| 1 | //===- PDBStringTable.cpp - PDB String Table ---------------------*- 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 "llvm/DebugInfo/PDB/Native/PDBStringTable.h" |
| 10 | |
| 11 | #include "llvm/DebugInfo/PDB/Native/Hash.h" |
| 12 | #include "llvm/DebugInfo/PDB/Native/RawError.h" |
| 13 | #include "llvm/DebugInfo/PDB/Native/RawTypes.h" |
| 14 | #include "llvm/Support/BinaryStreamReader.h" |
| 15 | #include "llvm/Support/Endian.h" |
| 16 | |
| 17 | using namespace llvm; |
| 18 | using namespace llvm::support; |
| 19 | using namespace llvm::pdb; |
| 20 | |
| 21 | uint32_t PDBStringTable::getByteSize() const { return Header->ByteSize; } |
| 22 | uint32_t PDBStringTable::getNameCount() const { return NameCount; } |
| 23 | uint32_t PDBStringTable::getHashVersion() const { return Header->HashVersion; } |
| 24 | uint32_t PDBStringTable::getSignature() const { return Header->Signature; } |
| 25 | |
| 26 | Error PDBStringTable::(BinaryStreamReader &Reader) { |
| 27 | if (auto EC = Reader.readObject(Dest&: Header)) |
| 28 | return EC; |
| 29 | |
| 30 | if (Header->Signature != PDBStringTableSignature) |
| 31 | return make_error<RawError>(Args: raw_error_code::corrupt_file, |
| 32 | Args: "Invalid hash table signature" ); |
| 33 | if (Header->HashVersion != 1 && Header->HashVersion != 2) |
| 34 | return make_error<RawError>(Args: raw_error_code::corrupt_file, |
| 35 | Args: "Unsupported hash version" ); |
| 36 | |
| 37 | assert(Reader.bytesRemaining() == 0); |
| 38 | return Error::success(); |
| 39 | } |
| 40 | |
| 41 | Error PDBStringTable::readStrings(BinaryStreamReader &Reader) { |
| 42 | BinaryStreamRef Stream; |
| 43 | if (auto EC = Reader.readStreamRef(Ref&: Stream)) |
| 44 | return EC; |
| 45 | |
| 46 | if (auto EC = Strings.initialize(Contents: Stream)) { |
| 47 | return joinErrors(E1: std::move(EC), |
| 48 | E2: make_error<RawError>(Args: raw_error_code::corrupt_file, |
| 49 | Args: "Invalid hash table byte length" )); |
| 50 | } |
| 51 | |
| 52 | assert(Reader.bytesRemaining() == 0); |
| 53 | return Error::success(); |
| 54 | } |
| 55 | |
| 56 | const codeview::DebugStringTableSubsectionRef & |
| 57 | PDBStringTable::getStringTable() const { |
| 58 | return Strings; |
| 59 | } |
| 60 | |
| 61 | Error PDBStringTable::readHashTable(BinaryStreamReader &Reader) { |
| 62 | const support::ulittle32_t *HashCount; |
| 63 | if (auto EC = Reader.readObject(Dest&: HashCount)) |
| 64 | return EC; |
| 65 | |
| 66 | if (auto EC = Reader.readArray(Array&: IDs, NumItems: *HashCount)) { |
| 67 | return joinErrors(E1: std::move(EC), |
| 68 | E2: make_error<RawError>(Args: raw_error_code::corrupt_file, |
| 69 | Args: "Could not read bucket array" )); |
| 70 | } |
| 71 | |
| 72 | return Error::success(); |
| 73 | } |
| 74 | |
| 75 | Error PDBStringTable::readEpilogue(BinaryStreamReader &Reader) { |
| 76 | if (auto EC = Reader.readInteger(Dest&: NameCount)) |
| 77 | return EC; |
| 78 | |
| 79 | assert(Reader.bytesRemaining() == 0); |
| 80 | return Error::success(); |
| 81 | } |
| 82 | |
| 83 | Error PDBStringTable::reload(BinaryStreamReader &Reader) { |
| 84 | |
| 85 | BinaryStreamReader SectionReader; |
| 86 | |
| 87 | std::tie(args&: SectionReader, args&: Reader) = Reader.split(Offset: sizeof(PDBStringTableHeader)); |
| 88 | if (auto EC = readHeader(Reader&: SectionReader)) |
| 89 | return EC; |
| 90 | |
| 91 | std::tie(args&: SectionReader, args&: Reader) = Reader.split(Offset: Header->ByteSize); |
| 92 | if (auto EC = readStrings(Reader&: SectionReader)) |
| 93 | return EC; |
| 94 | |
| 95 | // We don't know how long the hash table is until we parse it, so let the |
| 96 | // function responsible for doing that figure it out. |
| 97 | if (auto EC = readHashTable(Reader)) |
| 98 | return EC; |
| 99 | |
| 100 | std::tie(args&: SectionReader, args&: Reader) = Reader.split(Offset: sizeof(uint32_t)); |
| 101 | if (auto EC = readEpilogue(Reader&: SectionReader)) |
| 102 | return EC; |
| 103 | |
| 104 | assert(Reader.bytesRemaining() == 0); |
| 105 | return Error::success(); |
| 106 | } |
| 107 | |
| 108 | Expected<StringRef> PDBStringTable::getStringForID(uint32_t ID) const { |
| 109 | return Strings.getString(Offset: ID); |
| 110 | } |
| 111 | |
| 112 | Expected<uint32_t> PDBStringTable::getIDForString(StringRef Str) const { |
| 113 | uint32_t Hash = |
| 114 | (Header->HashVersion == 1) ? hashStringV1(Str) : hashStringV2(Str); |
| 115 | size_t Count = IDs.size(); |
| 116 | uint32_t Start = Hash % Count; |
| 117 | for (size_t I = 0; I < Count; ++I) { |
| 118 | // The hash is just a starting point for the search, but if it |
| 119 | // doesn't work we should find the string no matter what, because |
| 120 | // we iterate the entire array. |
| 121 | uint32_t Index = (Start + I) % Count; |
| 122 | |
| 123 | // If we find 0, it means the item isn't in the hash table. |
| 124 | uint32_t ID = IDs[Index]; |
| 125 | if (ID == 0) |
| 126 | return make_error<RawError>(Args: raw_error_code::no_entry); |
| 127 | auto ExpectedStr = getStringForID(ID); |
| 128 | if (!ExpectedStr) |
| 129 | return ExpectedStr.takeError(); |
| 130 | |
| 131 | if (*ExpectedStr == Str) |
| 132 | return ID; |
| 133 | } |
| 134 | return make_error<RawError>(Args: raw_error_code::no_entry); |
| 135 | } |
| 136 | |
| 137 | FixedStreamArray<support::ulittle32_t> PDBStringTable::name_ids() const { |
| 138 | return IDs; |
| 139 | } |
| 140 | |