1 //===- ObjectFileTransformer.cpp --------------------------------*- C++ -*-===//
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 <unordered_set>
10 
11 #include "llvm/Object/ELFObjectFile.h"
12 #include "llvm/Object/MachOUniversal.h"
13 #include "llvm/Object/ObjectFile.h"
14 #include "llvm/Support/DataExtractor.h"
15 #include "llvm/Support/raw_ostream.h"
16 
17 #include "llvm/DebugInfo/GSYM/ObjectFileTransformer.h"
18 #include "llvm/DebugInfo/GSYM/GsymCreator.h"
19 
20 using namespace llvm;
21 using namespace gsym;
22 
23 constexpr uint32_t NT_GNU_BUILD_ID_TAG = 0x03;
24 
25 static std::vector<uint8_t> getUUID(const object::ObjectFile &Obj) {
26   // Extract the UUID from the object file
27   std::vector<uint8_t> UUID;
28   if (auto *MachO = dyn_cast<object::MachOObjectFile>(&Obj)) {
29     const ArrayRef<uint8_t> MachUUID = MachO->getUuid();
30     if (!MachUUID.empty())
31       UUID.assign(MachUUID.data(), MachUUID.data() + MachUUID.size());
32   } else if (isa<object::ELFObjectFileBase>(&Obj)) {
33     const StringRef GNUBuildID(".note.gnu.build-id");
34     for (const object::SectionRef &Sect : Obj.sections()) {
35       Expected<StringRef> SectNameOrErr = Sect.getName();
36       if (!SectNameOrErr) {
37         consumeError(SectNameOrErr.takeError());
38         continue;
39       }
40       StringRef SectName(*SectNameOrErr);
41       if (SectName != GNUBuildID)
42         continue;
43       StringRef BuildIDData;
44       Expected<StringRef> E = Sect.getContents();
45       if (E)
46         BuildIDData = *E;
47       else {
48         consumeError(E.takeError());
49         continue;
50       }
51       DataExtractor Decoder(BuildIDData, Obj.makeTriple().isLittleEndian(), 8);
52       uint64_t Offset = 0;
53       const uint32_t NameSize = Decoder.getU32(&Offset);
54       const uint32_t PayloadSize = Decoder.getU32(&Offset);
55       const uint32_t PayloadType = Decoder.getU32(&Offset);
56       StringRef Name(Decoder.getFixedLengthString(&Offset, NameSize));
57       if (Name == "GNU" && PayloadType == NT_GNU_BUILD_ID_TAG) {
58         Offset = alignTo(Offset, 4);
59         StringRef UUIDBytes(Decoder.getBytes(&Offset, PayloadSize));
60         if (!UUIDBytes.empty()) {
61           auto Ptr = reinterpret_cast<const uint8_t *>(UUIDBytes.data());
62           UUID.assign(Ptr, Ptr + UUIDBytes.size());
63         }
64       }
65     }
66   }
67   return UUID;
68 }
69 
70 llvm::Error ObjectFileTransformer::convert(const object::ObjectFile &Obj,
71                                            raw_ostream &Log,
72                                            GsymCreator &Gsym) {
73   using namespace llvm::object;
74 
75   const bool IsMachO = isa<MachOObjectFile>(&Obj);
76   const bool IsELF = isa<ELFObjectFileBase>(&Obj);
77 
78   // Read build ID.
79   Gsym.setUUID(getUUID(Obj));
80 
81   // Parse the symbol table.
82   size_t NumBefore = Gsym.getNumFunctionInfos();
83   for (const object::SymbolRef &Sym : Obj.symbols()) {
84     Expected<SymbolRef::Type> SymType = Sym.getType();
85     if (!SymType) {
86       consumeError(SymType.takeError());
87       continue;
88     }
89     Expected<uint64_t> AddrOrErr = Sym.getValue();
90     if (!AddrOrErr)
91       // TODO: Test this error.
92       return AddrOrErr.takeError();
93 
94     if (SymType.get() != SymbolRef::Type::ST_Function ||
95         !Gsym.IsValidTextAddress(*AddrOrErr) ||
96         Gsym.hasFunctionInfoForAddress(*AddrOrErr))
97       continue;
98     // Function size for MachO files will be 0
99     constexpr bool NoCopy = false;
100     const uint64_t size = IsELF ? ELFSymbolRef(Sym).getSize() : 0;
101     Expected<StringRef> Name = Sym.getName();
102     if (!Name) {
103       logAllUnhandledErrors(Name.takeError(), Log, "ObjectFileTransformer: ");
104       continue;
105     }
106     // Remove the leading '_' character in any symbol names if there is one
107     // for mach-o files.
108     if (IsMachO)
109       Name->consume_front("_");
110     Gsym.addFunctionInfo(
111         FunctionInfo(*AddrOrErr, size, Gsym.insertString(*Name, NoCopy)));
112   }
113   size_t FunctionsAddedCount = Gsym.getNumFunctionInfos() - NumBefore;
114   Log << "Loaded " << FunctionsAddedCount << " functions from symbol table.\n";
115   return Error::success();
116 }
117