RFR: 8194312: Support parallel and concurrent JNI global handle processing

[serguei.spitsyn at oracle.com]

On 1/8/18 16:58, Kim Barrett wrote:
>> On Jan 8, 2018, at 7:20 PM, serguei.spitsyn at oracle.com wrote:
>>
>> Hi Kim,
>>
>> A couple of quick minor comments.
> Thanks.
>
>> http://cr.openjdk.java.net/~kbarrett/8194312/open.00/src/jdk.hotspot.agent/share/classes/sun/jvm/hotspot/utilities/PointerLocation.java.frames.html
>>   241     } else if (isInStrongGlobalJNIHandles() ||
>>   242                isInWeakGlobalJNIHandles()) {
>>
>>   243       tty.print("In ");
>>
>>   244       if (isInStrongGlobalJNIHandles()) {
>>
>>   245         tty.print("strong global");
>>
>>   246       } else if (isInWeakGlobalJNIHandles()) {
>>
>>   247         tty.print("weak global");
>>   248       }
>>
>>   A suggestion to simplify it a little bit as follows:
>>        } else if (isInStrongGlobalJNIHandles()) {
>>
>>          tty.print("In strong global");
>>
>>        } else if (isInWeakGlobalJNIHandles()) {
>>
>>          tty.print("In weak global");
>>        }
> Good idea.  Will do.
>
>> http://cr.openjdk.java.net/~kbarrett/8194312/open.00/src/hotspot/share/gc/shared/oopStorage.cpp.html
>>    61 void OopStorage::BlockList::push_back(const Block& block) {
>>    62   const Block* old = _tail;
>>    63   if (old == NULL) {
>>    64     assert(_tail == NULL, "invariant");
>>    65     _head = _tail = █
>>    66   } else {
>>    67     _get_entry(*old)._next = █
>>    68     _get_entry(block)._prev = old;
>>    69     _tail = █
>>    70   }
>>    71 }
>>
>>    It seems, the assert should check _head instead of _tail.
> Oops!  Will do.

Thanks!

Also, this fragment looks strange:

  354   if (block->is_empty()) {
  355     log_debug(oopstorage, blocks)("%s: block not empty " PTR_FORMAT, name(), p2i(block));
  356     --_empty_block_count;
  357   }


Should the message tell "block is empty" instead of "block not empty"?

Thanks,
Serguei