RFR(XXS) 8218558: NMT stack traces in output should show mt component for virtual memory allocations

coleen.phillimore at oracle.com coleen.phillimore at oracle.com
Thu Feb 7 16:58:01 UTC 2019


This looks good!
Coleen

On 2/6/19 2:27 PM, zgu at redhat.com wrote:
> On Wed, 2019-02-06 at 20:07 +0100, Thomas Stüfe wrote:
>> Hi Zhengyu,
>>
>> looks good.
> Thanks again.
>
>> Since _flag is immutable, you could make it const. But I leave that
>> up to you, the change looks fine to me otherwise. I do not need
>> another webrev.
> Unfortunately, it does not work. There are assignments that invoke its
> default assignment operator.
>
> -Zhengyu
>> ..Thomas
>>
>> On Wed, Feb 6, 2019 at 7:19 PM <zgu at redhat.com> wrote:
>>> On Wed, 2019-02-06 at 17:14 +0100, Thomas Stüfe wrote:
>>>> Hi Zhengyu,
>>>>
>>>> Looks good.
>>>>
>>>> Could you not move MEMFLAGS _flag; from the child classes up to
>>> class
>>>> AllocationSite ? Since AFAICS all child classes now have
>>> Memflags.
>>>
>>> Good point!
>>>
>>> Updated: http://cr.openjdk.java.net/~zgu/JDK-8218558/webrev.01/
>>>
>>> Reran hotspot_nmt, the same results.
>>>
>>> Thanks,
>>>
>>> -Zhengyu
>>>
>>>
>>>> Cheers, Thomas
>>>>
>>>>
>>>>
>>>>
>>>> On Wed, Feb 6, 2019 at 4:25 PM <zgu at redhat.com> wrote:
>>>>> Please review this small enhancement that outputs memory types
>>> for
>>>>> virtual memory allocations [1]. This mirrors JDK-8139673, which
>>>>> outputs
>>>>> memory types for malloc allocations.
>>>>>
>>>>> Bug: https://bugs.openjdk.java.net/browse/JDK-8218558
>>>>> Webrev: http://cr.openjdk.java.net/~zgu/JDK-8218558/webrev.00/i
>>> ndex
>>>>> .htm
>>>>> l
>>>>>
>>>>> Test:
>>>>>    hotspot_nmt on Linux x64 (fastdebug and release),
>>>>>    passed all except
>>>>> runtime/NMT/CheckForProperDetailStackTrace.java
>>>>>    (JDK-8218458) with release build.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> -Zhengyu
>>>>>
>>>>> [1] New outputs:
>>>>>
>>>>> detail:
>>>>> [0x00007fb9ea533218] SafepointMechanism::initialize()+0x48
>>>>> [0x00007fb9ea7cd6bc] Threads::create_vm(JavaVMInitArgs*,
>>>>> bool*)+0x19c
>>>>> [0x00007fb9e9ea5639] JNI_CreateJavaVM+0x69
>>>>> [0x00007fb9eb9c9176] JavaMain+0x86
>>>>>                               (reserved=8KB, committed=8KB
>>>>> Type=Safepoint)
>>>>>
>>>>> detail.diff:
>>>>> [0x00007ff269ad84b4] ReservedSpace::ReservedSpace(unsigned
>>> long,
>>>>> unsigned long)+0x1a4
>>>>> [0x00007ff268fd56a7] CodeHeap::reserve(ReservedSpace, unsigned
>>>>> long,
>>>>> unsigned long)+0x327
>>>>> [0x00007ff268bd84df] CodeCache::add_heap(ReservedSpace, char
>>>>> const*,
>>>>> int)+0x13f
>>>>> [0x00007ff268bd8bbe] CodeCache::initialize_heaps()+0x4be
>>>>>                               (mmap: reserved=940KB,
>>> committed=64KB
>>>>> +4KB
>>>>> Type=Code)



More information about the hotspot-runtime-dev mailing list