STRING
49#ifdef LLVM_REVISION
50 " " LLVM_REVISION
51#endif
52 ;
53 // Allows for testing of the irsymtab writer and upgrade mechanism. This
54 // environment variable should not be set by users.
55 if (char *OverrideName = getenv("LLVM_OVERRIDE_PRODUCER"))
56 return OverrideName;
57 return DefaultName;
58}
59
60const char *kExpectedProducerName = getExpectedProducerName();
61
Peter Collingbourne7b30f162017-03-31 04:47:07 +000062/// Stores the temporary state that is required to build an IR symbol table.
63struct Builder {
64 SmallVector<char, 0> &Symtab;
Peter Collingbourneafaeed52017-06-21 18:23:19 +000065 StringTableBuilder &StrtabBuilder;
66 StringSaver Saver;
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000067
Peter Collingbourneafaeed52017-06-21 18:23:19 +000068 // This ctor initializes a StringSaver using the passed in BumpPtrAllocator.
69 // The StringTableBuilder does not create a copy of any strings added to it,
70 // so this provides somewhere to store any strings that we create.
71 Builder(SmallVector<char, 0> &Symtab, StringTableBuilder &StrtabBuilder,
72 BumpPtrAllocator &Alloc)
73 : Symtab(Symtab), StrtabBuilder(StrtabBuilder), Saver(Alloc) {}
Peter Collingbourne7b30f162017-03-31 04:47:07 +000074
75 DenseMap<const Comdat *, unsigned> ComdatMap;
Peter Collingbourne7b30f162017-03-31 04:47:07 +000076 Mangler Mang;
77 Triple TT;
78
79 std::vector<storage::Comdat> Comdats;
80 std::vector<storage::Module> Mods;
81 std::vector<storage::Symbol> Syms;
82 std::vector<storage::Uncommon> Uncommons;
83
84 std::string COFFLinkerOpts;
85 raw_string_ostream COFFLinkerOptsOS{COFFLinkerOpts};
86
87 void setStr(storage::Str &S, StringRef Value) {
88 S.Offset = StrtabBuilder.add(Value);
Peter Collingbournec74cf062017-04-17 17:55:24 +000089 S.Size = Value.size();
Peter Collingbourne7b30f162017-03-31 04:47:07 +000090 }
Eugene Zelenko1df42fa2017-04-24 23:21:38 +000091
Peter Collingbourne7b30f162017-03-31 04:47:07 +000092 template <typename T>
93 void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
94 R.Offset = Symtab.size();
95 R.Size = Objs.size();
96 Symtab.insert(Symtab.end(), reinterpret_cast<const char *>(Objs.data()),
97 reinterpret_cast<const char *>(Objs.data() + Objs.size()));
98 }
99
100 Error addModule(Module *M);
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000101 Error addSymbol(const ModuleSymbolTable &Msymtab,
102 const SmallPtrSet<GlobalValue *, 8> &Used,
103 ModuleSymbolTable::Symbol Sym);
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000104
105 Error build(ArrayRef<Module *> Mods);
106};
107
108Error Builder::addModule(Module *M) {
Peter Collingbournedc8c0182017-06-08 22:04:24 +0000109 if (M->getDataLayoutStr().empty())
110 return make_error<StringError>("input module has no datalayout",
111 inconvertibleErrorCode());
112
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000113 SmallPtrSet<GlobalValue *, 8> Used;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000114 collectUsedGlobalVariables(*M, Used, /*CompilerUsed*/ false);
115
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000116 ModuleSymbolTable Msymtab;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000117 Msymtab.addModule(M);
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000118
119 storage::Module Mod;
120 Mod.Begin = Syms.size();
121 Mod.End = Syms.size() + Msymtab.symbols().size();
122 Mod.UncBegin = Uncommons.size();
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000123 Mods.push_back(Mod);
124
125 if (TT.isOSBinFormatCOFF()) {
126 if (auto E = M->materializeMetadata())
127 return E;
Peter Collingbourne89061b22017-06-12 20:10:48 +0000128 if (NamedMDNode *LinkerOptions =
129 M->getNamedMetadata("llvm.linker.options")) {
130 for (MDNode *MDOptions : LinkerOptions->operands())
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000131 for (const MDOperand &MDOption : cast<MDNode>(MDOptions)->operands())
132 COFFLinkerOptsOS << " " << cast<MDString>(MDOption)->getString();
133 }
134 }
135
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000136 for (ModuleSymbolTable::Symbol Msym : Msymtab.symbols())
137 if (Error Err = addSymbol(Msymtab, Used, Msym))
138 return Err;
139
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000140 return Error::success();
141}
142
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000143Error Builder::addSymbol(const ModuleSymbolTable &Msymtab,
144 const SmallPtrSet<GlobalValue *, 8> &Used,
145 ModuleSymbolTable::Symbol Msym) {
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000146 Syms.emplace_back();
147 storage::Symbol &Sym = Syms.back();
148 Sym = {};
149
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000150 storage::Uncommon *Unc = nullptr;
151 auto Uncommon = [&]() -> storage::Uncommon & {
152 if (Unc)
153 return *Unc;
Peter Collingbourne76423dc2017-04-17 23:43:49 +0000154 Sym.Flags |= 1 << storage::Symbol::FB_has_uncommon;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000155 Uncommons.emplace_back();
156 Unc = &Uncommons.back();
157 *Unc = {};
158 setStr(Unc->COFFWeakExternFallbackName, "");
Teresa Johnsona83c3f72017-07-25 19:42:32 +0000159 setStr(Unc->SectionName, "");
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000160 return *Unc;
161 };
162
163 SmallString<64> Name;
164 {
165 raw_svector_ostream OS(Name);
166 Msymtab.printSymbolName(OS, Msym);
167 }
168 setStr(Sym.Name, Saver.save(StringRef(Name)));
169
170 auto Flags = Msymtab.getSymbolFlags(Msym);
171 if (Flags & object::BasicSymbolRef::SF_Undefined)
172 Sym.Flags |= 1 << storage::Symbol::FB_undefined;
173 if (Flags & object::BasicSymbolRef::SF_Weak)
174 Sym.Flags |= 1 << storage::Symbol::FB_weak;
175 if (Flags & object::BasicSymbolRef::SF_Common)
176 Sym.Flags |= 1 << storage::Symbol::FB_common;
177 if (Flags & object::BasicSymbolRef::SF_Indirect)
178 Sym.Flags |= 1 << storage::Symbol::FB_indirect;
179 if (Flags & object::BasicSymbolRef::SF_Global)
180 Sym.Flags |= 1 << storage::Symbol::FB_global;
181 if (Flags & object::BasicSymbolRef::SF_FormatSpecific)
182 Sym.Flags |= 1 << storage::Symbol::FB_format_specific;
Tobias Edler von Koch90df1f482017-04-13 16:24:14 +0000183 if (Flags & object::BasicSymbolRef::SF_Executable)
184 Sym.Flags |= 1 << storage::Symbol::FB_executable;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000185
186 Sym.ComdatIndex = -1;
187 auto *GV = Msym.dyn_cast<GlobalValue *>();
188 if (!GV) {
Peter Collingbournefa58f752017-04-26 17:53:39 +0000189 // Undefined module asm symbols act as GC roots and are implicitly used.
190 if (Flags & object::BasicSymbolRef::SF_Undefined)
191 Sym.Flags |= 1 << storage::Symbol::FB_used;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000192 setStr(Sym.IRName, "");
193 return Error::success();
194 }
195
196 setStr(Sym.IRName, GV->getName());
197
198 if (Used.count(GV))
199 Sym.Flags |= 1 << storage::Symbol::FB_used;
200 if (GV->isThreadLocal())
201 Sym.Flags |= 1 << storage::Symbol::FB_tls;
202 if (GV->hasGlobalUnnamedAddr())
203 Sym.Flags |= 1 << storage::Symbol::FB_unnamed_addr;
204 if (canBeOmittedFromSymbolTable(GV))
205 Sym.Flags |= 1 << storage::Symbol::FB_may_omit;
206 Sym.Flags |= unsigned(GV->getVisibility()) << storage::Symbol::FB_visibility;
207
208 if (Flags & object::BasicSymbolRef::SF_Common) {
209 Uncommon().CommonSize = GV->getParent()->getDataLayout().getTypeAllocSize(
210 GV->getType()->getElementType());
211 Uncommon().CommonAlign = GV->getAlignment();
212 }
213
214 const GlobalObject *Base = GV->getBaseObject();
215 if (!Base)
216 return make_error<StringError>("Unable to determine comdat of alias!",
217 inconvertibleErrorCode());
218 if (const Comdat *C = Base->getComdat()) {
219 auto P = ComdatMap.insert(std::make_pair(C, Comdats.size()));
220 Sym.ComdatIndex = P.first->second;
221
222 if (P.second) {
223 storage::Comdat Comdat;
224 setStr(Comdat.Name, C->getName());
225 Comdats.push_back(Comdat);
226 }
227 }
228
229 if (TT.isOSBinFormatCOFF()) {
230 emitLinkerFlagsForGlobalCOFF(COFFLinkerOptsOS, GV, TT, Mang);
231
232 if ((Flags & object::BasicSymbolRef::SF_Weak) &&
233 (Flags & object::BasicSymbolRef::SF_Indirect)) {
234 std::string FallbackName;
235 raw_string_ostream OS(FallbackName);
236 Msymtab.printSymbolName(
237 OS, cast<GlobalValue>(
238 cast<GlobalAlias>(GV)->getAliasee()->stripPointerCasts()));
239 OS.flush();
240 setStr(Uncommon().COFFWeakExternFallbackName, Saver.save(FallbackName));
241 }
242 }
243
Teresa Johnsona83c3f72017-07-25 19:42:32 +0000244 if (!Base->getSection().empty())
245 setStr(Uncommon().SectionName, Saver.save(Base->getSection()));
246
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000247 return Error::success();
248}
249
250Error Builder::build(ArrayRef<Module *> IRMods) {
251 storage::Header Hdr;
252
253 assert(!IRMods.empty());
Peter Collingbourne53ed8672017-06-27 23:49:58 +0000254 Hdr.Version = storage::Header::kCurrentVersion;
255 setStr(Hdr.Producer, kExpectedProducerName);
Peter Collingbourne8446f1f2017-04-14 02:55:06 +0000256 setStr(Hdr.TargetTriple, IRMods[0]->getTargetTriple());
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000257 setStr(Hdr.SourceFileName, IRMods[0]->getSourceFileName());
258 TT = Triple(IRMods[0]->getTargetTriple());
259
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000260 for (auto *M : IRMods)
261 if (Error Err = addModule(M))
262 return Err;
263
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000264 COFFLinkerOptsOS.flush();
Peter Collingbourneafaeed52017-06-21 18:23:19 +0000265 setStr(Hdr.COFFLinkerOpts, Saver.save(COFFLinkerOpts));
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000266
267 // We are about to fill in the header's range fields, so reserve space for it
268 // and copy it in afterwards.
269 Symtab.resize(sizeof(storage::Header));
270 writeRange(Hdr.Modules, Mods);
271 writeRange(Hdr.Comdats, Comdats);
272 writeRange(Hdr.Symbols, Syms);
273 writeRange(Hdr.Uncommons, Uncommons);
274
275 *reinterpret_cast<storage::Header *>(Symtab.data()) = Hdr;
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000276 return Error::success();
277}
278
Eugene Zelenko1df42fa2017-04-24 23:21:38 +0000279} // end anonymous namespace
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000280
281Error irsymtab::build(ArrayRef<Module *> Mods, SmallVector<char, 0> &Symtab,
Peter Collingbourneafaeed52017-06-21 18:23:19 +0000282 StringTableBuilder &StrtabBuilder,
283 BumpPtrAllocator &Alloc) {
284 return Builder(Symtab, StrtabBuilder, Alloc).build(Mods);
Peter Collingbourne7b30f162017-03-31 04:47:07 +0000285}
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000286
Peter Collingbourne8dde4cb2017-06-08 22:00:24 +0000287// Upgrade a vector of bitcode modules created by an old version of LLVM by
288// creating an irsymtab for them in the current format.
289static Expected<FileContents> upgrade(ArrayRef<BitcodeModule> BMs) {
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000290 FileContents FC;
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000291
292 LLVMContext Ctx;
293 std::vector<Module *> Mods;
294 std::vector<std::unique_ptr<Module>> OwnedMods;
295 for (auto BM : BMs) {
296 Expected<std::unique_ptr<Module>> MOrErr =
297 BM.getLazyModule(Ctx, /*ShouldLazyLoadMetadata*/ true,
298 /*IsImporting*/ false);
299 if (!MOrErr)
300 return MOrErr.takeError();
301
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000302 Mods.push_back(MOrErr->get());
303 OwnedMods.push_back(std::move(*MOrErr));
304 }
305
Peter Collingbourneafaeed52017-06-21 18:23:19 +0000306 StringTableBuilder StrtabBuilder(StringTableBuilder::RAW);
307 BumpPtrAllocator Alloc;
308 if (Error E = build(Mods, FC.Symtab, StrtabBuilder, Alloc))
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000309 return std::move(E);
310
Peter Collingbourneafaeed52017-06-21 18:23:19 +0000311 StrtabBuilder.finalizeInOrder();
312 FC.Strtab.resize(StrtabBuilder.getSize());
313 StrtabBuilder.write((uint8_t *)FC.Strtab.data());
314
Peter Collingbournec00c2b22017-06-08 01:26:14 +0000315 FC.TheReader = {{FC.Symtab.data(), FC.Symtab.size()},
316 {FC.Strtab.data(), FC.Strtab.size()}};
317 return std::move(FC);
318}
Peter Collingbourne8dde4cb2017-06-08 22:00:24 +0000319
320Expected<FileContents> irsymtab::readBitcode(const BitcodeFileContents &BFC) {
321 if (BFC.Mods.empty())
322 return make_error<StringError>("Bitcode file does not contain any modules",
323 inconvertibleErrorCode());
324
Peter Collingbourne99b98c22017-06-27 23:50:24 +0000325 if (BFC.StrtabForSymtab.empty() ||
326 BFC.Symtab.size() < sizeof(storage::Header))
327 return upgrade(BFC.Mods);
328
329 // We cannot use the regular reader to read the version and producer, because
330 // it will expect the header to be in the current format. The only thing we
331 // can rely on is that the version and producer will be present as the first
332 // struct elements.
333 auto *Hdr = reinterpret_cast<const storage::Header *>(BFC.Symtab.data());
334 unsigned Version = Hdr->Version;
335 StringRef Producer = Hdr->Producer.get(BFC.StrtabForSymtab);
336 if (Version != storage::Header::kCurrentVersion ||
337 Producer != kExpectedProducerName)
338 return upgrade(BFC.Mods);
339
340 FileContents FC;
341 FC.TheReader = {{BFC.Symtab.data(), BFC.Symtab.size()},
342 {BFC.StrtabForSymtab.data(), BFC.StrtabForSymtab.size()}};
343
344 // Finally, make sure that the number of modules in the symbol table matches
345 // the number of modules in the bitcode file. If they differ, it may mean that
346 // the bitcode file was created by binary concatenation, so we need to create
347 // a new symbol table from scratch.
348 if (FC.TheReader.getNumModules() != BFC.Mods.size())
349 return upgrade(std::move(BFC.Mods));
350
351 return std::move(FC);
Peter Collingbourne8dde4cb2017-06-08 22:00:24 +0000352}