(S) RFR: 8147379: Investigate if ConvertSleepToYield still should be false by default on Sparc

David Holmes david.holmes at oracle.com
Wed Feb 10 21:12:54 UTC 2016


On 11/02/2016 1:10 AM, Daniel D. Daugherty wrote:
> On 2/9/16 5:45 PM, David Holmes wrote:
>> bug: https://bugs.openjdk.java.net/browse/JDK-8147379
>>
>> webrev:  http://cr.openjdk.java.net/~dholmes/8147379/webrev/
>
> .hgignore
>      No comments on your change. There be a bit of JVM/CI cruft here...

Oops that shouldn't have been in this webrev - please ignore it. 
Seperate bug filed.

Thanks for the Review Dan.

David
-----

> src/cpu/aarch64/vm/globals_aarch64.hpp
>      No comments.
>
> src/cpu/ppc/vm/globals_ppc.hpp
>      No comments.
>
> src/cpu/sparc/vm/globals_sparc.hpp
>      No comments.
>
> src/cpu/x86/vm/globals_x86.hpp
>      No comments.
>
> src/cpu/zero/vm/globals_zero.hpp
>      No comments.
>
> src/share/vm/runtime/arguments.cpp
>      No comments. (after doing a double take)
>
> src/share/vm/runtime/globals.hpp
>      No comments.
>
> test/runtime/CommandLine/VMDeprecatedOptions.java
>      No comments.
>
> Thumbs up!
>
> Dan
>
>
>>
>> ConvertSleepToYield is a product VM flag that should only have been
>> defined for Solaris, and was by "accident" only set differently
>> (false) on Solaris sparc. It relates to a GUI responsiveness issue in
>> some ancient Solaris version, the details of which have been lost (AWT
>> folk had no input). Changing this setting serves no useful purpose
>> today. So we remove the pd nature of this flag and just make it a
>> shared flag that defaults to true. We also deprecate it.
>>
>> ConvertYieldToSleep is a product VM flag that should only have been
>> provided for Windows, but was made available on all platforms. It's
>> role was to allow for backward compatibility with the Classic VM.
>> Changing this setting serves no useful purpose today. It is also
>> deprecated.
>>
>> These changes have been approved by our CCC process.
>>
>> Testing: - JPRT
>>          - updated the VMDeprecatedOptions test
>>
>> That was a lot of words for a very small change. :)
>>
>> Thanks,
>> David
>


More information about the hotspot-runtime-dev mailing list