1 //===- FunctionInfo.h -------------------------------------------*- 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 #ifndef LLVM_DEBUGINFO_GSYM_FUNCTIONINFO_H 10 #define LLVM_DEBUGINFO_GSYM_FUNCTIONINFO_H 11 12 #include "llvm/ADT/Optional.h" 13 #include "llvm/DebugInfo/GSYM/InlineInfo.h" 14 #include "llvm/DebugInfo/GSYM/LineTable.h" 15 #include "llvm/DebugInfo/GSYM/LookupResult.h" 16 #include "llvm/DebugInfo/GSYM/Range.h" 17 #include "llvm/DebugInfo/GSYM/StringTable.h" 18 #include <cstdint> 19 #include <tuple> 20 21 namespace llvm { 22 class raw_ostream; 23 24 namespace gsym { 25 26 class GsymReader; 27 /// Function information in GSYM files encodes information for one contiguous 28 /// address range. If a function has discontiguous address ranges, they will 29 /// need to be encoded using multiple FunctionInfo objects. 30 /// 31 /// ENCODING 32 /// 33 /// The function information gets the function start address as an argument 34 /// to the FunctionInfo::decode(...) function. This information is calculated 35 /// from the GSYM header and an address offset from the GSYM address offsets 36 /// table. The encoded FunctionInfo information must be aligned to a 4 byte 37 /// boundary. 38 /// 39 /// The encoded data for a FunctionInfo starts with fixed data that all 40 /// function info objects have: 41 /// 42 /// ENCODING NAME DESCRIPTION 43 /// ========= =========== ==================================================== 44 /// uint32_t Size The size in bytes of this function. 45 /// uint32_t Name The string table offset of the function name. 46 /// 47 /// The optional data in a FunctionInfo object follows this fixed information 48 /// and consists of a stream of tuples that consist of: 49 /// 50 /// ENCODING NAME DESCRIPTION 51 /// ========= =========== ==================================================== 52 /// uint32_t InfoType An "InfoType" enumeration that describes the type 53 /// of optional data that is encoded. 54 /// uint32_t InfoLength The size in bytes of the encoded data that 55 /// immediately follows this length if this value is 56 /// greater than zero. 57 /// uint8_t[] InfoData Encoded bytes that represent the data for the 58 /// "InfoType". These bytes are only present if 59 /// "InfoLength" is greater than zero. 60 /// 61 /// The "InfoType" is an enumeration: 62 /// 63 /// enum InfoType { 64 /// EndOfList = 0u, 65 /// LineTableInfo = 1u, 66 /// InlineInfo = 2u 67 /// }; 68 /// 69 /// This stream of tuples is terminated by a "InfoType" whose value is 70 /// InfoType::EndOfList and a zero for "InfoLength". This signifies the end of 71 /// the optional information list. This format allows us to add new optional 72 /// information data to a FunctionInfo object over time and allows older 73 /// clients to still parse the format and skip over any data that they don't 74 /// understand or want to parse. 75 /// 76 /// So the function information encoding essientially looks like: 77 /// 78 /// struct { 79 /// uint32_t Size; 80 /// uint32_t Name; 81 /// struct { 82 /// uint32_t InfoType; 83 /// uint32_t InfoLength; 84 /// uint8_t InfoData[InfoLength]; 85 /// }[N]; 86 /// } 87 /// 88 /// Where "N" is the number of tuples. 89 struct FunctionInfo { 90 AddressRange Range; 91 uint32_t Name; ///< String table offset in the string table. 92 llvm::Optional<LineTable> OptLineTable; 93 llvm::Optional<InlineInfo> Inline; 94 95 FunctionInfo(uint64_t Addr = 0, uint64_t Size = 0, uint32_t N = 0) 96 : Range(Addr, Addr + Size), Name(N) {} 97 98 /// Query if a FunctionInfo has rich debug info. 99 /// 100 /// \returns A bool that indicates if this object has something else than 101 /// range and name. When converting information from a symbol table and from 102 /// debug info, we might end up with multiple FunctionInfo objects for the 103 /// same range and we need to be able to tell which one is the better object 104 /// to use. hasRichInfoFunctionInfo105 bool hasRichInfo() const { 106 return OptLineTable.hasValue() || Inline.hasValue(); 107 } 108 109 /// Query if a FunctionInfo object is valid. 110 /// 111 /// Address and size can be zero and there can be no line entries for a 112 /// symbol so the only indication this entry is valid is if the name is 113 /// not zero. This can happen when extracting information from symbol 114 /// tables that do not encode symbol sizes. In that case only the 115 /// address and name will be filled in. 116 /// 117 /// \returns A boolean indicating if this FunctionInfo is valid. isValidFunctionInfo118 bool isValid() const { 119 return Name != 0; 120 } 121 122 /// Decode an object from a binary data stream. 123 /// 124 /// \param Data The binary stream to read the data from. This object must 125 /// have the data for the object starting at offset zero. The data 126 /// can contain more data than needed. 127 /// 128 /// \param BaseAddr The FunctionInfo's start address and will be used as the 129 /// base address when decoding any contained information like the line table 130 /// and the inline info. 131 /// 132 /// \returns An FunctionInfo or an error describing the issue that was 133 /// encountered during decoding. 134 static llvm::Expected<FunctionInfo> decode(DataExtractor &Data, 135 uint64_t BaseAddr); 136 137 /// Encode this object into FileWriter stream. 138 /// 139 /// \param O The binary stream to write the data to at the current file 140 /// position. 141 /// 142 /// \returns An error object that indicates failure or the offset of the 143 /// function info that was successfully written into the stream. 144 llvm::Expected<uint64_t> encode(FileWriter &O) const; 145 146 147 /// Lookup an address within a FunctionInfo object's data stream. 148 /// 149 /// Instead of decoding an entire FunctionInfo object when doing lookups, 150 /// we can decode only the information we need from the FunctionInfo's data 151 /// for the specific address. The lookup result information is returned as 152 /// a LookupResult. 153 /// 154 /// \param Data The binary stream to read the data from. This object must 155 /// have the data for the object starting at offset zero. The data 156 /// can contain more data than needed. 157 /// 158 /// \param GR The GSYM reader that contains the string and file table that 159 /// will be used to fill in information in the returned result. 160 /// 161 /// \param FuncAddr The function start address decoded from the GsymReader. 162 /// 163 /// \param Addr The address to lookup. 164 /// 165 /// \returns An LookupResult or an error describing the issue that was 166 /// encountered during decoding. An error should only be returned if the 167 /// address is not contained in the FunctionInfo or if the data is corrupted. 168 static llvm::Expected<LookupResult> lookup(DataExtractor &Data, 169 const GsymReader &GR, 170 uint64_t FuncAddr, 171 uint64_t Addr); 172 startAddressFunctionInfo173 uint64_t startAddress() const { return Range.Start; } endAddressFunctionInfo174 uint64_t endAddress() const { return Range.End; } sizeFunctionInfo175 uint64_t size() const { return Range.size(); } setStartAddressFunctionInfo176 void setStartAddress(uint64_t Addr) { Range.Start = Addr; } setEndAddressFunctionInfo177 void setEndAddress(uint64_t Addr) { Range.End = Addr; } setSizeFunctionInfo178 void setSize(uint64_t Size) { Range.End = Range.Start + Size; } 179 clearFunctionInfo180 void clear() { 181 Range = {0, 0}; 182 Name = 0; 183 OptLineTable = None; 184 Inline = None; 185 } 186 }; 187 188 inline bool operator==(const FunctionInfo &LHS, const FunctionInfo &RHS) { 189 return LHS.Range == RHS.Range && LHS.Name == RHS.Name && 190 LHS.OptLineTable == RHS.OptLineTable && LHS.Inline == RHS.Inline; 191 } 192 inline bool operator!=(const FunctionInfo &LHS, const FunctionInfo &RHS) { 193 return !(LHS == RHS); 194 } 195 /// This sorting will order things consistently by address range first, but then 196 /// followed by inlining being valid and line tables. We might end up with a 197 /// FunctionInfo from debug info that will have the same range as one from the 198 /// symbol table, but we want to quickly be able to sort and use the best version 199 /// when creating the final GSYM file. 200 inline bool operator<(const FunctionInfo &LHS, const FunctionInfo &RHS) { 201 // First sort by address range 202 if (LHS.Range != RHS.Range) 203 return LHS.Range < RHS.Range; 204 205 // Then sort by inline 206 if (LHS.Inline.hasValue() != RHS.Inline.hasValue()) 207 return RHS.Inline.hasValue(); 208 209 return LHS.OptLineTable < RHS.OptLineTable; 210 } 211 212 raw_ostream &operator<<(raw_ostream &OS, const FunctionInfo &R); 213 214 } // namespace gsym 215 } // namespace llvm 216 217 #endif // LLVM_DEBUGINFO_GSYM_FUNCTIONINFO_H 218