dwarf_line_to_module.h
1 // -*- mode: c++ -*- 2 3 // Copyright 2010 Google LLC 4 // 5 // Redistribution and use in source and binary forms, with or without 6 // modification, are permitted provided that the following conditions are 7 // met: 8 // 9 // * Redistributions of source code must retain the above copyright 10 // notice, this list of conditions and the following disclaimer. 11 // * Redistributions in binary form must reproduce the above 12 // copyright notice, this list of conditions and the following disclaimer 13 // in the documentation and/or other materials provided with the 14 // distribution. 15 // * Neither the name of Google LLC nor the names of its 16 // contributors may be used to endorse or promote products derived from 17 // this software without specific prior written permission. 18 // 19 // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 20 // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 21 // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 22 // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 23 // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 24 // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 25 // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 26 // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 27 // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 28 // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 29 // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 30 31 // Original author: Jim Blandy <jimb@mozilla.com> <jimb@red-bean.com> 32 33 // The DwarfLineToModule class accepts line number information from a 34 // DWARF parser and adds it to a google_breakpad::Module. The Module 35 // can write that data out as a Breakpad symbol file. 36 37 #ifndef COMMON_LINUX_DWARF_LINE_TO_MODULE_H 38 #define COMMON_LINUX_DWARF_LINE_TO_MODULE_H 39 40 #include <string> 41 42 #include "common/module.h" 43 #include "common/dwarf/dwarf2reader.h" 44 #include "common/using_std_string.h" 45 46 namespace google_breakpad { 47 48 // A class for producing a vector of google_breakpad::Module::Line 49 // instances from parsed DWARF line number data. 50 // 51 // An instance of this class can be provided as a handler to a 52 // LineInfo DWARF line number information parser. The 53 // handler accepts source location information from the parser and 54 // uses it to produce a vector of google_breakpad::Module::Line 55 // objects, referring to google_breakpad::Module::File objects added 56 // to a particular google_breakpad::Module. 57 // 58 // GNU toolchain omitted sections support: 59 // ====================================== 60 // 61 // Given the right options, the GNU toolchain will omit unreferenced 62 // functions from the final executable. Unfortunately, when it does so, it 63 // does not remove the associated portions of the DWARF line number 64 // program; instead, it gives the DW_LNE_set_address instructions referring 65 // to the now-deleted code addresses of zero. Given this input, the DWARF 66 // line parser will call AddLine with a series of lines starting at address 67 // zero. For example, here is the output from 'readelf -wl' for a program 68 // with four functions, the first three of which have been omitted: 69 // 70 // Line Number Statements: 71 // Extended opcode 2: set Address to 0x0 72 // Advance Line by 14 to 15 73 // Copy 74 // Special opcode 48: advance Address by 3 to 0x3 and Line by 1 to 16 75 // Special opcode 119: advance Address by 8 to 0xb and Line by 2 to 18 76 // Advance PC by 2 to 0xd 77 // Extended opcode 1: End of Sequence 78 // 79 // Extended opcode 2: set Address to 0x0 80 // Advance Line by 14 to 15 81 // Copy 82 // Special opcode 48: advance Address by 3 to 0x3 and Line by 1 to 16 83 // Special opcode 119: advance Address by 8 to 0xb and Line by 2 to 18 84 // Advance PC by 2 to 0xd 85 // Extended opcode 1: End of Sequence 86 // 87 // Extended opcode 2: set Address to 0x0 88 // Advance Line by 19 to 20 89 // Copy 90 // Special opcode 48: advance Address by 3 to 0x3 and Line by 1 to 21 91 // Special opcode 76: advance Address by 5 to 0x8 and Line by 1 to 22 92 // Advance PC by 2 to 0xa 93 // Extended opcode 1: End of Sequence 94 // 95 // Extended opcode 2: set Address to 0x80483a4 96 // Advance Line by 23 to 24 97 // Copy 98 // Special opcode 202: advance Address by 14 to 0x80483b2 and Line by 1 to 25 99 // Special opcode 76: advance Address by 5 to 0x80483b7 and Line by 1 to 26 100 // Advance PC by 6 to 0x80483bd 101 // Extended opcode 1: End of Sequence 102 // 103 // Instead of collecting runs of lines describing code that is not there, 104 // we try to recognize and drop them. Since the linker doesn't explicitly 105 // distinguish references to dropped sections from genuine references to 106 // code at address zero, we must use a heuristic. We have chosen: 107 // 108 // - If a line starts at address zero, omit it. (On the platforms 109 // breakpad targets, it is extremely unlikely that there will be code 110 // at address zero.) 111 // 112 // - If a line starts immediately after an omitted line, omit it too. 113 class DwarfLineToModule: public LineInfoHandler { 114 public: 115 // As the DWARF line info parser passes us line records, add source 116 // files to MODULE, and add all lines to the end of LINES. LINES 117 // need not be empty. If the parser hands us a zero-length line, we 118 // omit it. If the parser hands us a line that extends beyond the 119 // end of the address space, we clip it. It's up to our client to 120 // sort out which lines belong to which functions; we don't add them 121 // to any particular function in MODULE ourselves. 122 DwarfLineToModule(Module* module, 123 const string& compilation_dir, 124 vector<Module::Line>* lines, 125 std::map<uint32_t, Module::File*>* files) 126 : module_(module), 127 compilation_dir_(compilation_dir), 128 lines_(lines), 129 files_(files), 130 highest_file_number_(-1), 131 omitted_line_end_(0), 132 warned_bad_file_number_(false), 133 warned_bad_directory_number_(false) { } 134 135 ~DwarfLineToModule() { } 136 137 void DefineDir(const string& name, uint32_t dir_num); 138 void DefineFile(const string& name, int32_t file_num, 139 uint32_t dir_num, uint64_t mod_time, 140 uint64_t length); 141 void AddLine(uint64_t address, uint64_t length, 142 uint32_t file_num, uint32_t line_num, uint32_t column_num); 143 144 private: 145 146 typedef std::map<uint32_t, string> DirectoryTable; 147 typedef std::map<uint32_t, Module::File*> FileTable; 148 149 // The module we're contributing debugging info to. Owned by our 150 // client. 151 Module *module_; 152 153 // The compilation directory for the current compilation unit whose 154 // lines are being accumulated. 155 string compilation_dir_; 156 157 // The vector of lines we're accumulating. Owned by our client. 158 // 159 // In a Module, as in a breakpad symbol file, lines belong to 160 // specific functions, but DWARF simply assigns lines to addresses; 161 // one must infer the line/function relationship using the 162 // functions' beginning and ending addresses. So we can't add these 163 // to the appropriate function from module_ until we've read the 164 // function info as well. Instead, we accumulate lines here, and let 165 // whoever constructed this sort it all out. 166 vector<Module::Line>* lines_; 167 168 // A table mapping directory numbers to paths. 169 DirectoryTable directories_; 170 171 // A table mapping file numbers to Module::File pointers. 172 FileTable* files_; 173 174 // The highest file number we've seen so far, or -1 if we've seen 175 // none. Used for dynamically defined file numbers. 176 int32_t highest_file_number_; 177 178 // This is the ending address of the last line we omitted, or zero if we 179 // didn't omit the previous line. It is zero before we have received any 180 // AddLine calls. 181 uint64_t omitted_line_end_; 182 183 // True if we've warned about: 184 bool warned_bad_file_number_; // bad file numbers 185 bool warned_bad_directory_number_; // bad directory numbers 186 }; 187 188 } // namespace google_breakpad 189 190 #endif // COMMON_LINUX_DWARF_LINE_TO_MODULE_H