crash in JvmtiExport::post_compiled_method_load

serguei.spitsyn at oracle.com serguei.spitsyn at oracle.com
Wed May 9 08:02:32 UTC 2018


Hi Ioi and David,

Ioi, thank you for the disassembly.
It helps for sure!


Below are 3 call sites for the nmethod::post_compiled_method_load_event():

void ciEnv::register_method(ciMethod* target, . . .) {
   . . .
   if (nm != NULL) {
     // JVMTI -- compiled method notification (must be done outside lock)
nm->post_compiled_method_load_event(); <== !!
   } else {
     . . .
   }

}

JVMCIEnv::CodeInstallResult JVMCIEnv::register_method(. . .) {
   . . .
   // JVMTI -- compiled method notification (must be done outside lock)
   if (nm != NULL) {
nm->post_compiled_method_load_event(); <== !!
     . . .
   }
   return result;
}

void AdapterHandlerLibrary::create_native_wrapper(const methodHandle& 
method) {
   . . .
   if (nm != NULL) {
     . . .
nm->post_compiled_method_load_event(); <== !!
   }
}

The nmethod::post_compiled_method_load_event() creates a JvmtiDeferredEvent
compiled_load_event entry and enqueues it for the ServiceThread processing.

void nmethod::post_compiled_method_load_event() {

   Method* moop = method();
   HOTSPOT_COMPILED_METHOD_LOAD(
       (char *) moop->klass_name()->bytes(),
       moop->klass_name()->utf8_length(),
       (char *) moop->name()->bytes(),
       moop->name()->utf8_length(),
       (char *) moop->signature()->bytes(),
       moop->signature()->utf8_length(),
       insts_begin(), insts_size());

   if (JvmtiExport::should_post_compiled_method_load() ||
       JvmtiExport::should_post_compiled_method_unload()) {
     get_and_cache_jmethod_id();
   }

   if (JvmtiExport::should_post_compiled_method_load()) {
     // Let the Service thread (which is a real Java thread) post the event
     MutexLockerEx ml(Service_lock, Mutex::_no_safepoint_check_flag);
     JvmtiDeferredEventQueue::enqueue(
JvmtiDeferredEvent::compiled_method_load_event(this)); <== !!
   }
}


The vmtiDeferredEvent::compiled_method_load_event() below uses
nmethodLocker::lock_nmethod(nm) to protect the nmethod* nm
from being updated:

JvmtiDeferredEvent JvmtiDeferredEvent::compiled_method_load_event(
     nmethod* nm) {
   JvmtiDeferredEvent event = JvmtiDeferredEvent(TYPE_COMPILED_METHOD_LOAD);
   event._event_data.compiled_method_load = 
nm;                              <== !!
   // Keep the nmethod alive until the ServiceThread can process
   // this deferred event.
nmethodLocker::lock_nmethod(nm); <== !!
   return event;
}

void JvmtiDeferredEventQueue::enqueue(const JvmtiDeferredEvent& event) {
   assert(Service_lock->owned_by_self(), "Must own Service_lock");

   process_pending_events();

   // Events get added to the end of the queue (and are pulled off the 
front).
   QueueNode* node = new QueueNode(event);
   if (_queue_tail == NULL) {
     _queue_tail = _queue_head = node;
   } else {
     assert(_queue_tail->next() == NULL, "Must be the last element in 
the list");
     _queue_tail->set_next(node);
     _queue_tail = node;
   }

   Service_lock->notify_all();
   assert((_queue_head == NULL) == (_queue_tail == NULL),
          "Inconsistent queue markers");
}

JvmtiDeferredEvent JvmtiDeferredEventQueue::dequeue() {
   assert(Service_lock->owned_by_self(), "Must own Service_lock");

   process_pending_events();

   assert(_queue_head != NULL, "Nothing to dequeue");

   if (_queue_head == NULL) {
     // Just in case this happens in product; it shouldn't but let's not 
crash
     return JvmtiDeferredEvent();
   }

   QueueNode* node = _queue_head;
   _queue_head = _queue_head->next();
   if (_queue_head == NULL) {
     _queue_tail = NULL;
   }

   assert((_queue_head == NULL) == (_queue_tail == NULL),
          "Inconsistent queue markers");

   JvmtiDeferredEvent event = node->event();
   delete node;
   return event;
}

The ServiceThread (different thread which is a JavaThread)
takes the event from the JvmtiDeferredEventQueue queue
and posts it.

void ServiceThread::service_thread_entry(JavaThread* jt, TRAPS) {
   . . .
       if (has_jvmti_events) {
         jvmti_event = JvmtiDeferredEventQueue::dequeue();           <== !!
       }
     }

     if (has_jvmti_events) {
jvmti_event.post(); <== !!
     }
   . . .
}

The JvmtiDeferredEvent::post() below makes the call:
   JvmtiExport::post_compiled_method_load(nm);

and only then unlocks the nmethod nm with nmethodLocker::unlock_nmethod(nm).


void JvmtiDeferredEvent::post() {
   assert(ServiceThread::is_service_thread(Thread::current()),
          "Service thread must post enqueued events");
   switch(_type) {
     case TYPE_COMPILED_METHOD_LOAD: {
       nmethod* nm = _event_data.compiled_method_load;
       JvmtiExport::post_compiled_method_load(nm); <== !!
       // done with the deferred event so unlock the nmethod
       nmethodLocker::unlock_nmethod(nm); <== !!
       break;
     }
     . . .
     default:
       ShouldNotReachHere();
   }
}

The printed disassembly shows where we observe the crash in the 
functions below:

void JvmtiExport::post_compiled_method_load(nmethod *nm) {
   if (JvmtiEnv::get_phase() < JVMTI_PHASE_PRIMORDIAL) {
     return;
   }
   JavaThread* thread = JavaThread::current();

   EVT_TRIG_TRACE(JVMTI_EVENT_COMPILED_METHOD_LOAD,
                  ("[%s] method compile load event triggered",
                  JvmtiTrace::safe_get_thread_name(thread)));

   JvmtiEnvIterator it;
   for (JvmtiEnv* env = it.first(); env != NULL; env = it.next(env)) {
     if (env->is_enabled(JVMTI_EVENT_COMPILED_METHOD_LOAD)) {
       if (env->phase() == JVMTI_PHASE_PRIMORDIAL) {
         continue;
       }
       EVT_TRACE(JVMTI_EVENT_COMPILED_METHOD_LOAD,
                 ("[%s] class compile method load event sent %s.%s ",
                 JvmtiTrace::safe_get_thread_name(thread),
                 (nm->method() == NULL) ? "NULL" : 
nm->method()->klass_name()->as_C_string(),
                 (nm->method() == NULL) ? "NULL" : 
nm->method()->name()->as_C_string()));
       ResourceMark rm(thread);
       HandleMark hm(thread);

       // Add inlining information
       jvmtiCompiledMethodLoadInlineRecord* inlinerecord = 
create_inline_record(nm);
       // Pass inlining information through the void pointer
       JvmtiCompiledMethodLoadEventMark jem(thread, nm, 
inlinerecord);                     <== !!
       JvmtiJavaThreadEventTransition jet(thread);
       jvmtiEventCompiledMethodLoad callback = 
env->callbacks()->CompiledMethodLoad;
       if (callback != NULL) {
         (*callback)(env->jvmti_external(), jem.jni_methodID(),
                     jem.code_size(), jem.code_data(), jem.map_length(),
                     jem.map(), jem.compile_info());
       }
     }
   }
}

class JvmtiCompiledMethodLoadEventMark : public JvmtiMethodEventMark {
   . . .
  public:
   JvmtiCompiledMethodLoadEventMark(JavaThread *thread, nmethod *nm, 
void* compile_info_ptr = NULL)
           : JvmtiMethodEventMark(thread,methodHandle(thread, 
nm->method())) {           <== !!
      . . .
    }
   . . .
}

class JvmtiMethodEventMark : public JvmtiThreadEventMark {
   . . .
   JvmtiMethodEventMark(JavaThread *thread, methodHandle method) :
     JvmtiThreadEventMark(thread),
     _mid(to_jmethodID(method)) 
{};                                                     <== line 234 !!
   jmethodID jni_methodID() { return _mid; }
};


All the fragments above show the code path correctly operates with nmethod
by protecting it with the nmethodLocker::lock_nmethod(nm) until the event
has been posted.

There can be several guesses about the above:
   - the above is correct but some other code abuses the convention and
     updates the nmethod without grabbing the nmethod lock
   - the above operations with the nmethod is incorrect (miss some steps)
   - some other reason (I doubt it is the case)


One conclusion is that it is unlikely an issue in the JVMTI agent as I 
suspected initially.
It would still be valuable to have more details about a scenario that 
reproduced this issue.
It is probably enough information to file a bug.

Any opinions on the above?
Any ideas about what can be broken?

Thanks,
Serguei


On 5/8/18 23:00, David Holmes wrote:
> On 9/05/2018 3:06 PM, Ioi Lam wrote:
>> Serguei & Nezih,
>>
>> I've looked at the crash log and it seems like nm->method() is NULL, 
>> which caused the crash.
>>
>> I am not familiar with the JVMTI internals or nmethods, so I am just 
>> guessing here -- is it legal for nm->method() to be NULL? Or, is it 
>> possible that someone has cleared nm->method() before it's processed 
>> by post_compiled_method_load?
>
> I was wondering the same thing. Vladimir K. pointed me at:
>
> http://hg.openjdk.java.net/jdk/jdk/file/ae0ebd3cf949/src/hotspot/share/runtime/sweeper.hpp#l42 
>
>
> for the nmethod lifecycle. Given this is actually a deferred event, I 
> too wonder if the nmethod may have been made not-entrant or even 
> zombie, since the event was posted.
>
> David
> -----
>
>> Details from the crash log:
>>
>> =================================
>> Native frames: (J=compiled Java code, A=aot compiled Java code, 
>> j=interpreted, Vv=VM code, C=native code)
>> V  [libjvm.so+0x99e365] 
>> JvmtiExport::post_compiled_method_load(nmethod*)+0x275
>> V  [libjvm.so+0x9a6614]  JvmtiDeferredEvent::post()+0x44
>> V  [libjvm.so+0xc2c507] 
>> ServiceThread::service_thread_entry(JavaThread*, Thread*)+0x337
>> V  [libjvm.so+0xceee58]  JavaThread::thread_main_inner()+0xd8
>> V  [libjvm.so+0xb6de12]  thread_native_entry(Thread*)+0xf2
>> C  [libpthread.so.0+0x7e25]  start_thread+0xc5
>> ...
>> R14=0x0000000000000000 is an unknown value
>> =================================
>>
>> Here's a disassembly of JvmtiExport::post_compiled_method_load from 
>> JDK 9.0.1
>>
>> http://hg.openjdk.java.net/jdk9/jdk9/hotspot/file/b756e7a2ec33/src/share/vm/prims/jvmtiExport.cpp 
>>
>>
>> 2051     HandleMark hm(thread);
>>     0x00007ffff67dd239 <+329>:    mov    %rax,-0x140(%rbp)
>>     0x00007ffff67dd240 <+336>:    callq  0x7ffff66149d0 
>> <HandleMark::initialize(Thread*)>
>>
>> 2054  jvmtiCompiledMethodLoadInlineRecord* inlinerecord = 
>> create_inline_record(nm);
>>     0x00007ffff67dd245 <+341>:    mov    %r12,%rdi
>>     0x00007ffff67dd248 <+344>:    callq  0x7ffff67d4090 
>> <create_inline_record(nmethod*)>
>>
>> 2055      // Pass inlining information through the void pointer
>> 2056      JvmtiCompiledMethodLoadEventMark jem(thread, nm, 
>> inlinerecord);
>>     0x00007ffff67dd24d <+349>:    mov    %rax,-0x1a8(%rbp)
>>     0x00007ffff67dd254 <+356>:    mov    0x70(%r12),%rax
>>     0x00007ffff67dd259 <+361>:    mov    %rbx,-0x178(%rbp)
>>     0x00007ffff67dd260 <+368>:    test   %rax,%rax
>>     0x00007ffff67dd263 <+371>:    mov    %rax,-0x180(%rbp)
>>     0x00007ffff67dd26a <+378>:    je     0x7ffff67dd282 
>> <JvmtiExport::post_compiled_method_load(nmethod*)+402>
>>     0x00007ffff67dd26c <+380>:    mov    %rax,-0xa0(%rbp)
>>     0x00007ffff67dd273 <+387>:    mov    0x160(%rbx),%rdi
>>     0x00007ffff67dd27a <+394>:    mov    %r13,%rsi
>>     0x00007ffff67dd27d <+397>:    callq  0x7ffff671d5f0 
>> <GrowableArray<Metadata*>::append(Metadata* const&)>
>>     0x00007ffff67dd282 <+402>:    lea    0x218(%rbx),%rax
>>     0x00007ffff67dd289 <+409>:    mov    %rbx,-0xf0(%rbp)
>>     0x00007ffff67dd290 <+416>:    movl   $0x0,-0xe0(%rbp)
>>     0x00007ffff67dd29a <+426>:    mov    %rax,-0xe8(%rbp)
>>     0x00007ffff67dd2a1 <+433>:    mov    0x3c8(%rbx),%rax
>>     0x00007ffff67dd2a8 <+440>:    test   %rax,%rax
>>     0x00007ffff67dd2ab <+443>:    je     0x7ffff67dd2b6 
>> <JvmtiExport::post_compiled_method_load(nmethod*)+454>
>>     0x00007ffff67dd2ad <+445>:    mov    0x10(%rax),%eax
>>     0x00007ffff67dd2b0 <+448>:    mov    %eax,-0xe0(%rbp)
>>     0x00007ffff67dd2b6 <+454>:    mov    0x38(%rbx),%r14
>>     0x00007ffff67dd2ba <+458>:    mov    %rbx,%rdi
>>     0x00007ffff67dd2bd <+461>:    callq  0x7ffff6707d50 
>> <JNIHandleBlock::allocate_block(Thread*)>
>>     0x00007ffff67dd2c2 <+466>:    lea    0x1f8(%rbx),%rdi
>>     0x00007ffff67dd2c9 <+473>:    mov    %rbx,%rsi
>>     0x00007ffff67dd2cc <+476>:    mov    %r14,0x118(%rax)
>>     0x00007ffff67dd2d3 <+483>:    mov    %rax,0x38(%rbx)
>>     0x00007ffff67dd2d7 <+487>:    callq  0x7ffff6559410 
>> <JavaFrameAnchor::make_walkable(JavaThread*)>
>>     0x00007ffff67dd2dc <+492>:    mov    0x1f0(%rbx),%rsi
>>     0x00007ffff67dd2e3 <+499>:    mov    -0xf0(%rbp),%rdi
>>     0x00007ffff67dd2ea <+506>:    callq  0x7ffff6708520 
>> <JNIHandles::make_local(Thread*, oopDesc*)>
>>     0x00007ffff67dd2ef <+511>:    mov    -0x1a0(%rbp),%rsi
>>     0x00007ffff67dd2f6 <+518>:    mov    %rax,-0xd8(%rbp)
>>     0x00007ffff67dd2fd <+525>:    lea    -0x170(%rbp),%rax
>>     0x00007ffff67dd304 <+532>:    mov    %rax,%rdi
>>     0x00007ffff67dd307 <+535>:    mov    %rax,-0x1b0(%rbp)
>>
>>
>> //  JvmtiMethodEventMark(JavaThread *thread, methodHandle method) :
>> //    JvmtiThreadEventMark(thread),
>> 234    _mid(to_jmethodID(method)) {};  <-- this line
>>
>>     0x00007ffff67dd30e <+542>:    callq  0x7ffff66140a0 
>> <methodHandle::methodHandle(methodHandle const&)>
>>     0x00007ffff67dd313 <+547>:    mov    -0x170(%rbp),%r14
>>     0x00007ffff67dd31a <+554>:    movq   $0x0,-0x98(%rbp)
>>     0x00007ffff67dd325 <+565>:    test   %r14,%r14
>>     0x00007ffff67dd328 <+568>:    mov    %r14,-0xa0(%rbp)
>>     0x00007ffff67dd32f <+575>:    je     0x7ffff67dd365 
>> <JvmtiExport::post_compiled_method_load(nmethod*)+629>
>>
>>     0x00007ffff67dd331 <+577>:    data16 lea 
>> 0x946ca7(%rip),%rdi        # 0x7ffff7123fe0
>>     0x00007ffff67dd339 <+585>:    data16 data16 callq 0x7ffff60b11d0 
>> <__tls_get_addr at plt>
>>     0x00007ffff67dd341 <+593>:    lea    -0x188(%rbp),%rsi
>>     0x00007ffff67dd348 <+600>:    mov    %r14,-0x188(%rbp)
>>     0x00007ffff67dd34f <+607>:    mov    (%rax),%rax
>>     0x00007ffff67dd352 <+610>:    mov    %rax,-0x98(%rbp)
>>     0x00007ffff67dd359 <+617>:    mov    0x160(%rax),%rdi
>>     0x00007ffff67dd360 <+624>:    callq  0x7ffff671d5f0 
>> <GrowableArray<Metadata*>::append(Metadata* const&)>
>>
>>  >>> CRASH %r14 == 0x0
>> 198 jmethodID to_jmethodID(methodHandle method) { return 
>> method->jmethod_id(); }
>>
>>     0x00007ffff67dd365 <+629>:    mov    0x8(%r14),%rax
>>
>>
>> Hope this helps
>> - Ioi
>>
>> On 5/8/18 4:51 PM, serguei.spitsyn at oracle.com wrote:
>>> Hi Nezih,
>>>
>>> Crashing in the JvmtiExport::post_compiled_method_load function
>>> does not prove itself that it is a VM issue.
>>> And there are mo specific details in the log that help to understand 
>>> this.
>>> Filing a bug with this log is almost useless as it would not help to 
>>> get to the root cause.
>>> Also, there is still a pretty big chance the real problem is in the 
>>> agent.
>>>
>>> My advice is to get as more details about this crash as possible.
>>> The best you can do is to provide a stand alone test case.
>>> Even if the issue is in the agent it will help to figure it out.
>>>
>>> Thanks,
>>> Serguei
>>>
>>>
>>> On 5/8/18 16:38, nezih yigitbasi wrote:
>>>> Hi Serguei,
>>>> We don't have a repro unfortunately. We have seen this only once so 
>>>> far in our production environment.
>>>>
>>>> Do you think this can be an agent issue? I thought like if it was 
>>>> an agent issue the stack trace would show it. The stack trace, 
>>>> however, ends in the VM code -- the crash happens in the 
>>>> "JvmtiExport::post_compiled_method_load" method.
>>>>
>>>> Thanks,
>>>> Nezih
>>>>
>>>> 2018-05-08 16:34 GMT-07:00 serguei.spitsyn at oracle.com 
>>>> <mailto:serguei.spitsyn at oracle.com> <serguei.spitsyn at oracle.com 
>>>> <mailto:serguei.spitsyn at oracle.com>>:
>>>>
>>>>     Hi Nezih,
>>>>
>>>>     You error file with log does not help much.
>>>>     Could you, please, reproduce this issue with the debug or
>>>>     fastdebug build?
>>>>     Also, a standalone test case is better to provide to
>>>>     increase chances the issue to be investigated and fixed.
>>>>
>>>>     In general, a JVMTI agent has to be written very carefully as it
>>>>     can easily cause the VM to crash.
>>>>     So, it makes sense to look at the agent implementation first to
>>>>     make sure it does not do anything wrong.
>>>>     It is what we normally do first and why a test case that
>>>>     demonstrates the problem is needed.
>>>>
>>>>     Thanks,
>>>>     Serguei
>>>>
>>>>
>>>>
>>>>     On 5/8/18 16:00, nezih yigitbasi wrote:
>>>>>     Hi,
>>>>>
>>>>>     First of all, sorry for bringing up a crash issue in this dev
>>>>>     mailing list, but the crash report submission page
>>>>>     (https://bugreport.java.com/bugreport/crash.jsp
>>>>> <https://bugreport.java.com/bugreport/crash.jsp>) doesn't list
>>>>>     Java 9 in the release drop down, so I couldn't report it there.
>>>>>
>>>>>     We recently got a crash with Java 9.0.1+11 with an interesting
>>>>>     stack ending at "JvmtiExport::post_compiled_method_load()"
>>>>>     (entire error file is here
>>>>> <https://gist.github.com/nezihyigitbasi/52a58698cc9acfcab21c69d00bd0cef2>).
>>>>>     A google search didn't end up with much info, so I just wanted
>>>>>     to check with this mailing list to see whether anyone has any
>>>>>     ideas to investigate this further.
>>>>>
>>>>>     Thanks,
>>>>>     Nezih
>>>>>
>>>>
>>>>
>>>
>>



More information about the serviceability-dev mailing list