// Copyright (c) 2006, Google Inc. // All rights reserved. // // Redistribution and use in source and binary forms, with or without // modification, are permitted provided that the following conditions are // met: // // * Redistributions of source code must retain the above copyright // notice, this list of conditions and the following disclaimer. // * Redistributions in binary form must reproduce the above // copyright notice, this list of conditions and the following disclaimer // in the documentation and/or other materials provided with the // distribution. // * Neither the name of Google Inc. nor the names of its // contributors may be used to endorse or promote products derived from // this software without specific prior written permission. // // THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS // "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT // LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR // A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT // OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, // SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT // LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, // DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY // THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT // (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE // OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. // minidump_stackwalk.cc: Process a minidump with MinidumpProcessor, printing // the results, including stack traces. // // Author: Mark Mentovai #include #include #include #include #include #include "google_breakpad/processor/basic_source_line_resolver.h" #include "google_breakpad/processor/call_stack.h" #include "google_breakpad/processor/code_module.h" #include "google_breakpad/processor/code_modules.h" #include "google_breakpad/processor/minidump.h" #include "google_breakpad/processor/minidump_processor.h" #include "google_breakpad/processor/process_state.h" #include "google_breakpad/processor/stack_frame_cpu.h" #include "processor/logging.h" #include "processor/pathname_stripper.h" #include "processor/scoped_ptr.h" #include "processor/simple_symbol_supplier.h" namespace { using std::string; using std::vector; using google_breakpad::BasicSourceLineResolver; using google_breakpad::CallStack; using google_breakpad::CodeModule; using google_breakpad::CodeModules; using google_breakpad::MinidumpModule; using google_breakpad::MinidumpProcessor; using google_breakpad::PathnameStripper; using google_breakpad::ProcessState; using google_breakpad::scoped_ptr; using google_breakpad::SimpleSymbolSupplier; using google_breakpad::StackFrame; using google_breakpad::StackFramePPC; using google_breakpad::StackFrameSPARC; using google_breakpad::StackFrameX86; using google_breakpad::StackFrameAMD64; // Separator character for machine readable output. static const char kOutputSeparator = '|'; // PrintRegister prints a register's name and value to stdout. It will // print four registers on a line. For the first register in a set, // pass 0 for |sequence|. For registers in a set, pass the most recent // return value of PrintRegister. Note that PrintRegister will print a // newline before the first register (with |sequence| set to 0) is printed. // The caller is responsible for printing the final newline after a set // of registers is completely printed, regardless of the number of calls // to PrintRegister. static int PrintRegister(const char *name, u_int32_t value, int sequence) { if (sequence % 4 == 0) { printf("\n "); } printf(" %5s = 0x%08x", name, value); return ++sequence; } // StripSeparator takes a string |original| and returns a copy // of the string with all occurences of |kOutputSeparator| removed. static string StripSeparator(const string &original) { string result = original; string::size_type position = 0; while ((position = result.find(kOutputSeparator, position)) != string::npos) { result.erase(position, 1); } return result; } // PrintStack prints the call stack in |stack| to stdout, in a reasonably // useful form. Module, function, and source file names are displayed if // they are available. The code offset to the base code address of the // source line, function, or module is printed, preferring them in that // order. If no source line, function, or module information is available, // an absolute code offset is printed. // // If |cpu| is a recognized CPU name, relevant register state for each stack // frame printed is also output, if available. static void PrintStack(const CallStack *stack, const string &cpu) { int frame_count = stack->frames()->size(); for (int frame_index = 0; frame_index < frame_count; ++frame_index) { const StackFrame *frame = stack->frames()->at(frame_index); printf("%2d ", frame_index); if (frame->module) { printf("%s", PathnameStripper::File(frame->module->code_file()).c_str()); if (!frame->function_name.empty()) { printf("!%s", frame->function_name.c_str()); if (!frame->source_file_name.empty()) { string source_file = PathnameStripper::File(frame->source_file_name); printf(" [%s : %d + 0x%" PRIx64 "]", source_file.c_str(), frame->source_line, frame->instruction - frame->source_line_base); } else { printf(" + 0x%" PRIx64, frame->instruction - frame->function_base); } } else { printf(" + 0x%" PRIx64, frame->instruction - frame->module->base_address()); } } else { printf("0x%" PRIx64, frame->instruction); } int sequence = 0; if (cpu == "x86") { const StackFrameX86 *frame_x86 = reinterpret_cast(frame); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_EIP) sequence = PrintRegister("eip", frame_x86->context.eip, sequence); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_ESP) sequence = PrintRegister("esp", frame_x86->context.esp, sequence); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_EBP) sequence = PrintRegister("ebp", frame_x86->context.ebp, sequence); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_EBX) sequence = PrintRegister("ebx", frame_x86->context.ebx, sequence); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_ESI) sequence = PrintRegister("esi", frame_x86->context.esi, sequence); if (frame_x86->context_validity & StackFrameX86::CONTEXT_VALID_EDI) sequence = PrintRegister("edi", frame_x86->context.edi, sequence); if (frame_x86->context_validity == StackFrameX86::CONTEXT_VALID_ALL) { sequence = PrintRegister("eax", frame_x86->context.eax, sequence); sequence = PrintRegister("ecx", frame_x86->context.ecx, sequence); sequence = PrintRegister("edx", frame_x86->context.edx, sequence); sequence = PrintRegister("efl", frame_x86->context.eflags, sequence); } } else if (cpu == "ppc") { const StackFramePPC *frame_ppc = reinterpret_cast(frame); if (frame_ppc->context_validity & StackFramePPC::CONTEXT_VALID_SRR0) sequence = PrintRegister("srr0", frame_ppc->context.srr0, sequence); if (frame_ppc->context_validity & StackFramePPC::CONTEXT_VALID_GPR1) sequence = PrintRegister("r1", frame_ppc->context.gpr[1], sequence); } else if (cpu == "amd64") { const StackFrameAMD64 *frame_amd64 = reinterpret_cast(frame); if (frame_amd64->context_validity & StackFrameAMD64::CONTEXT_VALID_RIP) sequence = PrintRegister("rip", frame_amd64->context.rip, sequence); if (frame_amd64->context_validity & StackFrameAMD64::CONTEXT_VALID_RSP) sequence = PrintRegister("rsp", frame_amd64->context.rsp, sequence); if (frame_amd64->context_validity & StackFrameAMD64::CONTEXT_VALID_RBP) sequence = PrintRegister("rbp", frame_amd64->context.rbp, sequence); } else if (cpu == "sparc") { const StackFrameSPARC *frame_sparc = reinterpret_cast(frame); if (frame_sparc->context_validity & StackFrameSPARC::CONTEXT_VALID_SP) sequence = PrintRegister("sp", frame_sparc->context.g_r[14], sequence); if (frame_sparc->context_validity & StackFrameSPARC::CONTEXT_VALID_FP) sequence = PrintRegister("fp", frame_sparc->context.g_r[30], sequence); if (frame_sparc->context_validity & StackFrameSPARC::CONTEXT_VALID_PC) sequence = PrintRegister("pc", frame_sparc->context.pc, sequence); } printf("\n"); } } // PrintStackMachineReadable prints the call stack in |stack| to stdout, // in the following machine readable pipe-delimited text format: // thread number|frame number|module|function|source file|line|offset // // Module, function, source file, and source line may all be empty // depending on availability. The code offset follows the same rules as // PrintStack above. static void PrintStackMachineReadable(int thread_num, const CallStack *stack) { int frame_count = stack->frames()->size(); for (int frame_index = 0; frame_index < frame_count; ++frame_index) { const StackFrame *frame = stack->frames()->at(frame_index); printf("%d%c%d%c", thread_num, kOutputSeparator, frame_index, kOutputSeparator); if (frame->module) { assert(!frame->module->code_file().empty()); printf("%s", StripSeparator(PathnameStripper::File( frame->module->code_file())).c_str()); if (!frame->function_name.empty()) { printf("%c%s", kOutputSeparator, StripSeparator(frame->function_name).c_str()); if (!frame->source_file_name.empty()) { printf("%c%s%c%d%c0x%" PRIx64, kOutputSeparator, StripSeparator(frame->source_file_name).c_str(), kOutputSeparator, frame->source_line, kOutputSeparator, frame->instruction - frame->source_line_base); } else { printf("%c%c%c0x%" PRIx64, kOutputSeparator, // empty source file kOutputSeparator, // empty source line kOutputSeparator, frame->instruction - frame->function_base); } } else { printf("%c%c%c%c0x%" PRIx64, kOutputSeparator, // empty function name kOutputSeparator, // empty source file kOutputSeparator, // empty source line kOutputSeparator, frame->instruction - frame->module->base_address()); } } else { // the printf before this prints a trailing separator for module name printf("%c%c%c%c0x%" PRIx64, kOutputSeparator, // empty function name kOutputSeparator, // empty source file kOutputSeparator, // empty source line kOutputSeparator, frame->instruction); } printf("\n"); } } static void PrintModules(const CodeModules *modules) { if (!modules) return; printf("\n"); printf("Loaded modules:\n"); u_int64_t main_address = 0; const CodeModule *main_module = modules->GetMainModule(); if (main_module) { main_address = main_module->base_address(); } unsigned int module_count = modules->module_count(); for (unsigned int module_sequence = 0; module_sequence < module_count; ++module_sequence) { const CodeModule *module = modules->GetModuleAtSequence(module_sequence); u_int64_t base_address = module->base_address(); printf("0x%08" PRIx64 " - 0x%08" PRIx64 " %s %s%s\n", base_address, base_address + module->size() - 1, PathnameStripper::File(module->code_file()).c_str(), module->version().empty() ? "???" : module->version().c_str(), main_module != NULL && base_address == main_address ? " (main)" : ""); } } // PrintModulesMachineReadable outputs a list of loaded modules, // one per line, in the following machine-readable pipe-delimited // text format: // Module|{Module Filename}|{Version}|{Debug Filename}|{Debug Identifier}| // {Base Address}|{Max Address}|{Main} static void PrintModulesMachineReadable(const CodeModules *modules) { if (!modules) return; u_int64_t main_address = 0; const CodeModule *main_module = modules->GetMainModule(); if (main_module) { main_address = main_module->base_address(); } unsigned int module_count = modules->module_count(); for (unsigned int module_sequence = 0; module_sequence < module_count; ++module_sequence) { const CodeModule *module = modules->GetModuleAtSequence(module_sequence); u_int64_t base_address = module->base_address(); printf("Module%c%s%c%s%c%s%c%s%c0x%08" PRIx64 "%c0x%08" PRIx64 "%c%d\n", kOutputSeparator, StripSeparator(PathnameStripper::File(module->code_file())).c_str(), kOutputSeparator, StripSeparator(module->version()).c_str(), kOutputSeparator, StripSeparator(PathnameStripper::File(module->debug_file())).c_str(), kOutputSeparator, StripSeparator(module->debug_identifier()).c_str(), kOutputSeparator, base_address, kOutputSeparator, base_address + module->size() - 1, kOutputSeparator, main_module != NULL && base_address == main_address ? 1 : 0); } } static void PrintProcessState(const ProcessState& process_state) { // Print OS and CPU information. string cpu = process_state.system_info()->cpu; string cpu_info = process_state.system_info()->cpu_info; printf("Operating system: %s\n", process_state.system_info()->os.c_str()); printf(" %s\n", process_state.system_info()->os_version.c_str()); printf("CPU: %s\n", cpu.c_str()); if (!cpu_info.empty()) { // This field is optional. printf(" %s\n", cpu_info.c_str()); } printf(" %d CPU%s\n", process_state.system_info()->cpu_count, process_state.system_info()->cpu_count != 1 ? "s" : ""); printf("\n"); // Print crash information. if (process_state.crashed()) { printf("Crash reason: %s\n", process_state.crash_reason().c_str()); printf("Crash address: 0x%" PRIx64 "\n", process_state.crash_address()); } else { printf("No crash\n"); } // If the thread that requested the dump is known, print it first. int requesting_thread = process_state.requesting_thread(); if (requesting_thread != -1) { printf("\n"); printf("Thread %d (%s)\n", requesting_thread, process_state.crashed() ? "crashed" : "requested dump, did not crash"); PrintStack(process_state.threads()->at(requesting_thread), cpu); } // Print all of the threads in the dump. int thread_count = process_state.threads()->size(); for (int thread_index = 0; thread_index < thread_count; ++thread_index) { if (thread_index != requesting_thread) { // Don't print the crash thread again, it was already printed. printf("\n"); printf("Thread %d\n", thread_index); PrintStack(process_state.threads()->at(thread_index), cpu); } } PrintModules(process_state.modules()); } static void PrintProcessStateMachineReadable(const ProcessState& process_state) { // Print OS and CPU information. // OS|{OS Name}|{OS Version} // CPU|{CPU Name}|{CPU Info}|{Number of CPUs} printf("OS%c%s%c%s\n", kOutputSeparator, StripSeparator(process_state.system_info()->os).c_str(), kOutputSeparator, StripSeparator(process_state.system_info()->os_version).c_str()); printf("CPU%c%s%c%s%c%d\n", kOutputSeparator, StripSeparator(process_state.system_info()->cpu).c_str(), kOutputSeparator, // this may be empty StripSeparator(process_state.system_info()->cpu_info).c_str(), kOutputSeparator, process_state.system_info()->cpu_count); int requesting_thread = process_state.requesting_thread(); // Print crash information. // Crash|{Crash Reason}|{Crash Address}|{Crashed Thread} printf("Crash%c", kOutputSeparator); if (process_state.crashed()) { printf("%s%c0x%" PRIx64 "%c", StripSeparator(process_state.crash_reason()).c_str(), kOutputSeparator, process_state.crash_address(), kOutputSeparator); } else { printf("No crash%c%c", kOutputSeparator, kOutputSeparator); } if (requesting_thread != -1) { printf("%d\n", requesting_thread); } else { printf("\n"); } PrintModulesMachineReadable(process_state.modules()); // blank line to indicate start of threads printf("\n"); // If the thread that requested the dump is known, print it first. if (requesting_thread != -1) { PrintStackMachineReadable(requesting_thread, process_state.threads()->at(requesting_thread)); } // Print all of the threads in the dump. int thread_count = process_state.threads()->size(); for (int thread_index = 0; thread_index < thread_count; ++thread_index) { if (thread_index != requesting_thread) { // Don't print the crash thread again, it was already printed. PrintStackMachineReadable(thread_index, process_state.threads()->at(thread_index)); } } } // Processes |minidump_file| using MinidumpProcessor. |symbol_path|, if // non-empty, is the base directory of a symbol storage area, laid out in // the format required by SimpleSymbolSupplier. If such a storage area // is specified, it is made available for use by the MinidumpProcessor. // // Returns the value of MinidumpProcessor::Process. If processing succeeds, // prints identifying OS and CPU information from the minidump, crash // information if the minidump was produced as a result of a crash, and // call stacks for each thread contained in the minidump. All information // is printed to stdout. static bool PrintMinidumpProcess(const string &minidump_file, const vector &symbol_paths, bool machine_readable) { scoped_ptr symbol_supplier; if (!symbol_paths.empty()) { // TODO(mmentovai): check existence of symbol_path if specified? symbol_supplier.reset(new SimpleSymbolSupplier(symbol_paths)); } BasicSourceLineResolver resolver; MinidumpProcessor minidump_processor(symbol_supplier.get(), &resolver); // Process the minidump. ProcessState process_state; if (minidump_processor.Process(minidump_file, &process_state) != MinidumpProcessor::PROCESS_OK) { BPLOG(ERROR) << "MinidumpProcessor::Process failed"; return false; } if (machine_readable) { PrintProcessStateMachineReadable(process_state); } else { PrintProcessState(process_state); } return true; } } // namespace static void usage(const char *program_name) { fprintf(stderr, "usage: %s [-m] [symbol-path ...]\n" " -m : Output in machine-readable format\n", program_name); } int main(int argc, char **argv) { BPLOG_INIT(&argc, &argv); if (argc < 2) { usage(argv[0]); return 1; } const char *minidump_file; bool machine_readable; int symbol_path_arg; if (strcmp(argv[1], "-m") == 0) { if (argc < 3) { usage(argv[0]); return 1; } machine_readable = true; minidump_file = argv[2]; symbol_path_arg = 3; } else { machine_readable = false; minidump_file = argv[1]; symbol_path_arg = 2; } // extra arguments are symbol paths std::vector symbol_paths; if (argc > symbol_path_arg) { for (int argi = symbol_path_arg; argi < argc; ++argi) symbol_paths.push_back(argv[argi]); } return PrintMinidumpProcess(minidump_file, symbol_paths, machine_readable) ? 0 : 1; }