Skip to content

java jdk 为什么一个线程空指针不退出

Posted on:September 6, 2023 at 08:03 PM

背景

了解jdk异常的捕获原理

堆栈

(gdb) bt
#0  PosixSignals::pd_hotspot_signal_handler (sig=sig@entry=11, info=info@entry=0x7ffff7bfd330, uc=uc@entry=0x7ffff7bfd200, thread=0x7ffff00295a0) at /home/ubuntu/jdk/src/hotspot/os_cpu/linux_x86/os_linux_x86.cpp:201
#1  0x00007ffff7090f7d in JVM_handle_linux_signal (abort_if_unrecognized=1, ucVoid=0x7ffff7bfd200, info=0x7ffff7bfd330, sig=11) at /home/ubuntu/jdk/src/hotspot/os/posix/signals_posix.cpp:656
#2  JVM_handle_linux_signal (sig=11, info=0x7ffff7bfd330, ucVoid=0x7ffff7bfd200, abort_if_unrecognized=1) at /home/ubuntu/jdk/src/hotspot/os/posix/signals_posix.cpp:557
#3  <signal handler called>
#4  0x00007fffe8537640 in ?? ()
#5  0x0000000000000246 in ?? ()
#6  0x00007fffe8537734 in ?? ()
#7  0x00007ffff79f1858 in ?? () from /home/ubuntu/jdk/build/linux-x86_64-server-fastdebug/jdk/lib/server/libjvm.so
#8  0x00007ffff7bfe290 in ?? ()
#9  0x00007ffff734777a in VM_Version::get_processor_features () at /home/ubuntu/jdk/src/hotspot/cpu/x86/vm_version_x86.cpp:803

这里会返回true , 然后就跳过jdk的退出

bool PosixSignals::pd_hotspot_signal_handler(int sig, siginfo_t* info,
                                             ucontext_t* uc, JavaThread* thread) {

  /*
  NOTE: does not seem to work on linux.
  if (info == NULL || info->si_code <= 0 || info->si_code == SI_NOINFO) {
    // can't decode this kind of signal
    info = NULL;
  } else {
    assert(sig == info->si_signo, "bad siginfo");
  }
*/
  // decide if this trap can be handled by a stub
  address stub = NULL;

  address pc          = NULL;

  //%note os_trap_1
  if (info != NULL && uc != NULL && thread != NULL) {
    pc = (address) os::Posix::ucontext_get_pc(uc);

    if (sig == SIGSEGV && info->si_addr == 0 && info->si_code == SI_KERNEL) {
      // An irrecoverable SI_KERNEL SIGSEGV has occurred.
      // It's likely caused by dereferencing an address larger than TASK_SIZE.
      return false;
    }

    // Handle ALL stack overflow variations here
    if (sig == SIGSEGV) {
      address addr = (address) info->si_addr;

      // check if fault address is within thread stack
      if (thread->is_in_full_stack(addr)) {
        // stack overflow
        if (os::Posix::handle_stack_overflow(thread, addr, pc, uc, &stub)) {
          return true; // continue
        }
      }
    }

    if ((sig == SIGSEGV) && VM_Version::is_cpuinfo_segv_addr(pc)) {
      // Verify that OS save/restore AVX registers.
      stub = VM_Version::cpuinfo_cont_addr();
    }

    if (thread->thread_state() == _thread_in_Java) {
      // Java thread running in Java code => find exception handler if any
      // a fault inside compiled code, the interpreter, or a stub

      if (sig == SIGSEGV && SafepointMechanism::is_poll_address((address)info->si_addr)) {
        stub = SharedRuntime::get_poll_stub(pc);
      } else if (sig == SIGBUS /* && info->si_code == BUS_OBJERR */) {
        // BugId 4454115: A read from a MappedByteBuffer can fault
        // here if the underlying file has been truncated.
        // Do not crash the VM in such a case.
        CodeBlob* cb = CodeCache::find_blob(pc);
        CompiledMethod* nm = (cb != NULL) ? cb->as_compiled_method_or_null() : NULL;
        bool is_unsafe_arraycopy = thread->doing_unsafe_access() && UnsafeCopyMemory::contains_pc(pc);
        if ((nm != NULL && nm->has_unsafe_access()) || is_unsafe_arraycopy) {
          address next_pc = Assembler::locate_next_instruction(pc);
          if (is_unsafe_arraycopy) {
            next_pc = UnsafeCopyMemory::page_error_continue_pc(pc);
          }
          stub = SharedRuntime::handle_unsafe_access(thread, next_pc);
        }
      }
      else

#ifdef AMD64
      if (sig == SIGFPE  &&
          (info->si_code == FPE_INTDIV || info->si_code == FPE_FLTDIV)) {
        stub =
          SharedRuntime::
          continuation_for_implicit_exception(thread,
                                              pc,
                                              SharedRuntime::
                                              IMPLICIT_DIVIDE_BY_ZERO);
#else
      if (sig == SIGFPE /* && info->si_code == FPE_INTDIV */) {
        // HACK: si_code does not work on linux 2.2.12-20!!!
        int op = pc[0];
        if (op == 0xDB) {
          // FIST
          // TODO: The encoding of D2I in x86_32.ad can cause an exception
          // prior to the fist instruction if there was an invalid operation
          // pending. We want to dismiss that exception. From the win_32
          // side it also seems that if it really was the fist causing
          // the exception that we do the d2i by hand with different
          // rounding. Seems kind of weird.
          // NOTE: that we take the exception at the NEXT floating point instruction.
          assert(pc[0] == 0xDB, "not a FIST opcode");
          assert(pc[1] == 0x14, "not a FIST opcode");
          assert(pc[2] == 0x24, "not a FIST opcode");
          return true;
        } else if (op == 0xF7) {
          // IDIV
          stub = SharedRuntime::continuation_for_implicit_exception(thread, pc, SharedRuntime::IMPLICIT_DIVIDE_BY_ZERO);
        } else {
          // TODO: handle more cases if we are using other x86 instructions
          //   that can generate SIGFPE signal on linux.
          tty->print_cr("unknown opcode 0x%X with SIGFPE.", op);
          fatal("please update this code.");
        }
#endif // AMD64
      } else if (sig == SIGSEGV &&
                 MacroAssembler::uses_implicit_null_check(info->si_addr)) {
          // Determination of interpreter/vtable stub/compiled code null exception
          stub = SharedRuntime::continuation_for_implicit_exception(thread, pc, SharedRuntime::IMPLICIT_NULL);
      }
    } else if ((thread->thread_state() == _thread_in_vm ||
                thread->thread_state() == _thread_in_native) &&
               (sig == SIGBUS && /* info->si_code == BUS_OBJERR && */
               thread->doing_unsafe_access())) {
        address next_pc = Assembler::locate_next_instruction(pc);
        if (UnsafeCopyMemory::contains_pc(pc)) {
          next_pc = UnsafeCopyMemory::page_error_continue_pc(pc);
        }
        stub = SharedRuntime::handle_unsafe_access(thread, next_pc);
    }

    // jni_fast_Get<Primitive>Field can trap at certain pc's if a GC kicks in
    // and the heap gets shrunk before the field access.
    if ((sig == SIGSEGV) || (sig == SIGBUS)) {
      address addr = JNI_FastGetField::find_slowcase_pc(pc);
      if (addr != (address)-1) {
        stub = addr;
      }
    }
  }

#ifndef AMD64
  // Execution protection violation
  //
  // This should be kept as the last step in the triage.  We don't
  // have a dedicated trap number for a no-execute fault, so be
  // conservative and allow other handlers the first shot.
  //
  // Note: We don't test that info->si_code == SEGV_ACCERR here.
  // this si_code is so generic that it is almost meaningless; and
  // the si_code for this condition may change in the future.
  // Furthermore, a false-positive should be harmless.
  if (UnguardOnExecutionViolation > 0 &&
      stub == NULL &&
      (sig == SIGSEGV || sig == SIGBUS) &&
      uc->uc_mcontext.gregs[REG_TRAPNO] == trap_page_fault) {
    int page_size = os::vm_page_size();
    address addr = (address) info->si_addr;
    address pc = os::Posix::ucontext_get_pc(uc);
    // Make sure the pc and the faulting address are sane.
    //
    // If an instruction spans a page boundary, and the page containing
    // the beginning of the instruction is executable but the following
    // page is not, the pc and the faulting address might be slightly
    // different - we still want to unguard the 2nd page in this case.
    //
    // 15 bytes seems to be a (very) safe value for max instruction size.
    bool pc_is_near_addr =
      (pointer_delta((void*) addr, (void*) pc, sizeof(char)) < 15);
    bool instr_spans_page_boundary =
      (align_down((intptr_t) pc ^ (intptr_t) addr,
                       (intptr_t) page_size) > 0);

    if (pc == addr || (pc_is_near_addr && instr_spans_page_boundary)) {
      static volatile address last_addr =
        (address) os::non_memory_address_word();

      // In conservative mode, don't unguard unless the address is in the VM
      if (addr != last_addr &&
          (UnguardOnExecutionViolation > 1 || os::address_is_in_vm(addr))) {

        // Set memory to RWX and retry
        address page_start = align_down(addr, page_size);
        bool res = os::protect_memory((char*) page_start, page_size,
                                      os::MEM_PROT_RWX);

        log_debug(os)("Execution protection violation "
                      "at " INTPTR_FORMAT
                      ", unguarding " INTPTR_FORMAT ": %s, errno=%d", p2i(addr),
                      p2i(page_start), (res ? "success" : "failed"), errno);
        stub = pc;

        // Set last_addr so if we fault again at the same address, we don't end
        // up in an endless loop.
        //
        // There are two potential complications here.  Two threads trapping at
        // the same address at the same time could cause one of the threads to
        // think it already unguarded, and abort the VM.  Likely very rare.
        //
        // The other race involves two threads alternately trapping at
        // different addresses and failing to unguard the page, resulting in
        // an endless loop.  This condition is probably even more unlikely than
        // the first.
        //
        // Although both cases could be avoided by using locks or thread local
        // last_addr, these solutions are unnecessary complication: this
        // handler is a best-effort safety net, not a complete solution.  It is
        // disabled by default and should only be used as a workaround in case
        // we missed any no-execute-unsafe VM code.

        last_addr = addr;
      }
    }
  }
#endif // !AMD64

  if (stub != NULL) {
    // save all thread context in case we need to restore it
    if (thread != NULL) thread->set_saved_exception_pc(pc);

    os::Posix::ucontext_set_pc(uc, stub);
    return true;     /////////////////////////////////////////  这里会是true
  }

  return false;
}

相关阅读