RFR(XXS) 8068592 Remove unused code in objectMonitor.hpp

Daniel D. Daugherty daniel.daugherty at oracle.com
Thu Jun 9 16:05:14 UTC 2016


Claes and Jerry,

It would be good if you guys could check out the latest version
of this fix... Thanks!

Dan


On 6/8/16 7:57 PM, Daniel D. Daugherty wrote:
> On 6/8/16 6:23 PM, David Holmes wrote:
>> On 9/06/2016 6:37 AM, Daniel D. Daugherty wrote:
>>> Done pulling the loose thread:
>>>
>>> Webrev URL:
>>> http://cr.openjdk.java.net/~dcubed/8068592-webrev/1-jdk9-hs-hotspot/
>>
>> I'm glad it didn't unravel too far :)
>
> I had to exercise great restraint in order to avoid making any more
> white space fixes... :-)
>
>
>> Changes look good!
>
> Thanks for the re-review!
>
> Dan
>
>
>>
>> Thanks,
>> David
>>
>>> This version changes three files:
>>>
>>> src/share/vm/runtime/objectMonitor.cpp
>>> src/share/vm/runtime/objectMonitor.hpp
>>> src/share/vm/runtime/synchronizer.hpp
>>>
>>> We've gone from 12 lines changed (mostly deletes)
>>> to 34 lines changed (still mostly deletes); still
>>> considering this one to "XXS"...
>>>
>>> Passes JPRT test builds...
>>>
>>> Thanks, in advance, for any comments, suggestions or feedback.
>>>
>>> Dan
>>>
>>>
>>>
>>>
>>> On 6/8/16 8:54 AM, Daniel D. Daugherty wrote:
>>>> At the risk of pulling the loose thread on the sweater, I'll take a 
>>>> look
>>>> at the two BTWs below... :-)
>>>>
>>>> Dan
>>>>
>>>>
>>>> On 6/7/16 5:01 PM, David Holmes wrote:
>>>>> BTW you could (but don't have to) get rid of this in 
>>>>> objectMonitor.cpp
>>>>>
>>>>> #define TrySpin TrySpin_VaryDuration
>>>>>
>>>>> and rename TrySpin_Duration to TrySpin, now there are no other
>>>>> options to select.
>>>>>
>>>>> BTW2: You might also grep for SpinCallback - there seems to be some
>>>>> infrastructure for that unused facility as well.
>>>>>
>>>>> David
>>>>>
>>>>> On 8/06/2016 7:01 AM, David Holmes wrote:
>>>>>> Hi Dan,
>>>>>>
>>>>>> Looks good. I agree this is all unused.
>>>>>>
>>>>>> Thanks,
>>>>>> David
>>>>>>
>>>>>> On 8/06/2016 6:55 AM, Daniel D. Daugherty wrote:
>>>>>>> Greetings,
>>>>>>>
>>>>>>> I have a fix ready for the following RFE:
>>>>>>>
>>>>>>>     JDK-8068592 Remove unused code in objectMonitor.hpp
>>>>>>>     https://bugs.openjdk.java.net/browse/JDK-8068592
>>>>>>>
>>>>>>> Yes, this an RFE and I'm fixing it now for a few reasons:
>>>>>>>
>>>>>>> 1) I promised Claes Redestad back on 04.25 that this would be the
>>>>>>> first
>>>>>>>    issue I worked on when I got my Solaris-X64 environment back.
>>>>>>> mt-haku
>>>>>>>    is now running Solaris 11.2 SRU5.5 so it is back!
>>>>>>>
>>>>>>> 2) This is the simplest change in my queue with little to no risk
>>>>>>> so it
>>>>>>>    makes the perfect test case for my upgrade of mt-haku from 
>>>>>>> Solaris
>>>>>>>    10u11 -> Solaris 11.2 SRU5.5.
>>>>>>>
>>>>>>> 3) This change gets rid of unused fields, unused macro assembler
>>>>>>> functions
>>>>>>>    and a couple of declared but not defined functions from
>>>>>>> ObjectMonitor.
>>>>>>>    Slimming down ObjectMonitor is never a bad thing!
>>>>>>>
>>>>>>> Yes, I know that the process for getting approval for RFEs is 
>>>>>>> not yet
>>>>>>> defined. However, I want to get the code review out of the way 
>>>>>>> so that
>>>>>>> this changeset can go in if it is approved.
>>>>>>>
>>>>>>> Sanity check: I searched a JDK9-hs clone for the following 
>>>>>>> patterns:
>>>>>>>
>>>>>>>     FreeNext_offset_in_bytes
>>>>>>>     Responsible_offset_in_bytes
>>>>>>>     Spinner_offset_in_bytes
>>>>>>>     TrySpin_Fixed
>>>>>>>     TrySpin_VaryFrequency
>>>>>>>     WaitSet_offset_in_bytes
>>>>>>>     _SpinClock
>>>>>>>     _SpinFreq
>>>>>>>     _SpinState
>>>>>>>
>>>>>>> and found no mentions outside of the ones that I'm deleting from
>>>>>>> src/share/vm/runtime/objectMonitor.hpp.
>>>>>>>
>>>>>>> Webrev URL:
>>>>>>> http://cr.openjdk.java.net/~dcubed/8068592-webrev/0-jdk9-hs-hotspot/ 
>>>>>>>
>>>>>>>
>>>>>>> Testing: JPRT test builds are in process.
>>>>>>>
>>>>>>> Thanks, in advance, for any comments, suggestions or feedback.
>>>>>>>
>>>>>>> Dan
>>>>
>>>>
>>>
>
>



More information about the hotspot-runtime-dev mailing list