Details | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
14 | pmbaty | 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/DebugInfo/GSYM/ExtractRanges.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/StringTable.h" |
||
17 | #include <cstdint> |
||
18 | #include <tuple> |
||
19 | |||
20 | namespace llvm { |
||
21 | class raw_ostream; |
||
22 | |||
23 | namespace gsym { |
||
24 | |||
25 | class GsymReader; |
||
26 | /// Function information in GSYM files encodes information for one contiguous |
||
27 | /// address range. If a function has discontiguous address ranges, they will |
||
28 | /// need to be encoded using multiple FunctionInfo objects. |
||
29 | /// |
||
30 | /// ENCODING |
||
31 | /// |
||
32 | /// The function information gets the function start address as an argument |
||
33 | /// to the FunctionInfo::decode(...) function. This information is calculated |
||
34 | /// from the GSYM header and an address offset from the GSYM address offsets |
||
35 | /// table. The encoded FunctionInfo information must be aligned to a 4 byte |
||
36 | /// boundary. |
||
37 | /// |
||
38 | /// The encoded data for a FunctionInfo starts with fixed data that all |
||
39 | /// function info objects have: |
||
40 | /// |
||
41 | /// ENCODING NAME DESCRIPTION |
||
42 | /// ========= =========== ==================================================== |
||
43 | /// uint32_t Size The size in bytes of this function. |
||
44 | /// uint32_t Name The string table offset of the function name. |
||
45 | /// |
||
46 | /// The optional data in a FunctionInfo object follows this fixed information |
||
47 | /// and consists of a stream of tuples that consist of: |
||
48 | /// |
||
49 | /// ENCODING NAME DESCRIPTION |
||
50 | /// ========= =========== ==================================================== |
||
51 | /// uint32_t InfoType An "InfoType" enumeration that describes the type |
||
52 | /// of optional data that is encoded. |
||
53 | /// uint32_t InfoLength The size in bytes of the encoded data that |
||
54 | /// immediately follows this length if this value is |
||
55 | /// greater than zero. |
||
56 | /// uint8_t[] InfoData Encoded bytes that represent the data for the |
||
57 | /// "InfoType". These bytes are only present if |
||
58 | /// "InfoLength" is greater than zero. |
||
59 | /// |
||
60 | /// The "InfoType" is an enumeration: |
||
61 | /// |
||
62 | /// enum InfoType { |
||
63 | /// EndOfList = 0u, |
||
64 | /// LineTableInfo = 1u, |
||
65 | /// InlineInfo = 2u |
||
66 | /// }; |
||
67 | /// |
||
68 | /// This stream of tuples is terminated by a "InfoType" whose value is |
||
69 | /// InfoType::EndOfList and a zero for "InfoLength". This signifies the end of |
||
70 | /// the optional information list. This format allows us to add new optional |
||
71 | /// information data to a FunctionInfo object over time and allows older |
||
72 | /// clients to still parse the format and skip over any data that they don't |
||
73 | /// understand or want to parse. |
||
74 | /// |
||
75 | /// So the function information encoding essientially looks like: |
||
76 | /// |
||
77 | /// struct { |
||
78 | /// uint32_t Size; |
||
79 | /// uint32_t Name; |
||
80 | /// struct { |
||
81 | /// uint32_t InfoType; |
||
82 | /// uint32_t InfoLength; |
||
83 | /// uint8_t InfoData[InfoLength]; |
||
84 | /// }[N]; |
||
85 | /// } |
||
86 | /// |
||
87 | /// Where "N" is the number of tuples. |
||
88 | struct FunctionInfo { |
||
89 | AddressRange Range; |
||
90 | uint32_t Name; ///< String table offset in the string table. |
||
91 | std::optional<LineTable> OptLineTable; |
||
92 | std::optional<InlineInfo> Inline; |
||
93 | |||
94 | FunctionInfo(uint64_t Addr = 0, uint64_t Size = 0, uint32_t N = 0) |
||
95 | : Range(Addr, Addr + Size), Name(N) {} |
||
96 | |||
97 | /// Query if a FunctionInfo has rich debug info. |
||
98 | /// |
||
99 | /// \returns A bool that indicates if this object has something else than |
||
100 | /// range and name. When converting information from a symbol table and from |
||
101 | /// debug info, we might end up with multiple FunctionInfo objects for the |
||
102 | /// same range and we need to be able to tell which one is the better object |
||
103 | /// to use. |
||
104 | bool hasRichInfo() const { return OptLineTable || Inline; } |
||
105 | |||
106 | /// Query if a FunctionInfo object is valid. |
||
107 | /// |
||
108 | /// Address and size can be zero and there can be no line entries for a |
||
109 | /// symbol so the only indication this entry is valid is if the name is |
||
110 | /// not zero. This can happen when extracting information from symbol |
||
111 | /// tables that do not encode symbol sizes. In that case only the |
||
112 | /// address and name will be filled in. |
||
113 | /// |
||
114 | /// \returns A boolean indicating if this FunctionInfo is valid. |
||
115 | bool isValid() const { |
||
116 | return Name != 0; |
||
117 | } |
||
118 | |||
119 | /// Decode an object from a binary data stream. |
||
120 | /// |
||
121 | /// \param Data The binary stream to read the data from. This object must |
||
122 | /// have the data for the object starting at offset zero. The data |
||
123 | /// can contain more data than needed. |
||
124 | /// |
||
125 | /// \param BaseAddr The FunctionInfo's start address and will be used as the |
||
126 | /// base address when decoding any contained information like the line table |
||
127 | /// and the inline info. |
||
128 | /// |
||
129 | /// \returns An FunctionInfo or an error describing the issue that was |
||
130 | /// encountered during decoding. |
||
131 | static llvm::Expected<FunctionInfo> decode(DataExtractor &Data, |
||
132 | uint64_t BaseAddr); |
||
133 | |||
134 | /// Encode this object into FileWriter stream. |
||
135 | /// |
||
136 | /// \param O The binary stream to write the data to at the current file |
||
137 | /// position. |
||
138 | /// |
||
139 | /// \returns An error object that indicates failure or the offset of the |
||
140 | /// function info that was successfully written into the stream. |
||
141 | llvm::Expected<uint64_t> encode(FileWriter &O) const; |
||
142 | |||
143 | |||
144 | /// Lookup an address within a FunctionInfo object's data stream. |
||
145 | /// |
||
146 | /// Instead of decoding an entire FunctionInfo object when doing lookups, |
||
147 | /// we can decode only the information we need from the FunctionInfo's data |
||
148 | /// for the specific address. The lookup result information is returned as |
||
149 | /// a LookupResult. |
||
150 | /// |
||
151 | /// \param Data The binary stream to read the data from. This object must |
||
152 | /// have the data for the object starting at offset zero. The data |
||
153 | /// can contain more data than needed. |
||
154 | /// |
||
155 | /// \param GR The GSYM reader that contains the string and file table that |
||
156 | /// will be used to fill in information in the returned result. |
||
157 | /// |
||
158 | /// \param FuncAddr The function start address decoded from the GsymReader. |
||
159 | /// |
||
160 | /// \param Addr The address to lookup. |
||
161 | /// |
||
162 | /// \returns An LookupResult or an error describing the issue that was |
||
163 | /// encountered during decoding. An error should only be returned if the |
||
164 | /// address is not contained in the FunctionInfo or if the data is corrupted. |
||
165 | static llvm::Expected<LookupResult> lookup(DataExtractor &Data, |
||
166 | const GsymReader &GR, |
||
167 | uint64_t FuncAddr, |
||
168 | uint64_t Addr); |
||
169 | |||
170 | uint64_t startAddress() const { return Range.start(); } |
||
171 | uint64_t endAddress() const { return Range.end(); } |
||
172 | uint64_t size() const { return Range.size(); } |
||
173 | |||
174 | void clear() { |
||
175 | Range = {0, 0}; |
||
176 | Name = 0; |
||
177 | OptLineTable = std::nullopt; |
||
178 | Inline = std::nullopt; |
||
179 | } |
||
180 | }; |
||
181 | |||
182 | inline bool operator==(const FunctionInfo &LHS, const FunctionInfo &RHS) { |
||
183 | return LHS.Range == RHS.Range && LHS.Name == RHS.Name && |
||
184 | LHS.OptLineTable == RHS.OptLineTable && LHS.Inline == RHS.Inline; |
||
185 | } |
||
186 | inline bool operator!=(const FunctionInfo &LHS, const FunctionInfo &RHS) { |
||
187 | return !(LHS == RHS); |
||
188 | } |
||
189 | /// This sorting will order things consistently by address range first, but then |
||
190 | /// followed by inlining being valid and line tables. We might end up with a |
||
191 | /// FunctionInfo from debug info that will have the same range as one from the |
||
192 | /// symbol table, but we want to quickly be able to sort and use the best version |
||
193 | /// when creating the final GSYM file. |
||
194 | inline bool operator<(const FunctionInfo &LHS, const FunctionInfo &RHS) { |
||
195 | // First sort by address range |
||
196 | if (LHS.Range != RHS.Range) |
||
197 | return LHS.Range < RHS.Range; |
||
198 | |||
199 | // Then sort by inline |
||
200 | if (LHS.Inline.has_value() != RHS.Inline.has_value()) |
||
201 | return RHS.Inline.has_value(); |
||
202 | |||
203 | return LHS.OptLineTable < RHS.OptLineTable; |
||
204 | } |
||
205 | |||
206 | raw_ostream &operator<<(raw_ostream &OS, const FunctionInfo &R); |
||
207 | |||
208 | } // namespace gsym |
||
209 | } // namespace llvm |
||
210 | |||
211 | #endif // LLVM_DEBUGINFO_GSYM_FUNCTIONINFO_H |