michael@0: // Copyright (c) 2011 Google Inc.
michael@0: // All rights reserved.
michael@0: //
michael@0: // Redistribution and use in source and binary forms, with or without
michael@0: // modification, are permitted provided that the following conditions are
michael@0: // met:
michael@0: //
michael@0: //     * Redistributions of source code must retain the above copyright
michael@0: // notice, this list of conditions and the following disclaimer.
michael@0: //     * Redistributions in binary form must reproduce the above
michael@0: // copyright notice, this list of conditions and the following disclaimer
michael@0: // in the documentation and/or other materials provided with the
michael@0: // distribution.
michael@0: //     * Neither the name of Google Inc. nor the names of its
michael@0: // contributors may be used to endorse or promote products derived from
michael@0: // this software without specific prior written permission.
michael@0: //
michael@0: // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
michael@0: // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
michael@0: // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
michael@0: // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
michael@0: // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
michael@0: // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
michael@0: // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
michael@0: // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
michael@0: // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
michael@0: // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
michael@0: // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
michael@0: 
michael@0: // Original author: Jim Blandy <jimb@mozilla.com> <jimb@red-bean.com>
michael@0: 
michael@0: // module.cc: Implement google_breakpad::Module.  See module.h.
michael@0: 
michael@0: #include "common/module.h"
michael@0: 
michael@0: #include <assert.h>
michael@0: #include <errno.h>
michael@0: #include <stdio.h>
michael@0: #include <string.h>
michael@0: 
michael@0: #include <algorithm>
michael@0: #include <iostream>
michael@0: #include <utility>
michael@0: 
michael@0: namespace google_breakpad {
michael@0: 
michael@0: using std::dec;
michael@0: using std::endl;
michael@0: using std::hex;
michael@0: 
michael@0: 
michael@0: Module::Module(const string &name, const string &os,
michael@0:                const string &architecture, const string &id) :
michael@0:     name_(name),
michael@0:     os_(os),
michael@0:     architecture_(architecture),
michael@0:     id_(id),
michael@0:     load_address_(0) { }
michael@0: 
michael@0: Module::~Module() {
michael@0:   for (FileByNameMap::iterator it = files_.begin(); it != files_.end(); ++it)
michael@0:     delete it->second;
michael@0:   for (FunctionSet::iterator it = functions_.begin();
michael@0:        it != functions_.end(); ++it) {
michael@0:     delete *it;
michael@0:   }
michael@0:   for (StackFrameEntrySet::iterator it = stack_frame_entries_.begin();
michael@0:        it != stack_frame_entries_.end(); ++it) {
michael@0:     delete *it;
michael@0:   }
michael@0:   for (ExternSet::iterator it = externs_.begin(); it != externs_.end(); ++it)
michael@0:     delete *it;
michael@0: }
michael@0: 
michael@0: void Module::SetLoadAddress(Address address) {
michael@0:   load_address_ = address;
michael@0: }
michael@0: 
michael@0: void Module::AddFunction(Function *function) {
michael@0:   // FUNC lines must not hold an empty name, so catch the problem early if
michael@0:   // callers try to add one.
michael@0:   assert(!function->name.empty());
michael@0:   std::pair<FunctionSet::iterator,bool> ret = functions_.insert(function);
michael@0:   if (!ret.second) {
michael@0:     // Free the duplicate that was not inserted because this Module
michael@0:     // now owns it.
michael@0:     delete function;
michael@0:   }
michael@0: }
michael@0: 
michael@0: void Module::AddFunctions(vector<Function *>::iterator begin,
michael@0:                           vector<Function *>::iterator end) {
michael@0:   for (vector<Function *>::iterator it = begin; it != end; ++it)
michael@0:     AddFunction(*it);
michael@0: }
michael@0: 
michael@0: void Module::AddStackFrameEntry(StackFrameEntry* stack_frame_entry) {
michael@0:   std::pair<StackFrameEntrySet::iterator,bool> ret =
michael@0:       stack_frame_entries_.insert(stack_frame_entry);
michael@0:   if (!ret.second) {
michael@0:     // Free the duplicate that was not inserted because this Module
michael@0:     // now owns it.
michael@0:     delete stack_frame_entry;
michael@0:   }
michael@0: }
michael@0: 
michael@0: void Module::AddExtern(Extern *ext) {
michael@0:   std::pair<ExternSet::iterator,bool> ret = externs_.insert(ext);
michael@0:   if (!ret.second) {
michael@0:     // Free the duplicate that was not inserted because this Module
michael@0:     // now owns it.
michael@0:     delete ext;
michael@0:   }
michael@0: }
michael@0: 
michael@0: void Module::GetFunctions(vector<Function *> *vec,
michael@0:                           vector<Function *>::iterator i) {
michael@0:   vec->insert(i, functions_.begin(), functions_.end());
michael@0: }
michael@0: 
michael@0: template<typename T>
michael@0: bool EntryContainsAddress(T entry, Module::Address address) {
michael@0:   return entry->address <= address && address < entry->address + entry->size;
michael@0: }
michael@0: 
michael@0: Module::Function* Module::FindFunctionByAddress(Address address) {
michael@0:   Function search;
michael@0:   search.address = address;
michael@0:   // Ensure that name always sorts higher than the function name,
michael@0:   // so that upper_bound always returns the function just after
michael@0:   // the function containing this address.
michael@0:   search.name = "\xFF";
michael@0:   FunctionSet::iterator it = functions_.upper_bound(&search);
michael@0:   if (it == functions_.begin())
michael@0:     return NULL;
michael@0: 
michael@0:   it--;
michael@0: 
michael@0:   if (EntryContainsAddress(*it, address))
michael@0:     return *it;
michael@0: 
michael@0:   return NULL;
michael@0: }
michael@0: 
michael@0: void Module::GetExterns(vector<Extern *> *vec,
michael@0:                         vector<Extern *>::iterator i) {
michael@0:   vec->insert(i, externs_.begin(), externs_.end());
michael@0: }
michael@0: 
michael@0: Module::Extern* Module::FindExternByAddress(Address address) {
michael@0:   Extern search;
michael@0:   search.address = address;
michael@0:   ExternSet::iterator it = externs_.upper_bound(&search);
michael@0: 
michael@0:   if (it == externs_.begin())
michael@0:     return NULL;
michael@0: 
michael@0:   it--;
michael@0:   if ((*it)->address > address)
michael@0:     return NULL;
michael@0: 
michael@0:   return *it;
michael@0: }
michael@0: 
michael@0: Module::File *Module::FindFile(const string &name) {
michael@0:   // A tricky bit here.  The key of each map entry needs to be a
michael@0:   // pointer to the entry's File's name string.  This means that we
michael@0:   // can't do the initial lookup with any operation that would create
michael@0:   // an empty entry for us if the name isn't found (like, say,
michael@0:   // operator[] or insert do), because such a created entry's key will
michael@0:   // be a pointer the string passed as our argument.  Since the key of
michael@0:   // a map's value type is const, we can't fix it up once we've
michael@0:   // created our file.  lower_bound does the lookup without doing an
michael@0:   // insertion, and returns a good hint iterator to pass to insert.
michael@0:   // Our "destiny" is where we belong, whether we're there or not now.
michael@0:   FileByNameMap::iterator destiny = files_.lower_bound(&name);
michael@0:   if (destiny == files_.end()
michael@0:       || *destiny->first != name) {  // Repeated string comparison, boo hoo.
michael@0:     File *file = new File;
michael@0:     file->name = name;
michael@0:     file->source_id = -1;
michael@0:     destiny = files_.insert(destiny,
michael@0:                             FileByNameMap::value_type(&file->name, file));
michael@0:   }
michael@0:   return destiny->second;
michael@0: }
michael@0: 
michael@0: Module::File *Module::FindFile(const char *name) {
michael@0:   string name_string = name;
michael@0:   return FindFile(name_string);
michael@0: }
michael@0: 
michael@0: Module::File *Module::FindExistingFile(const string &name) {
michael@0:   FileByNameMap::iterator it = files_.find(&name);
michael@0:   return (it == files_.end()) ? NULL : it->second;
michael@0: }
michael@0: 
michael@0: void Module::GetFiles(vector<File *> *vec) {
michael@0:   vec->clear();
michael@0:   for (FileByNameMap::iterator it = files_.begin(); it != files_.end(); ++it)
michael@0:     vec->push_back(it->second);
michael@0: }
michael@0: 
michael@0: void Module::GetStackFrameEntries(vector<StackFrameEntry *>* vec) {
michael@0:   vec->clear();
michael@0:   vec->insert(vec->begin(), stack_frame_entries_.begin(),
michael@0:               stack_frame_entries_.end());
michael@0: }
michael@0: 
michael@0: Module::StackFrameEntry* Module::FindStackFrameEntryByAddress(Address address) {
michael@0:   StackFrameEntry search;
michael@0:   search.address = address;
michael@0:   StackFrameEntrySet::iterator it = stack_frame_entries_.upper_bound(&search);
michael@0: 
michael@0:   if (it == stack_frame_entries_.begin())
michael@0:     return NULL;
michael@0: 
michael@0:   it--;
michael@0:   if (EntryContainsAddress(*it, address))
michael@0:     return *it;
michael@0: 
michael@0:   return NULL;
michael@0: }
michael@0: 
michael@0: void Module::AssignSourceIds() {
michael@0:   // First, give every source file an id of -1.
michael@0:   for (FileByNameMap::iterator file_it = files_.begin();
michael@0:        file_it != files_.end(); ++file_it) {
michael@0:     file_it->second->source_id = -1;
michael@0:   }
michael@0: 
michael@0:   // Next, mark all files actually cited by our functions' line number
michael@0:   // info, by setting each one's source id to zero.
michael@0:   for (FunctionSet::const_iterator func_it = functions_.begin();
michael@0:        func_it != functions_.end(); ++func_it) {
michael@0:     Function *func = *func_it;
michael@0:     for (vector<Line>::iterator line_it = func->lines.begin();
michael@0:          line_it != func->lines.end(); ++line_it)
michael@0:       line_it->file->source_id = 0;
michael@0:   }
michael@0: 
michael@0:   // Finally, assign source ids to those files that have been marked.
michael@0:   // We could have just assigned source id numbers while traversing
michael@0:   // the line numbers, but doing it this way numbers the files in
michael@0:   // lexicographical order by name, which is neat.
michael@0:   int next_source_id = 0;
michael@0:   for (FileByNameMap::iterator file_it = files_.begin();
michael@0:        file_it != files_.end(); ++file_it) {
michael@0:     if (!file_it->second->source_id)
michael@0:       file_it->second->source_id = next_source_id++;
michael@0:   }
michael@0: }
michael@0: 
michael@0: bool Module::ReportError() {
michael@0:   fprintf(stderr, "error writing symbol file: %s\n",
michael@0:           strerror(errno));
michael@0:   return false;
michael@0: }
michael@0: 
michael@0: std::ostream& operator<<(std::ostream& stream, const Module::Expr& expr) {
michael@0:   assert(!expr.isExprInvalid());
michael@0:   switch (expr.how_) {
michael@0:     case Module::kExprSimple:
michael@0:       stream << FromUniqueString(expr.ident_) << " " << expr.offset_ << " +";
michael@0:       break;
michael@0:     case Module::kExprSimpleMem:
michael@0:       stream << FromUniqueString(expr.ident_) << " " << expr.offset_ << " + ^";
michael@0:       break;
michael@0:     case Module::kExprPostfix:
michael@0:       stream << expr.postfix_; break;
michael@0:     case Module::kExprInvalid:
michael@0:     default:
michael@0:       break;
michael@0:   }
michael@0:   return stream;
michael@0: }
michael@0: 
michael@0: bool Module::WriteRuleMap(const RuleMap &rule_map, std::ostream &stream) {
michael@0:   // Visit the register rules in alphabetical order.  Because
michael@0:   // rule_map has the elements in some arbitrary order,
michael@0:   // get the names out into a vector, sort them, and visit in
michael@0:   // sorted order.
michael@0:   std::vector<const UniqueString*> rr_names;
michael@0:   for (RuleMap::const_iterator it = rule_map.begin();
michael@0:        it != rule_map.end(); ++it) {
michael@0:     rr_names.push_back(it->first);
michael@0:   }
michael@0: 
michael@0:   std::sort(rr_names.begin(), rr_names.end(), LessThan_UniqueString);
michael@0: 
michael@0:   // Now visit the register rules in alphabetical order.
michael@0:   for (std::vector<const UniqueString*>::const_iterator name = rr_names.begin();
michael@0:        name != rr_names.end();
michael@0:        ++name) {
michael@0:     if (name != rr_names.begin())
michael@0:       stream << " ";
michael@0:     stream << FromUniqueString(*name) << ": " << rule_map.find(*name)->second;
michael@0:   }
michael@0:   return stream.good();
michael@0: }
michael@0: 
michael@0: bool Module::Write(std::ostream &stream, SymbolData symbol_data) {
michael@0:   stream << "MODULE " << os_ << " " << architecture_ << " "
michael@0:          << id_ << " " << name_ << endl;
michael@0:   if (!stream.good())
michael@0:     return ReportError();
michael@0: 
michael@0:   if (symbol_data != ONLY_CFI) {
michael@0:     AssignSourceIds();
michael@0: 
michael@0:     // Write out files.
michael@0:     for (FileByNameMap::iterator file_it = files_.begin();
michael@0:          file_it != files_.end(); ++file_it) {
michael@0:       File *file = file_it->second;
michael@0:       if (file->source_id >= 0) {
michael@0:         stream << "FILE " << file->source_id << " " <<  file->name << endl;
michael@0:         if (!stream.good())
michael@0:           return ReportError();
michael@0:       }
michael@0:     }
michael@0: 
michael@0:     // Write out functions and their lines.
michael@0:     for (FunctionSet::const_iterator func_it = functions_.begin();
michael@0:          func_it != functions_.end(); ++func_it) {
michael@0:       Function *func = *func_it;
michael@0:       stream << "FUNC " << hex
michael@0:              << (func->address - load_address_) << " "
michael@0:              << func->size << " "
michael@0:              << func->parameter_size << " "
michael@0:              << func->name << dec << endl;
michael@0:       if (!stream.good())
michael@0:         return ReportError();
michael@0: 
michael@0:       for (vector<Line>::iterator line_it = func->lines.begin();
michael@0:            line_it != func->lines.end(); ++line_it) {
michael@0:         stream << hex
michael@0:                << (line_it->address - load_address_) << " "
michael@0:                << line_it->size << " "
michael@0:                << dec
michael@0:                << line_it->number << " "
michael@0:                << line_it->file->source_id << endl;
michael@0:         if (!stream.good())
michael@0:           return ReportError();
michael@0:       }
michael@0:     }
michael@0: 
michael@0:     // Write out 'PUBLIC' records.
michael@0:     for (ExternSet::const_iterator extern_it = externs_.begin();
michael@0:          extern_it != externs_.end(); ++extern_it) {
michael@0:       Extern *ext = *extern_it;
michael@0:       stream << "PUBLIC " << hex
michael@0:              << (ext->address - load_address_) << " 0 "
michael@0:              << ext->name << dec << endl;
michael@0:     }
michael@0:   }
michael@0: 
michael@0:   if (symbol_data != NO_CFI) {
michael@0:     // Write out 'STACK CFI INIT' and 'STACK CFI' records.
michael@0:     StackFrameEntrySet::const_iterator frame_it;
michael@0:     for (frame_it = stack_frame_entries_.begin();
michael@0:          frame_it != stack_frame_entries_.end(); ++frame_it) {
michael@0:       StackFrameEntry *entry = *frame_it;
michael@0:       stream << "STACK CFI INIT " << hex
michael@0:              << (entry->address - load_address_) << " "
michael@0:              << entry->size << " " << dec;
michael@0:       if (!stream.good()
michael@0:           || !WriteRuleMap(entry->initial_rules, stream))
michael@0:         return ReportError();
michael@0: 
michael@0:       stream << endl;
michael@0: 
michael@0:       // Write out this entry's delta rules as 'STACK CFI' records.
michael@0:       for (RuleChangeMap::const_iterator delta_it = entry->rule_changes.begin();
michael@0:            delta_it != entry->rule_changes.end(); ++delta_it) {
michael@0:         stream << "STACK CFI " << hex
michael@0:                << (delta_it->first - load_address_) << " " << dec;
michael@0:         if (!stream.good()
michael@0:             || !WriteRuleMap(delta_it->second, stream))
michael@0:           return ReportError();
michael@0: 
michael@0:         stream << endl;
michael@0:       }
michael@0:     }
michael@0:   }
michael@0: 
michael@0:   return true;
michael@0: }
michael@0: 
michael@0: }  // namespace google_breakpad