/ src / common / dwarf_cfi_to_module.h
dwarf_cfi_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  // dwarf_cfi_to_module.h: Define the DwarfCFIToModule class, which
 34  // accepts parsed DWARF call frame info and adds it to a
 35  // google_breakpad::Module object, which can write that information to
 36  // a Breakpad symbol file.
 37  
 38  #ifndef COMMON_LINUX_DWARF_CFI_TO_MODULE_H
 39  #define COMMON_LINUX_DWARF_CFI_TO_MODULE_H
 40  
 41  #include <assert.h>
 42  #include <stdio.h>
 43  
 44  #include <set>
 45  #include <string>
 46  #include <memory>
 47  #include <vector>
 48  
 49  #include "common/module.h"
 50  #include "common/dwarf/dwarf2reader.h"
 51  #include "common/using_std_string.h"
 52  
 53  namespace google_breakpad {
 54  
 55  using google_breakpad::Module;
 56  using std::set;
 57  using std::vector;
 58  
 59  // A class that accepts parsed call frame information from the DWARF
 60  // CFI parser and populates a google_breakpad::Module object with the
 61  // contents.
 62  class DwarfCFIToModule: public CallFrameInfo::Handler {
 63   public:
 64  
 65    // DwarfCFIToModule uses an instance of this class to report errors
 66    // detected while converting DWARF CFI to Breakpad STACK CFI records.
 67    class Reporter {
 68     public:
 69      // Create a reporter that writes messages to the standard error
 70      // stream. FILE is the name of the file we're processing, and
 71      // SECTION is the name of the section within that file that we're
 72      // looking at (.debug_frame, .eh_frame, etc.).
 73      Reporter(const string& file, const string& section)
 74        : file_(file), section_(section) { }
 75      virtual ~Reporter() { }
 76  
 77      // The DWARF CFI entry at OFFSET cites register REG, but REG is not
 78      // covered by the vector of register names passed to the
 79      // DwarfCFIToModule constructor, nor does it match the return
 80      // address column number for this entry.
 81      virtual void UnnamedRegister(size_t offset, int reg);
 82  
 83      // The DWARF CFI entry at OFFSET says that REG is undefined, but the
 84      // Breakpad symbol file format cannot express this.
 85      virtual void UndefinedNotSupported(size_t offset, const string& reg);
 86  
 87      // The DWARF CFI entry at OFFSET says that REG uses a DWARF
 88      // expression to find its value, but DwarfCFIToModule is not
 89      // capable of translating DWARF expressions to Breakpad postfix
 90      // expressions.
 91      virtual void ExpressionsNotSupported(size_t offset, const string& reg);
 92  
 93    protected:
 94      string file_, section_;
 95    };
 96  
 97    // Register name tables. If TABLE is a vector returned by one of these
 98    // functions, then TABLE[R] is the name of the register numbered R in
 99    // DWARF call frame information.
100    class RegisterNames {
101     public:
102      // Intel's "x86" or IA-32.
103      static vector<string> I386();
104  
105      // AMD x86_64, AMD64, Intel EM64T, or Intel 64
106      static vector<string> X86_64();
107  
108      // ARM.
109      static vector<string> ARM();
110  
111      // ARM64, aka AARCH64.
112      static vector<string> ARM64();
113  
114      // MIPS.
115      static vector<string> MIPS();
116  
117      // RISC-V.
118      static vector<string> RISCV();
119  
120     private:
121      // Given STRINGS, an array of C strings with SIZE elements, return an
122      // equivalent vector<string>.
123      static vector<string> MakeVector(const char* const* strings, size_t size);
124    };
125  
126    // Create a handler for the CallFrameInfo parser that
127    // records the stack unwinding information it receives in MODULE.
128    //
129    // Use REGISTER_NAMES[I] as the name of register number I; *this
130    // keeps a reference to the vector, so the vector should remain
131    // alive for as long as the DwarfCFIToModule does.
132    //
133    // Use REPORTER for reporting problems encountered in the conversion
134    // process.
135    DwarfCFIToModule(Module* module, const vector<string>& register_names,
136                     Reporter* reporter)
137        : module_(module), register_names_(register_names), reporter_(reporter),
138          return_address_(-1), cfa_name_(".cfa"), ra_name_(".ra") {
139    }
140    virtual ~DwarfCFIToModule() = default;
141  
142    virtual bool Entry(size_t offset, uint64_t address, uint64_t length,
143                       uint8_t version, const string& augmentation,
144                       unsigned return_address);
145    virtual bool UndefinedRule(uint64_t address, int reg);
146    virtual bool SameValueRule(uint64_t address, int reg);
147    virtual bool OffsetRule(uint64_t address, int reg,
148                            int base_register, long offset);
149    virtual bool ValOffsetRule(uint64_t address, int reg,
150                               int base_register, long offset);
151    virtual bool RegisterRule(uint64_t address, int reg, int base_register);
152    virtual bool ExpressionRule(uint64_t address, int reg,
153                                const string& expression);
154    virtual bool ValExpressionRule(uint64_t address, int reg,
155                                   const string& expression);
156    virtual bool End();
157  
158    virtual string Architecture();
159  
160   private:
161    // Return the name to use for register REG.
162    string RegisterName(int i);
163  
164    // Record RULE for register REG at ADDRESS.
165    void Record(Module::Address address, int reg, const string& rule);
166  
167    // The module to which we should add entries.
168    Module* module_;
169  
170    // Map from register numbers to register names.
171    const vector<string>& register_names_;
172  
173    // The reporter to use to report problems.
174    Reporter* reporter_;
175  
176    // The current entry we're constructing.
177    std::unique_ptr<Module::StackFrameEntry> entry_;
178  
179    // The section offset of the current frame description entry, for
180    // use in error messages.
181    size_t entry_offset_;
182  
183    // The return address column for that entry.
184    unsigned return_address_;
185  
186    // The names of the return address and canonical frame address. Putting
187    // these here instead of using string literals allows us to share their
188    // texts in reference-counted string implementations (all the
189    // popular ones). Many, many rules cite these strings.
190    string cfa_name_, ra_name_;
191  
192    // A set of strings used by this CFI. Before storing a string in one of
193    // our data structures, insert it into this set, and then use the string
194    // from the set.
195    //
196    // Because string uses reference counting internally, simply using
197    // strings from this set, even if passed by value, assigned, or held
198    // directly in structures and containers (map<string, ...>, for example),
199    // causes those strings to share a single instance of each distinct piece
200    // of text.
201    set<string> common_strings_;
202  };
203  
204  } // namespace google_breakpad
205  
206  #endif // COMMON_LINUX_DWARF_CFI_TO_MODULE_H