RFR (trivial) 8068593: Remove or repurpose unused PerfCounters from objectMonitor
coleen.phillimore at oracle.com
coleen.phillimore at oracle.com
Wed Feb 21 19:01:41 UTC 2018
On 2/21/18 12:55 PM, Daniel D. Daugherty wrote:
> On 2/21/18 11:55 AM, coleen.phillimore at oracle.com wrote:
>>
>> Also, still trivial, I removed two now unused "Knobs".
>>
>> open webrev at http://cr.openjdk.java.net/~coleenp/8068593.02/webrev
>
> src/hotspot/share/runtime/objectMonitor.cpp
> No comments.
>
> src/hotspot/share/runtime/objectMonitor.hpp
> No comments.
>
> Thumbs up!
>
> Reminder: Please don't forget to run your copyright update script.
Thanks Dan, no problem. I will run my script.
>
> Dan
>
> P.S.
> Here's the count of remaining uses of the OM_PERFDATA_OP macro:
>
> $ grep -r OM_PERFDATA_OP src/hotspot/share/runtime | \
> sed -e '/#define/d' -e 's/OM_PERFDATA_OP(//' -e 's/,.*//' -e 's/.*:
> *//' | \
> sort | uniq -c
> 1 ContendedLockAttempts
> 1 Deflations
> 2 FutileWakeups
> 2 Inflations
> 1 MonExtant
> 3 Notifications
> 1 Parks
>
Not sure what this means but these other counters are incremented at least.
runtime/objectMonitor.hpp: static PerfCounter *
_sync_ContendedLockAttempts;
runtime/objectMonitor.cpp: OM_PERFDATA_OP(ContendedLockAttempts, inc());
runtime/objectMonitor.cpp:PerfCounter *
ObjectMonitor::_sync_ContendedLockAttempts = NULL;
runtime/objectMonitor.cpp: NEWPERFCOUNTER(_sync_ContendedLockAttempts);
Thanks,
Coleen
>>
>> thanks,
>> Coleen
>>
>> On 2/21/18 9:40 AM, coleen.phillimore at oracle.com wrote:
>>>
>>>
>>> On 2/21/18 9:37 AM, Claes Redestad wrote:
>>>> Hi,
>>>>
>>>> objectMonitor.cpp, maybe remove the now dead block?
>>>>
>>>> if (Knob_LogSpins == 0) {
>>>> - ObjectMonitor::_sync_FailedSpins = NULL;
>>>> }
>>>>
>>>> Otherwise looks good to me, but you may want a review from someone
>>>> less biased. :-)
>>>
>>> Yes, I do. thanks for noticing this.
>>>
>>> Coleen
>>>
>>>>
>>>> /Claes
>>>>
>>>>
>>>> On 2018-02-21 15:34, coleen.phillimore at oracle.com wrote:
>>>>> Just removed them and ran tier1 tests. The ones remaining are
>>>>> deemed useful. See bug for details.
>>>>>
>>>>> open webrev at http://cr.openjdk.java.net/~coleenp/8068593.01/webrev
>>>>> bug link https://bugs.openjdk.java.net/browse/JDK-8068593
>>>>>
>>>>> Thanks,
>>>>> Coleen
>>>>
>>>
>>
>
More information about the hotspot-runtime-dev
mailing list