1 //===- SymbolizableObjectFile.cpp -----------------------------------------===//
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 // Implementation of SymbolizableObjectFile class.
10 //
11 //===----------------------------------------------------------------------===//
12 
13 #include "SymbolizableObjectFile.h"
14 #include "llvm/ADT/STLExtras.h"
15 #include "llvm/ADT/StringRef.h"
16 #include "llvm/ADT/Triple.h"
17 #include "llvm/BinaryFormat/COFF.h"
18 #include "llvm/DebugInfo/DWARF/DWARFContext.h"
19 #include "llvm/DebugInfo/Symbolize/SymbolizableModule.h"
20 #include "llvm/Object/COFF.h"
21 #include "llvm/Object/ObjectFile.h"
22 #include "llvm/Object/SymbolSize.h"
23 #include "llvm/Support/Casting.h"
24 #include "llvm/Support/DataExtractor.h"
25 #include "llvm/Support/Error.h"
26 #include <algorithm>
27 #include <cstdint>
28 #include <memory>
29 #include <string>
30 #include <system_error>
31 #include <utility>
32 #include <vector>
33 
34 using namespace llvm;
35 using namespace object;
36 using namespace symbolize;
37 
38 static DILineInfoSpecifier
39 getDILineInfoSpecifier(FunctionNameKind FNKind) {
40   return DILineInfoSpecifier(
41       DILineInfoSpecifier::FileLineInfoKind::AbsoluteFilePath, FNKind);
42 }
43 
44 ErrorOr<std::unique_ptr<SymbolizableObjectFile>>
45 SymbolizableObjectFile::create(object::ObjectFile *Obj,
46                                std::unique_ptr<DIContext> DICtx) {
47   std::unique_ptr<SymbolizableObjectFile> res(
48       new SymbolizableObjectFile(Obj, std::move(DICtx)));
49   std::unique_ptr<DataExtractor> OpdExtractor;
50   uint64_t OpdAddress = 0;
51   // Find the .opd (function descriptor) section if any, for big-endian
52   // PowerPC64 ELF.
53   if (Obj->getArch() == Triple::ppc64) {
54     for (section_iterator Section : Obj->sections()) {
55       StringRef Name;
56       StringRef Data;
57       if (auto EC = Section->getName(Name))
58         return EC;
59       if (Name == ".opd") {
60         if (auto EC = Section->getContents(Data))
61           return EC;
62         OpdExtractor.reset(new DataExtractor(Data, Obj->isLittleEndian(),
63                                              Obj->getBytesInAddress()));
64         OpdAddress = Section->getAddress();
65         break;
66       }
67     }
68   }
69   std::vector<std::pair<SymbolRef, uint64_t>> Symbols =
70       computeSymbolSizes(*Obj);
71   for (auto &P : Symbols)
72     res->addSymbol(P.first, P.second, OpdExtractor.get(), OpdAddress);
73 
74   // If this is a COFF object and we didn't find any symbols, try the export
75   // table.
76   if (Symbols.empty()) {
77     if (auto *CoffObj = dyn_cast<COFFObjectFile>(Obj))
78       if (auto EC = res->addCoffExportSymbols(CoffObj))
79         return EC;
80   }
81   return std::move(res);
82 }
83 
84 SymbolizableObjectFile::SymbolizableObjectFile(ObjectFile *Obj,
85                                                std::unique_ptr<DIContext> DICtx)
86     : Module(Obj), DebugInfoContext(std::move(DICtx)) {}
87 
88 namespace {
89 
90 struct OffsetNamePair {
91   uint32_t Offset;
92   StringRef Name;
93 
94   bool operator<(const OffsetNamePair &R) const {
95     return Offset < R.Offset;
96   }
97 };
98 
99 } // end anonymous namespace
100 
101 std::error_code SymbolizableObjectFile::addCoffExportSymbols(
102     const COFFObjectFile *CoffObj) {
103   // Get all export names and offsets.
104   std::vector<OffsetNamePair> ExportSyms;
105   for (const ExportDirectoryEntryRef &Ref : CoffObj->export_directories()) {
106     StringRef Name;
107     uint32_t Offset;
108     if (auto EC = Ref.getSymbolName(Name))
109       return EC;
110     if (auto EC = Ref.getExportRVA(Offset))
111       return EC;
112     ExportSyms.push_back(OffsetNamePair{Offset, Name});
113   }
114   if (ExportSyms.empty())
115     return std::error_code();
116 
117   // Sort by ascending offset.
118   array_pod_sort(ExportSyms.begin(), ExportSyms.end());
119 
120   // Approximate the symbol sizes by assuming they run to the next symbol.
121   // FIXME: This assumes all exports are functions.
122   uint64_t ImageBase = CoffObj->getImageBase();
123   for (auto I = ExportSyms.begin(), E = ExportSyms.end(); I != E; ++I) {
124     OffsetNamePair &Export = *I;
125     // FIXME: The last export has a one byte size now.
126     uint32_t NextOffset = I != E ? I->Offset : Export.Offset + 1;
127     uint64_t SymbolStart = ImageBase + Export.Offset;
128     uint64_t SymbolSize = NextOffset - Export.Offset;
129     SymbolDesc SD = {SymbolStart, SymbolSize};
130     Functions.insert(std::make_pair(SD, Export.Name));
131   }
132   return std::error_code();
133 }
134 
135 std::error_code SymbolizableObjectFile::addSymbol(const SymbolRef &Symbol,
136                                                   uint64_t SymbolSize,
137                                                   DataExtractor *OpdExtractor,
138                                                   uint64_t OpdAddress) {
139   Expected<SymbolRef::Type> SymbolTypeOrErr = Symbol.getType();
140   if (!SymbolTypeOrErr)
141     return errorToErrorCode(SymbolTypeOrErr.takeError());
142   SymbolRef::Type SymbolType = *SymbolTypeOrErr;
143   if (SymbolType != SymbolRef::ST_Function && SymbolType != SymbolRef::ST_Data)
144     return std::error_code();
145   Expected<uint64_t> SymbolAddressOrErr = Symbol.getAddress();
146   if (!SymbolAddressOrErr)
147     return errorToErrorCode(SymbolAddressOrErr.takeError());
148   uint64_t SymbolAddress = *SymbolAddressOrErr;
149   if (OpdExtractor) {
150     // For big-endian PowerPC64 ELF, symbols in the .opd section refer to
151     // function descriptors. The first word of the descriptor is a pointer to
152     // the function's code.
153     // For the purposes of symbolization, pretend the symbol's address is that
154     // of the function's code, not the descriptor.
155     uint64_t OpdOffset = SymbolAddress - OpdAddress;
156     uint32_t OpdOffset32 = OpdOffset;
157     if (OpdOffset == OpdOffset32 &&
158         OpdExtractor->isValidOffsetForAddress(OpdOffset32))
159       SymbolAddress = OpdExtractor->getAddress(&OpdOffset32);
160   }
161   Expected<StringRef> SymbolNameOrErr = Symbol.getName();
162   if (!SymbolNameOrErr)
163     return errorToErrorCode(SymbolNameOrErr.takeError());
164   StringRef SymbolName = *SymbolNameOrErr;
165   // Mach-O symbol table names have leading underscore, skip it.
166   if (Module->isMachO() && !SymbolName.empty() && SymbolName[0] == '_')
167     SymbolName = SymbolName.drop_front();
168   // FIXME: If a function has alias, there are two entries in symbol table
169   // with same address size. Make sure we choose the correct one.
170   auto &M = SymbolType == SymbolRef::ST_Function ? Functions : Objects;
171   SymbolDesc SD = { SymbolAddress, SymbolSize };
172   M.insert(std::make_pair(SD, SymbolName));
173   return std::error_code();
174 }
175 
176 // Return true if this is a 32-bit x86 PE COFF module.
177 bool SymbolizableObjectFile::isWin32Module() const {
178   auto *CoffObject = dyn_cast<COFFObjectFile>(Module);
179   return CoffObject && CoffObject->getMachine() == COFF::IMAGE_FILE_MACHINE_I386;
180 }
181 
182 uint64_t SymbolizableObjectFile::getModulePreferredBase() const {
183   if (auto *CoffObject = dyn_cast<COFFObjectFile>(Module))
184     return CoffObject->getImageBase();
185   return 0;
186 }
187 
188 bool SymbolizableObjectFile::getNameFromSymbolTable(SymbolRef::Type Type,
189                                                     uint64_t Address,
190                                                     std::string &Name,
191                                                     uint64_t &Addr,
192                                                     uint64_t &Size) const {
193   const auto &SymbolMap = Type == SymbolRef::ST_Function ? Functions : Objects;
194   if (SymbolMap.empty())
195     return false;
196   SymbolDesc SD = { Address, Address };
197   auto SymbolIterator = SymbolMap.upper_bound(SD);
198   if (SymbolIterator == SymbolMap.begin())
199     return false;
200   --SymbolIterator;
201   if (SymbolIterator->first.Size != 0 &&
202       SymbolIterator->first.Addr + SymbolIterator->first.Size <= Address)
203     return false;
204   Name = SymbolIterator->second.str();
205   Addr = SymbolIterator->first.Addr;
206   Size = SymbolIterator->first.Size;
207   return true;
208 }
209 
210 bool SymbolizableObjectFile::shouldOverrideWithSymbolTable(
211     FunctionNameKind FNKind, bool UseSymbolTable) const {
212   // When DWARF is used with -gline-tables-only / -gmlt, the symbol table gives
213   // better answers for linkage names than the DIContext. Otherwise, we are
214   // probably using PEs and PDBs, and we shouldn't do the override. PE files
215   // generally only contain the names of exported symbols.
216   return FNKind == FunctionNameKind::LinkageName && UseSymbolTable &&
217          isa<DWARFContext>(DebugInfoContext.get());
218 }
219 
220 DILineInfo SymbolizableObjectFile::symbolizeCode(uint64_t ModuleOffset,
221                                                  FunctionNameKind FNKind,
222                                                  bool UseSymbolTable) const {
223   DILineInfo LineInfo;
224   if (DebugInfoContext) {
225     LineInfo = DebugInfoContext->getLineInfoForAddress(
226         ModuleOffset, getDILineInfoSpecifier(FNKind));
227   }
228   // Override function name from symbol table if necessary.
229   if (shouldOverrideWithSymbolTable(FNKind, UseSymbolTable)) {
230     std::string FunctionName;
231     uint64_t Start, Size;
232     if (getNameFromSymbolTable(SymbolRef::ST_Function, ModuleOffset,
233                                FunctionName, Start, Size)) {
234       LineInfo.FunctionName = FunctionName;
235     }
236   }
237   return LineInfo;
238 }
239 
240 DIInliningInfo SymbolizableObjectFile::symbolizeInlinedCode(
241     uint64_t ModuleOffset, FunctionNameKind FNKind, bool UseSymbolTable) const {
242   DIInliningInfo InlinedContext;
243 
244   if (DebugInfoContext)
245     InlinedContext = DebugInfoContext->getInliningInfoForAddress(
246         ModuleOffset, getDILineInfoSpecifier(FNKind));
247   // Make sure there is at least one frame in context.
248   if (InlinedContext.getNumberOfFrames() == 0)
249     InlinedContext.addFrame(DILineInfo());
250 
251   // Override the function name in lower frame with name from symbol table.
252   if (shouldOverrideWithSymbolTable(FNKind, UseSymbolTable)) {
253     std::string FunctionName;
254     uint64_t Start, Size;
255     if (getNameFromSymbolTable(SymbolRef::ST_Function, ModuleOffset,
256                                FunctionName, Start, Size)) {
257       InlinedContext.getMutableFrame(InlinedContext.getNumberOfFrames() - 1)
258           ->FunctionName = FunctionName;
259     }
260   }
261 
262   return InlinedContext;
263 }
264 
265 DIGlobal SymbolizableObjectFile::symbolizeData(uint64_t ModuleOffset) const {
266   DIGlobal Res;
267   getNameFromSymbolTable(SymbolRef::ST_Data, ModuleOffset, Res.Name, Res.Start,
268                          Res.Size);
269   return Res;
270 }
271